Datasets:

LatestEval

/

arxiv-latest

like 0

Follow

LatestEval 1

arxiv

2312.11734v1

INTRODUCTION Parabolic graded refractive index (GRIN) fibers have a smoothly varying refractive index profile that decreases radially from the fiber axis that provides reduced modal dispersion compared to step-index fibers . Their normal modes are well-understood with the analytic description of their normal modes, propagation constants, and cladding-mode cutoffs . Beyond telecommunications, they find utility in strain and temperature sensing , as well as medical imaging to mention a few examples. In the ideal scenario of infinite parabolic GRIN media, the propagation of Hermite- and Laguerre- , as well as Ince-Gauss modes have been studied, leading to the discovery of hypergeometric breathing and pendulum-type modes. The reason for focusing on realistic low-contrast parabolic GRIN fibers in this research is specified here. Employing the weak guidance approximation, we build analytic propagation invariant modes with constant radial and azimuthal numbers, find their propagation constants and cut-off relations that are in good agreement with numerical results from finite element method (FEM). We identify symmetries related to the conservation of circular and azimuthal numbers to construct two families of harmonic motion modes using analogies to coherent states in quantum optics. One family are accelerating light fields whose intensity distribution center follows a circular trajectory in the plane transverse to propagation. The other are breathing light fields whose intensity distribution width varies with propagation.

Why is the focus on realistic low-contrast parabolic GRIN fibers in this research?

The focus on realistic low-contrast parabolic GRIN fibers is to propose electromagnetic field modes that solve Maxwell equations for inhomogeneous media.

purpose

arxiv

2312.11734v1

INTRODUCTION Parabolic graded refractive index (GRIN) fibers have a smoothly varying refractive index profile that decreases radially from the fiber axis that provides reduced modal dispersion compared to step-index fibers . Their normal modes are well-understood with the analytic description of their normal modes, propagation constants, and cladding-mode cutoffs . Beyond telecommunications, they find utility in strain and temperature sensing , as well as medical imaging to mention a few examples. In the ideal scenario of infinite parabolic GRIN media, the propagation of Hermite- and Laguerre- , as well as Ince-Gauss modes have been studied, leading to the discovery of hypergeometric breathing and pendulum-type modes. Here, we focus on realistic low-contrast parabolic GRIN fibers to propose electromagnetic field modes that solve Maxwell equations for inhomogeneous media. The purpose of using the weak guidance approximation in this research is outlined here. We identify symmetries related to the conservation of circular and azimuthal numbers to construct two families of harmonic motion modes using analogies to coherent states in quantum optics. One family are accelerating light fields whose intensity distribution center follows a circular trajectory in the plane transverse to propagation. The other are breathing light fields whose intensity distribution width varies with propagation.

What is the purpose of employing the weak guidance approximation in the context of parabolic GRIN fibers?

The weak guidance approximation is employed to build analytic propagation invariant modes with constant radial and azimuthal numbers, and to find their propagation constants and cut-off relations in good agreement with numerical results from finite element method (FEM).

purpose

arxiv

2312.11734v1

INTRODUCTION Parabolic graded refractive index (GRIN) fibers have a smoothly varying refractive index profile that decreases radially from the fiber axis that provides reduced modal dispersion compared to step-index fibers . Their normal modes are well-understood with the analytic description of their normal modes, propagation constants, and cladding-mode cutoffs . Beyond telecommunications, they find utility in strain and temperature sensing , as well as medical imaging to mention a few examples. In the ideal scenario of infinite parabolic GRIN media, the propagation of Hermite- and Laguerre- , as well as Ince-Gauss modes have been studied, leading to the discovery of hypergeometric breathing and pendulum-type modes. Here, we focus on realistic low-contrast parabolic GRIN fibers to propose electromagnetic field modes that solve Maxwell equations for inhomogeneous media. Employing the weak guidance approximation, we build analytic propagation invariant modes with constant radial and azimuthal numbers, find their propagation constants and cut-off relations that are in good agreement with numerical results from finite element method (FEM). The identified symmetries and their utilization in the research are discussed here. One family are accelerating light fields whose intensity distribution center follows a circular trajectory in the plane transverse to propagation. The other are breathing light fields whose intensity distribution width varies with propagation.

What symmetries are identified in the research and how are they utilized?

The research identifies symmetries related to the conservation of circular and azimuthal numbers to construct two families of harmonic motion modes through analogies to coherent states in quantum optics.

summary

arxiv

2312.11734v1

INTRODUCTION Parabolic graded refractive index (GRIN) fibers have a smoothly varying refractive index profile that decreases radially from the fiber axis that provides reduced modal dispersion compared to step-index fibers . Their normal modes are well-understood with the analytic description of their normal modes, propagation constants, and cladding-mode cutoffs . Beyond telecommunications, they find utility in strain and temperature sensing , as well as medical imaging to mention a few examples. In the ideal scenario of infinite parabolic GRIN media, the propagation of Hermite- and Laguerre- , as well as Ince-Gauss modes have been studied, leading to the discovery of hypergeometric breathing and pendulum-type modes. Here, we focus on realistic low-contrast parabolic GRIN fibers to propose electromagnetic field modes that solve Maxwell equations for inhomogeneous media. Employing the weak guidance approximation, we build analytic propagation invariant modes with constant radial and azimuthal numbers, find their propagation constants and cut-off relations that are in good agreement with numerical results from finite element method (FEM). We identify symmetries related to the conservation of circular and azimuthal numbers to construct two families of harmonic motion modes using analogies to coherent states in quantum optics. A description of one family of the harmonic motion modes is given here. The other are breathing light fields whose intensity distribution width varies with propagation.

Can you describe one of the families of harmonic motion modes identified in parabolic GRIN fiber research?

One family of the harmonic motion modes are accelerating light fields whose intensity distribution center follows a circular trajectory in the plane transverse to propagation.

example

arxiv

2312.11734v1

INTRODUCTION Parabolic graded refractive index (GRIN) fibers have a smoothly varying refractive index profile that decreases radially from the fiber axis that provides reduced modal dispersion compared to step-index fibers . Their normal modes are well-understood with the analytic description of their normal modes, propagation constants, and cladding-mode cutoffs . Beyond telecommunications, they find utility in strain and temperature sensing , as well as medical imaging to mention a few examples. In the ideal scenario of infinite parabolic GRIN media, the propagation of Hermite- and Laguerre- , as well as Ince-Gauss modes have been studied, leading to the discovery of hypergeometric breathing and pendulum-type modes. Here, we focus on realistic low-contrast parabolic GRIN fibers to propose electromagnetic field modes that solve Maxwell equations for inhomogeneous media. Employing the weak guidance approximation, we build analytic propagation invariant modes with constant radial and azimuthal numbers, find their propagation constants and cut-off relations that are in good agreement with numerical results from finite element method (FEM). We identify symmetries related to the conservation of circular and azimuthal numbers to construct two families of harmonic motion modes using analogies to coherent states in quantum optics. One family are accelerating light fields whose intensity distribution center follows a circular trajectory in the plane transverse to propagation. An explanation of the other family of harmonic motion modes is found here.

What is the other family of harmonic motion modes discovered in the research on parabolic GRIN fibers?

The other family of harmonic motion modes are breathing light fields whose intensity distribution width varies with propagation.

example

arxiv

2312.11821v1

ANTI-DEUTERON BEAM STUDY AT J-PARC HEF K1.8 BEAM LINE M. Ukai^1,2, Y. Ishikawa^2, T. Takahashi^1, K. Tanida^3 and T. O. Yamamoto^3 ^1 Institute of Particle and Nuclear Studies, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan ^2 Department of Physics, Tohoku University, Sendai, Miyagi 980-8578, Japan ^3 Advanced Science Research Center, JAEA, Tokai, Ibaraki 319-1195, Japan (Dataed: December 19, 2023) Abstract [h]11.5cm We performed a d beam study at the K1.8 beam line of J-PARC Hadron Experimental Facility. The yield of the deuterium beam in the experiment is stated here. Corresponding p beam yield is roughly estimated to be ∼ 0.3 Mcounts/spill for the same slit condition. Then, the d/p production ratio at extraction angle of 6 degrees is estimated to be ∼ 10^-6. This is the first time measurement of the d beam yield and d/p production ratio at J-PARC. Further beam line tuning may increase the d beam yield.

What was the measured yield of the 1.8 GeV/c deuterium beam in the study?

The 1.8 GeV/c deuterium beam yield was measured to be 0.30 ± 0.04 counts/spill when a beam of 30 GeV protons at 70 × 10^12 protons/spill was irradiated on a 66 mm thick gold target.

summary

arxiv

2312.11821v1

ANTI-DEUTERON BEAM STUDY AT J-PARC HEF K1.8 BEAM LINE M. Ukai^1,2, Y. Ishikawa^2, T. Takahashi^1, K. Tanida^3 and T. O. Yamamoto^3 ^1 Institute of Particle and Nuclear Studies, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan ^2 Department of Physics, Tohoku University, Sendai, Miyagi 980-8578, Japan ^3 Advanced Science Research Center, JAEA, Tokai, Ibaraki 319-1195, Japan (Dataed: December 19, 2023) Abstract [h]11.5cm We performed a d beam study at the K1.8 beam line of J-PARC Hadron Experimental Facility. 1.8 GeV/c d beam yield was measured to be 0.30 ± 0.04 counts/spill for 30 GeV 70 × 10^12 protons/spill irradiated on a 66 mm thick of gold target with the vertical slit opening widths of 2.2 mm, 5 mm and 5 mm for intermediate focus (IFV), mass slit 1 (MS1) and 2 (MS2), respectively. Corresponding p beam yield is roughly estimated to be ∼ 0.3 Mcounts/spill for the same slit condition. The estimated d/p production ratio is discussed in this sentence. This is the first time measurement of the d beam yield and d/p production ratio at J-PARC. Further beam line tuning may increase the d beam yield.

What is the estimated deuterium to proton production ratio at an extraction angle of 6 degrees?

The deuterium to proton (d/p) production ratio at an extraction angle of 6 degrees is estimated to be approximately 10^-6.

summary

arxiv

2312.11821v1

ANTI-DEUTERON BEAM STUDY AT J-PARC HEF K1.8 BEAM LINE M. Ukai^1,2, Y. Ishikawa^2, T. Takahashi^1, K. Tanida^3 and T. O. Yamamoto^3 ^1 Institute of Particle and Nuclear Studies, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan ^2 Department of Physics, Tohoku University, Sendai, Miyagi 980-8578, Japan ^3 Advanced Science Research Center, JAEA, Tokai, Ibaraki 319-1195, Japan (Dataed: December 19, 2023) Abstract [h]11.5cm We performed a d beam study at the K1.8 beam line of J-PARC Hadron Experimental Facility. 1.8 GeV/c d beam yield was measured to be 0.30 ± 0.04 counts/spill for 30 GeV 70 × 10^12 protons/spill irradiated on a 66 mm thick of gold target with the vertical slit opening widths of 2.2 mm, 5 mm and 5 mm for intermediate focus (IFV), mass slit 1 (MS1) and 2 (MS2), respectively. Corresponding p beam yield is roughly estimated to be ∼ 0.3 Mcounts/spill for the same slit condition. Then, the d/p production ratio at extraction angle of 6 degrees is estimated to be ∼ 10^-6. The significance of the measurements performed in this study is mentioned here. Further beam line tuning may increase the d beam yield.

What is significant about the deuterium beam yield and production ratio measurements conducted at J-PARC?

This study represents the first time measurement of the deuterium beam yield and deuterium to proton production ratio at J-PARC.

summary

arxiv

2312.11821v1

ANTI-DEUTERON BEAM STUDY AT J-PARC HEF K1.8 BEAM LINE M. Ukai^1,2, Y. Ishikawa^2, T. Takahashi^1, K. Tanida^3 and T. O. Yamamoto^3 ^1 Institute of Particle and Nuclear Studies, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan ^2 Department of Physics, Tohoku University, Sendai, Miyagi 980-8578, Japan ^3 Advanced Science Research Center, JAEA, Tokai, Ibaraki 319-1195, Japan (Dataed: December 19, 2023) Abstract [h]11.5cm We performed a d beam study at the K1.8 beam line of J-PARC Hadron Experimental Facility. 1.8 GeV/c d beam yield was measured to be 0.30 ± 0.04 counts/spill for 30 GeV 70 × 10^12 protons/spill irradiated on a 66 mm thick of gold target with the vertical slit opening widths of 2.2 mm, 5 mm and 5 mm for intermediate focus (IFV), mass slit 1 (MS1) and 2 (MS2), respectively. Corresponding p beam yield is roughly estimated to be ∼ 0.3 Mcounts/spill for the same slit condition. Then, the d/p production ratio at extraction angle of 6 degrees is estimated to be ∼ 10^-6. This is the first time measurement of the d beam yield and d/p production ratio at J-PARC. A possible future action to improve the deuterium beam yield is suggested in this sentence.

What potential future action is suggested to possibly enhance the deuterium beam yield?

Further beam line tuning may increase the deuterium beam yield.

future

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. The summary of the recent analyses of scalar glueball data is provided here. A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

According to the most recent analyses of BESIII data, where is the lightest realization of a scalar glueball located in terms of energy?

The most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV.

summary

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. An example of how mixing arises in a certain approach is illustrated here. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

How does mixing arise in the functional approach to gluonic and quark-antiquark states?

In the functional approach, mixing of gluonic and quark-antiquark states arises naturally by the coupling of their amplitudes.

example

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. The terminology 'pure QCD' is defined in this section. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

What is meant by 'pure QCD' in the context of researching gluonic bound states?

Pure QCD refers to the situation when quarks are made infinitely heavy, leaving only purely gluonic bound states.

terminology

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. Examples of the success of functional methods in describing certain spectra are given here. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

Can you give an example of what functional methods are successful in describing?

Functional methods are successful in describing the spectrum of baryons or tetraquarks.

example

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. An explanation of the challenges faced by functional methods in reproducing the gluonic spectrum is given here. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

Why do functional methods face challenges in reproducing the gluonic spectrum?

One of the main reasons functional methods struggle to reproduce the gluonic spectrum is the truncations necessary for functional bound state equations.

purpose

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. The purpose of using a specific truncation in the quark sector is discussed here. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

Why is the rainbow-ladder truncation often used in the quark sector?

A simple truncation called rainbow-ladder truncation works well in many cases for the quark sector.

purpose

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. A summary of what is needed to complement the rainbow-ladder truncation is provided here. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

What does the rainbow-ladder truncation require to be effective in studies of the quark sector?

The rainbow-ladder truncation only needs to be complemented by an effective quark-gluon interaction.

summary

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . A synopsis of what this contribution offers is given here. The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. We close with a summary.

What does this contribution specifically provide in relation to the functional setup?

This contribution provides an overview of the functional setup including tests for extensions of it.

summary

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. The reason behind the discussion of the importance of higher diagrams in the truncation is presented here. We close with a summary.

Why is the importance of higher diagrams in the truncation discussed?

The importance of higher diagrams in the truncation is discussed to evaluate their impact in the calculation.

purpose

arxiv

2312.12029v1

INTRODUCTION Glueballs are elusive objects that were suggested as bound states of gluons already in the early days of QCD and searched for since then. Most recent analyses of BESIII data put their lightest realization, the scalar glueball, in the region around 1700 MeV . A complication in its identification is that several states with scalar quantum numbers exist above 1 GeV. This mixing makes analyses of experimental data challenging as well as theoretical calculations. In the functional approach used here, such a mixing arises naturally by the coupling of the amplitudes of two-gluon and quark-antiquark states, see for the corresponding equations. However, there is no solution for full QCD yet. For pure QCD, on the other hand, viz., when quarks are made infinitely heavy and thus only purely gluonic bound states remain, the situation is as follows. From lattice calculations the spectrum of the lightest states is known for quite some time . More recent calculations improved the errors and added some further states , see also for the current status. For functional methods it was challenging to reproduce this spectrum, although they are successful in describing, for example, the spectrum of baryons or tetraquarks, see and references therein. One of the main reasons is the truncations necessary for functional bound state equations. For the quark sector, a particularly simple truncation, called rainbow-ladder truncation, works in many cases quite well. It only needs to be complemented by an effective quark-gluon interaction. This truncation works reliably and produces stable results. In the gluonic sector, first of all a rainbow-like truncation as for the quark propagator does not exist. Secondly, attempts to model the interactions did not lead to predictive results . The situation is different when input is used that is calculated from a high-level truncation as done in . This contribution provides an overview of the functional setup including tests for extensions of it in Sec. . The resulting spectrum is presented in Sec. where also the importance of higher diagrams in the truncation is discussed. A closing note for the passage is mentioned here.

What does the passage conclude with?

The passage concludes with a summary of the discussed topics.

summary

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology is described here. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

What does ISAC technology refer to?

Integrated sensing and communication (ISAC) technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and aiming for mutual benefits.

terminology

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. This sentence introduces a new modulation scheme example. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

Can you provide an example of a new modulation scheme suitable for high-mobility scenarios that modulate data symbols in the delay-Doppler (DD) domain?

Orthogonal time frequency space (OTFS) is a new modulation scheme which modulates data symbols in the delay-Doppler (DD) domain.

example

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. The purpose of exploiting the quasi-static and sparse channel in the DD domain by OTFS is discussed here. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

Why is OTFS modulation considered capable of supporting multi-path and high-mobility communications?

OTFS is capable of supporting multi-path and high-mobility communications by exploiting the quasi-static and sparse channel in the DD domain.

purpose

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. This paper's primary goal is explained in this sentence. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

What is the main goal of developing the optimal precoder in OTFS-based ISAC systems?

The paper devises the optimal precoder for OTFS-based ISAC systems with the goal of evaluating communication performance and sensing accuracy using BER and CRB.

purpose

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Constraints for the BER minimization problem are detailed here. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

What constraints is the BER minimization problem subject to in the paper?

The BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold.

purpose

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. The methodology for deriving the optimal precoder structure is summarized in this part. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

How is the structure of the optimal precoder for the OTFS-based ISAC systems derived in this paper?

The structure of the optimal precoder is derived by leveraging singular value decomposition (SVD), based on the Karush-Kuhn-Tucker (KKT) conditions.

summary

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. The method used to obtain the minimum BER precoder is summarized here. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

What algorithm is used to solve the dual problem for obtaining the minimum BER precoder?

The minimum BER precoder is obtained by solving the dual problem using the gradient descent algorithm.

summary

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. The simulation results are summarized in terms of the proposed precoder's performance here. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

How do the simulation results characterize the performance of the proposed precoder compared to benchmarks?

The proposed precoder can achieve a considerable BER performance compared with the benchmarks, according to simulation results.

summary

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. The notation for some mathematical operations and matrices is defined in this sentence. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). circ(·), diag(·) represent a cyclic shift matrix and diagonal matrix, respectively.

In the context of this paper, what do the symbols ⊗, 𝐅_M, and 𝐈_M represent?

The Kronecker product is denoted by ⊗, the M-dimensional discrete Fourier transform (DFT) matrix by 𝐅_M , and the identity matrix by 𝐈_M.

terminology

arxiv

2312.11921v1

INTRODUCTION Integrated sensing and communication (ISAC) has attracted significant attention due to the fact that both sensing and communication functionalities are highly expected in the emerging next-generation wireless networks . The so-called ISAC technology refers to the joint design of sensing and communications in a unified device exploiting the same resources and eventually pursuing mutual benefits. By doing so, the system can achieve higher utilization efficiency while saving the cost of spectral resources and power. Recently, various interesting schemes have been studied for realizing practical ISAC systems including the joint waveform design and resource allocation . Meanwhile, motivated by decades of research on orthogonal frequency division multiplexing (OFDM)-based communications, it is a natural progression to design ISAC systems by exploiting OFDM waveforms. Although OFDM shows its superiority in robust synchronization and resilience against multi-path fading, it suffers from a high peak-to-average-power ratio (PAPR) and has seriously degraded performance due to the severe Doppler spread in high-mobility scenarios. To circumvent this challenge, there is a trend to seek a new modulation scheme, e.g., orthogonal time frequency space (OTFS) , which modulates data symbols in the delay-Doppler (DD) domain, different from OFDM that modulates in the time-frequency (TF) domain. In particular, by exploiting the quasi-static and sparse channel in the DD domain, OTFS is capable of supporting multi-path and high-mobility communications. Furthermore, OTFS modulation directly links the transmitted signals and the DD domain channel characteristics, e.g., delay and Doppler shifts, enabling OTFS as a promising candidate for realizing ISAC . As such, various works, e.g., have studied radar-assisted communication systems adopting OTFS modulation for both uplink and downlink transmissions. However, these contributions mainly focus on signal processing in the TF domain, which does not fully exploit the the characteristics of DD domain. Meanwhile, The OTFS-based ISAC system design oriented DD domain has not been reported, yet. Motivated by the above discussions, in this paper, we devise the optimal precoder for OTFS-based ISAC systems in which the bit error rate (BER) and CRB are employed to evaluate the communication performance and sensing accuracy, respectively. We first derive the unified BER expressions for both the cases adopting zero-forcing (ZF) and minimum mean square error (MMSE) equalizations as well as the CRB with respect to (w.r.t) Doppler estimation. Then, the BER minimization problem is formulated subject to the maximum transmit power constraint and required CRB threshold. To address the problem at hand, we rely on the minimization of the average BER lower bound at moderate-to-high SNRs. By leveraging the singular value decomposition (SVD), we derive the structure of optimal precoder based on the Karush-Kuhn-Tucker (KKT) conditions. As such, the minimum BER precoder is obtained by solving the dual problem adopting the gradient descent algorithm. Our simulation results show that our proposed precoder can achieve a considerable BER performance compared with the benchmarks. Notations: The boldface lowercase letter and boldface capital letter denote the vector and the matrix, respectively. The superscript (·)^T and (·)^H denote the transposition, Hermitian operations, respectively. We adopt |·|, 𝔼(·), and tr(·) to denote the modulus of a complex variable, the expectation operation, and the trace operation, respectively; ⊗, 𝐅_M, and 𝐈_M represent the Kronecker product, the M-dimensional discrete Fourier transform (DFT) matrix and identity matrix, respectively. erfc (·) is the complementary error function and we define [x]^+≜max(0,x). ℂ^M× N denotes the set of all M × N matrices with complex entries and 𝐀_m,m is the (m,m)-th entry of 𝐀. The Gaussian distribution with mean μ and covariance matrix Σ is denoted by 𝒞𝒩 (μ, Σ). The definitions for certain matrix operations are provided here.

What do the terms circ(·) and diag(·) represent in the paper?

circ(·) represents a cyclic shift matrix and diag(·) represents a diagonal matrix.

terminology

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. Here, the summary of the number of osculating lines to a hypersurface is provided. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

What is the quantity of osculating lines to a hypersurface in the projective space as stated in the summary of the passage?

The number of osculating lines to a hypersurface Y in P^(n+1)_ℂ is n!.

summary

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . The use of 𝔸^1-homotopy in the generalization process is discussed here. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

What is the purpose of using 𝔸^1-homotopy as mentioned in the passage?

In the paper, the authors generalize the classic result to any perfect field k using 𝔸^1-homotopy.

purpose

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. The representation of the solution to an enumerative problem in 𝔸^1-homotopy theory is mentioned here. §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

In 𝔸^1-homotopy theory, what represents the solution to an enumerative problem?

The solution of an enumerative problem in 𝔸^1-homotopy theory is an element of the Grothendieck–Witt ring of k.

terminology

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). The definition of J(Y,p) is outlined in this sentence. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

What does J(Y,p) denote in the context of this paper?

J(Y,p) is the sum over points l in the subscheme F of ⟨ J(l) ⟩ times the trace from l to k(p).

terminology

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. The main result of the sums involving traces and signed volumes is summarized here. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

What is the main result summarized when a general hypersurface Y meets a k-rational line transversely as described in the passage?

For a general hypersurface Y of degree d≥ n in P^(n+1)_k and a general k-rational line ℒ meeting Y transversely, the sum of certain traces multiplied by signed volumes is dn!/2ℍ.

summary

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). An example of the contribution of an osculating line to the sum is given here. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

In the case of k=ℂ, what contribution does each osculating line provide to the computed sum?

Each osculating line contributes with ⟨ 1 ⟩ to the sum.

example

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. The example of the number of osculating conics to a hypersurface is demonstrated here. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

How many osculating conics are there to a hypersurface at a point in the complex projective space as given by the example?

The number of osculating conics to a hypersurface Y at a point p in P^(n+1)_ℂ is (2n+2)!/2^n+2 - ((n+1)!)^2/2.

example

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . The specific number of osculating conics for n=2 is stated in this part. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

What is the specific number of osculating conics when n equals 2, as provided by the example?

If n=2, there are 27 osculating conics.

example

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . Future research plans related to the count of osculating curves are revealed here. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. This work was supported by FCT - Fundação para a Ciência e a Tecnologia, under the project: UIDP/04561/2020. The author is a member of GNSAGA (INdAM).

What does the author plan to address in future research regarding osculating curves?

The problem of the enriched count of osculating curves of higher degree will be addressed in the future.

future

arxiv

2312.12129v1

INTRODUCTION A classical Salmon's result states that, over k=ℂ, for a general smooth surface Y⊂ℙ^3_ℂ and a general point p∈ Y there are exactly two lines meeting Y at p with contact order 3 . That result does not depend on the degree of Y. A possible generalization of Salmon's result consists on the number of rational curves C⊂ℙ^n+1_ℂ meeting a hypersurface with contact order (n+2) (C)-1 at a fixed general point. These curves have recently been used in a number of interesting applications , and their number has been computed recursively using Gromov–Witten invariants in . They are called osculating curves. In particular, the number of osculating lines to a hypersurface Y⊂ℙ^n+1_ℂ is n!. This result was probably first proved in . In this paper, we generalize Equation () to any perfect field k using 𝔸^1-homotopy. In this theory, the solution of an enumerative problem is an element of the Grothendieck–Witt ring (k) of k. It is the completion of the semi-ring of isomorphism classes of non-degenerate, symmetric bilinear forms on finite dimensional vector spaces over k. This technique has been used, for example, to compute the number of lines in the cubic surface , in the quintic threefold , and the degree of the Grassmannian of lines . §.§ Statement of the main result Let a∈ k∖{0}. We denote by ⟨ a ⟩∈(k) the class of the bilinear form k× k→ k given by (x,y)↦ axy. Given a finite extension K/k, the group morphism Tr_K/k(K)→(k) is the natural map induced by the field trace K→ k. We denote by ℍ:=⟨ 1⟩+⟨ -1⟩. Let Y⊂ℙ^n+1_k be a hypersurface of degree d≥ n given by a homogeneous polynomial f, and p∈ Y be a point. If p is not k-rational, we extend the scalars and consider f_k(p) with a lift p̃ of p. If U is an affine open set centered at p̃, the restriction of f_k(p) to U has a unique Taylor series, that is it decomposes as sum of homogeneous polynomials f^(i) of degree i: f_k(p)|U=f^(1)+f^(2)+⋯+f^(d). If the subscheme F:= f^(1)∩⋯∩ f^(n) of ℙ^n_k(p) is finite, for each point l∈ F we denote by J(l) the signed volume of the parallelepiped determined by the gradient vectors of f^(1),f^(2),⋯,f^(n) at l, and J(Y,p):=∑_l∈ f^(1)∩⋯∩ f^(n)Tr_k(l)/k(p)⟨ J(l)⟩. Our main result is the following. Let n and d be positive integers such that the following conditions are satisfied * n≡ 2 mod 4, * d≡ 0 mod 2. Let Y⊂ℙ^n+1_k be a general hypersurface of degree d≥ n, and ℒ be a general k-rational line meeting Y transversely. Then ∑_p∈ Y∩ℒTr_k(p)/k⟨ J(Y,p) ⟩=dn!/2ℍ. When k=ℂ, the map rank(ℂ)→ℤ is an isomorphism, so the theorem implies Equation (). Indeed, each osculating line contributes with ⟨ 1⟩ to the sum. By , the number of complex osculating lines is the same for each of the general d points {p_1,…,p_d}=Y∩ℒ. Hence there are n! lines through each of these points. Our result is related to Bézout–McKean Theorem, see Remark . §.§ Possible generalizations One may consider conics instead of lines. For example, the number of osculating conics to a hypersurface Y⊂ℙ^n+1_ℂ at p is (2n+2)!/2^n+2-((n+1)!)^2/2. This is proved in . See also the tree-formula of . Hence, if n=2, there are 27 osculating conics. Darboux noted that, when (Y)=3, each line of Y is coplanar to a unique osculating conic, and vice versa. This explains why they are 27. Recent examples of enriched counts of conics are . We hope to address the problem of the enriched count of osculating curves of higher degree in the near future. We do not expect that the solution will be a multiple of the hyperbolic class ℍ. §.§ Outline We introduce 𝔸^1-homotopy in Section . Sections and contain many preliminary results. They follow closely the style of . Section defines the relavant vector bundle, and the bundle of principal parts. The last section proves the main results. The author thanks Ethan Cotterill, Gabriele Degano, Stephen McKean, Kyler Siegel, and Israel Vainsencher for many useful discussions. He also thank Michael Stillman and Matthias Zach for their computational help. Details of the organization that supported the research are given in this sentence. The author is a member of GNSAGA (INdAM).

Which organization provided support for the research work mentioned in the passage?

This work was supported by FCT - Fundação para a Ciência e a Tecnologia under the project: UIDP/04561/2020.

terminology

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The reason for developing neural operators is outlined here. This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. In this paper, we propose an approach that employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. We have referred to this task as “super-fidelity”, drawing inspiration from the super-resolution concept in image processing, which enhances a low-resolution image to achieve higher resolution . The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. In this work, we utilize our recently developed neural operator, the vector-cloud neural network with equivariance (VCNN-e) , to establish the desired mapping in super-fidelity. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. Our numerical experiments on two distinct types of steady-state simulations, one with weak nonlinearity and the other with strong nonlinearity, illustrate the effectiveness of the super-fidelity approach. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. This is particularly beneficial to problems involving multiple simulations under varying conditions, a common need in addressing optimal design or control problems in diverse engineering applications . For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. In Section , we present the neural operator predictions and investigate the resulting acceleration effects for numerical solvers in two testing cases. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. We also discuss the difficulties and limitations of developing reliable neural network surrogates for PDEs. Section provides a summary of this work and some concluding remarks.

What purpose does the development of neural operators serve in the context of neural networks for PDE solutions?

The latest development of neural operators has further extended the capacity of neural networks to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations.

purpose

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The latest development of neural operators has further extended this capability to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations; see, e.g., . This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. The purpose of using neural network predictions in the proposed approach is discussed in this section. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. We have referred to this task as “super-fidelity”, drawing inspiration from the super-resolution concept in image processing, which enhances a low-resolution image to achieve higher resolution . The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. In this work, we utilize our recently developed neural operator, the vector-cloud neural network with equivariance (VCNN-e) , to establish the desired mapping in super-fidelity. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. Our numerical experiments on two distinct types of steady-state simulations, one with weak nonlinearity and the other with strong nonlinearity, illustrate the effectiveness of the super-fidelity approach. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. This is particularly beneficial to problems involving multiple simulations under varying conditions, a common need in addressing optimal design or control problems in diverse engineering applications . For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. In Section , we present the neural operator predictions and investigate the resulting acceleration effects for numerical solvers in two testing cases. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. We also discuss the difficulties and limitations of developing reliable neural network surrogates for PDEs. Section provides a summary of this work and some concluding remarks.

For what purpose does the proposed approach in the paper use neural network predictions?

The proposed approach employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs.

purpose

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The latest development of neural operators has further extended this capability to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations; see, e.g., . This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. In this paper, we propose an approach that employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. The term 'super-fidelity' is defined here. The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. In this work, we utilize our recently developed neural operator, the vector-cloud neural network with equivariance (VCNN-e) , to establish the desired mapping in super-fidelity. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. Our numerical experiments on two distinct types of steady-state simulations, one with weak nonlinearity and the other with strong nonlinearity, illustrate the effectiveness of the super-fidelity approach. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. This is particularly beneficial to problems involving multiple simulations under varying conditions, a common need in addressing optimal design or control problems in diverse engineering applications . For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. In Section , we present the neural operator predictions and investigate the resulting acceleration effects for numerical solvers in two testing cases. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. We also discuss the difficulties and limitations of developing reliable neural network surrogates for PDEs. Section provides a summary of this work and some concluding remarks.

What does the term 'super-fidelity' refer to in the context of the study?

“Super-fidelity” refers to the task of bridging the gap between low and high-fidelity models to enhance efficiency and stability of steady-state simulations in complex flow investigations.

terminology

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The latest development of neural operators has further extended this capability to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations; see, e.g., . This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. In this paper, we propose an approach that employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. We have referred to this task as “super-fidelity”, drawing inspiration from the super-resolution concept in image processing, which enhances a low-resolution image to achieve higher resolution . The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. An example of a neural operator used in this study is mentioned here. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. Our numerical experiments on two distinct types of steady-state simulations, one with weak nonlinearity and the other with strong nonlinearity, illustrate the effectiveness of the super-fidelity approach. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. This is particularly beneficial to problems involving multiple simulations under varying conditions, a common need in addressing optimal design or control problems in diverse engineering applications . For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. In Section , we present the neural operator predictions and investigate the resulting acceleration effects for numerical solvers in two testing cases. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. We also discuss the difficulties and limitations of developing reliable neural network surrogates for PDEs. Section provides a summary of this work and some concluding remarks.

Can you give an example of a neural operator used in the study to establish the desired super-fidelity mapping?

The vector-cloud neural network with equivariance (VCNN-e) is utilized to establish the desired mapping in super-fidelity.

example

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The latest development of neural operators has further extended this capability to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations; see, e.g., . This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. In this paper, we propose an approach that employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. We have referred to this task as “super-fidelity”, drawing inspiration from the super-resolution concept in image processing, which enhances a low-resolution image to achieve higher resolution . The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. In this work, we utilize our recently developed neural operator, the vector-cloud neural network with equivariance (VCNN-e) , to establish the desired mapping in super-fidelity. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. The main finding from the numerical experiments is summarized here. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. This is particularly beneficial to problems involving multiple simulations under varying conditions, a common need in addressing optimal design or control problems in diverse engineering applications . For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. In Section , we present the neural operator predictions and investigate the resulting acceleration effects for numerical solvers in two testing cases. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. We also discuss the difficulties and limitations of developing reliable neural network surrogates for PDEs. Section provides a summary of this work and some concluding remarks.

What is the main finding from the numerical experiments on the steady-state simulations conducted in this study?

Numerical experiments illustrate that in both cases of steady-state simulations with weak and strong nonlinearity, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy.

summary

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The latest development of neural operators has further extended this capability to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations; see, e.g., . This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. In this paper, we propose an approach that employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. We have referred to this task as “super-fidelity”, drawing inspiration from the super-resolution concept in image processing, which enhances a low-resolution image to achieve higher resolution . The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. In this work, we utilize our recently developed neural operator, the vector-cloud neural network with equivariance (VCNN-e) , to establish the desired mapping in super-fidelity. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. Our numerical experiments on two distinct types of steady-state simulations, one with weak nonlinearity and the other with strong nonlinearity, illustrate the effectiveness of the super-fidelity approach. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. The advantages of using trained neural networks for such engineering problems are explained here. For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. In Section , we present the neural operator predictions and investigate the resulting acceleration effects for numerical solvers in two testing cases. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. We also discuss the difficulties and limitations of developing reliable neural network surrogates for PDEs. Section provides a summary of this work and some concluding remarks.

Why could trained neural networks be specifically advantageous for problems involving multiple simulations under varying conditions in engineering applications?

Trained neural networks can deliver better initial conditions, thereby speeding up the solving process, which is particularly beneficial to problems involving multiple simulations under varying conditions common in optimal design or control problems in diverse engineering applications.

purpose

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The latest development of neural operators has further extended this capability to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations; see, e.g., . This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. In this paper, we propose an approach that employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. We have referred to this task as “super-fidelity”, drawing inspiration from the super-resolution concept in image processing, which enhances a low-resolution image to achieve higher resolution . The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. In this work, we utilize our recently developed neural operator, the vector-cloud neural network with equivariance (VCNN-e) , to establish the desired mapping in super-fidelity. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. Our numerical experiments on two distinct types of steady-state simulations, one with weak nonlinearity and the other with strong nonlinearity, illustrate the effectiveness of the super-fidelity approach. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. This is particularly beneficial to problems involving multiple simulations under varying conditions, a common need in addressing optimal design or control problems in diverse engineering applications . For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. The content of Section is summarized here. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. We also discuss the difficulties and limitations of developing reliable neural network surrogates for PDEs. Section provides a summary of this work and some concluding remarks.

What is covered in Section as mentioned in the passage?

Neural operator predictions and the resulting acceleration effects for numerical solvers in two testing cases are presented.

summary

arxiv

2312.11842v1

INTRODUCTION Partial differential equations (PDEs) serve as a fundamental and ubiquitous tool to describe how various properties change continuously in space and time. Their significance extends to numerous fields, including physics, engineering, finance, and beyond. While the development of numerical methods has enabled efficient solutions for certain types of PDEs, there remains a subset of PDEs for which traditional techniques are computationally intensive, thus limiting their applicability in tackling complex problems. Recent years have witnessed a promising new paradigm in scientific computing, as neural networks have emerged as promising tools for PDE solutions. Neural networks, with the capacity to approximate complex functions, can deliver near-ground-truth predictions quickly once trained well. The latest development of neural operators has further extended this capability to infinite-dimensional function spaces, enabling trained models to adapt to inputs with arbitrary resolutions/discretizations; see, e.g., . This paradigm shift in computational science has brought us more insights and new possibilities for solving PDEs. However, relying on neural network predictions as the ultimate solution to PDEs often triggers a trust crisis, especially for tasks that prioritize computational precision and outcome determinism. This trust issue exists in both two prevalent neural network-based approaches for PDE solving, including data-driven methods such as supervised learning of solution operators, and physics-driven methods such as physics-informed neural networks (PINNs). In the case of data-driven methods, training an accurate and reliable network model is highly challenging, which significantly depends on data volume and quality, as well as training algorithms. While studies have shown that well-trained models can produce results with desired accuracy in certain cases , the inherent black-box nature of neural networks and the absence of physical constraints (e.g., conservation laws) during training render these predictions impractical for researchers who rely on first principles in their work. On the other hand, PINNs aim to solve PDEs by training neural networks to satisfy physical constraints . However, comparative studies between PINNs and traditional solvers indicate that PINNs do not exhibit a clear advantage either in terms of accuracy or computational speed for various PDEs in low-dimensional cases . Hence, for scientific computing tasks that require solution precision, traditional solvers should continue to occupy a central role. It is essential to mitigate the risk of misusing neural network predictions and to ensure the accuracy of the outcomes, at least at the current stage. In this paper, we propose an approach that employs neural network predictions as initial conditions to accelerate the solution of steady-state PDEs. We have chosen steady-state fluid flow problems to demonstrate the efficacy of our approach, considering their industrial relevance and the inherent complexity associated with computational models based on the Navier–Stokes (N–S) equations. Specifically, our proposed solution strategy involves the development of an approximate mapping from the solution of a low-fidelity model (i.e., Laplace's equation for potential flow) to the solution of a high-fidelity model (i.e., the target PDEs). This mapping, which is straightforward and fast to evaluate, provides a solution that is much closer to the final solution than the low-fidelity model. By utilizing the output of this mapping as an improved initial guess, we can effectively warm-start the iterative solving process, further reducing the total computational time. We have referred to this task as “super-fidelity”, drawing inspiration from the super-resolution concept in image processing, which enhances a low-resolution image to achieve higher resolution . The objective of super-fidelity is to bridge the gap between low and high-fidelity models, thereby enhancing the efficiency and stability of steady-state simulations in complex flow investigations. In the context of warm-starting existing iterative solvers for high-fidelity models, our focus lies in constructing a mapping from low-fidelity solutions to high-fidelity ones – the core component of our super-fidelity approach. This mapping, mathematically, is also referred to as an operator between two function spaces. The field of operator learning, which has received significant research interest in recent years, offers such a way forward. This approach aims to develop a data-driven approximation of mappings/operators between function spaces to provide efficient evaluations of the target output after an initial training phase on available data. In particular, neural networks with specialized architectures, also known as neural operators, have shown considerable promise in approximating such mappings. In this work, we utilize our recently developed neural operator, the vector-cloud neural network with equivariance (VCNN-e) , to establish the desired mapping in super-fidelity. The VCNN-e guarantees the physical symmetry in flow simulations, allowing the operator to be applied in different coordinate systems and discretizations seamlessly. Furthermore, it considers the nonlocal property of flow simulations, enabling learning to be done with just a few solutions by decomposing the whole solution field into many patches for efficient learning. The entire workflow of our super-fidelity approach includes all four panels shown in Fig. . The second panel, referred to as the warm-start, involves the training of a super-fidelity model based on neural operators. Our numerical experiments on two distinct types of steady-state simulations, one with weak nonlinearity and the other with strong nonlinearity, illustrate the effectiveness of the super-fidelity approach. In both cases, the warm-start models achieve acceleration ratios of no less than two folds without any compromise on accuracy. For practitioners in computational fluid dynamics (CFD) who typically employ uniform fields or solutions from lower-fidelity models as initial conditions (see “traditional workflow” in Fig. , including the (a), (c), and (d) panels), trained neural networks can deliver better initial conditions, thereby speeding up the solving process. This is particularly beneficial to problems involving multiple simulations under varying conditions, a common need in addressing optimal design or control problems in diverse engineering applications . For researchers devoted to accelerating PDE solving through machine learning (see “surrogate model" in Fig. , including the (a), (b), and (d) panels), integrating traditional solvers as the final touch ensures the accuracy of results. The remainder of this paper is organized as follows. Section introduces the computational models adopted in this study, including the governing equations for both high-fidelity and low-fidelity models, and describes the neural operator architecture. In Section , we present the neural operator predictions and investigate the resulting acceleration effects for numerical solvers in two testing cases. In Section , we introduce an efficient strategy for data generation and explore potential application scenarios to achieve a “real" acceleration by reducing overall computation time, which includes both data generation and training phases. The potential future application scenarios are discussed in Section . Section provides a summary of this work and some concluding remarks.

What is the future aspect of research discussed in Section of the paper?

Potential application scenarios are explored to achieve a “real' acceleration by reducing overall computation time, which includes both data generation and training phases.

future

arxiv

2312.10893v1

The definition of Generative artificial intelligence (GenAI) systems is provided here. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

What are Generative artificial intelligence (GenAI) systems?

Generative artificial intelligence (GenAI) systems are models like Large Language Models (LLMs) that can generate artefacts by using extensive parameters and training data to model and sample from a feature space.

terminology

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . The concept of metacognition is elaborated on in this sentence. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

How is metacognition defined within the context of psychology?

Metacognition refers to the psychological ability to monitor, understand, and control one’s own thought processes.

terminology

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. A summary of GenAI systems' distinctive features compared to search engines is summarized here. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

In what way do GenAI systems differ from search engines?

GenAI systems present features that are distinct from that of search engines.

summary

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Examples of GenAI systems' flexibility are listed here. Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

Can you give examples of how GenAI systems are flexible compared to search engines?

GenAI systems are flexible in their responsiveness to user prompts, in the range of available parameters to users in their prompts, and in the multi-modality of their input/output space.

example

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Functions that GenAI systems can perform are detailed here. Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

What are some of the general-purpose functions that GenAI systems can perform?

GenAI systems can perform content generation, discrimination, and editing, among other functions.

example

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). The capability of GenAI systems to generate content is mentioned here. Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

What is an example of the content-related capability of GenAI systems?

GenAI systems are capable of generating original content.

example

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Examples of metacognitive support strategies are given in this sentence. Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

Could you provide examples of metacognitive support strategies?

Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management.

example

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. A summary of the main contributions of the work is outlined here. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Further, we provide research and design opportunities that leverage the unique properties of GenAI to augment GenAI system usability. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

What are the main contributions of the presented work?

The work conceptualizes and grounds the usability challenges of GenAI in an understanding of human metacognition, draws on research from psychological and cognitive science and recent GenAI user studies, proposes two directions for addressing the metacognitive demands of GenAI, and provides research and design opportunities that leverage the unique properties of GenAI to augment system usability.

summary

arxiv

2312.10893v1

INTRODUCTION Generative artificial intelligence (GenAI) systems—models, like Large Language Models (LLMs), that can generate artefacts by using extensive parameters and training data to model and sample from a feature space —have the potential to transform personal and professional work. Their potential stems from a unique combination of model flexibility (in their input/output space), generality (in their applicability to a wide range of tasks), and originality (in their ability to generate novel content) . However, these same properties also pose a challenge for designing GenAI systems to be human-centered . Recent user studies reveal a range of usability challenges around prompting , evaluating and relying on outputs , and deciding on an automation strategy: whether and how to integrate GenAI into workflows . Recent work has sought to characterize the unique properties of GenAI and their potential effects on users . Other work offers technical or design roadmaps for designing human-centered GenAI . However, there is not yet a coherent understanding of the usability challenges of GenAI, much less one grounded in a theory of human cognition. Indeed, recent work has called for foundational research to understand how people interact with GenAI and AI more broadly . Here, we argue that metacognition—the psychological ability to monitor, understand, and control one’s own thought processes —offers a valuable and unexplored perspective to understand and design for the usability challenges of GenAI. Firstly, we suggest that current GenAI systems impose multiple metacognitive demands on users; understanding these demands can help interpret and further probe the identified usability challenges, and identify additional challenges for future research. Secondly, we suggest that the perspective of metacognitive demands offers new research and design opportunities for human-GenAI interaction. In many ways, the metacognitive demands of working with GenAI systems parallel those of a manager delegating tasks to a team. The manager needs to clearly understand and formulate their goals, break down those goals into communicable tasks, confidently assess the quality of the team's output, and adjust plans accordingly along the way. Among others, these abilities comprise metacognitive monitoring and control of one's thought processes and behavior . Analogously, current GenAI systems often require verbalized prompting, demanding self-awareness of task goals, and decomposition of tasks into sub-tasks. System outputs then need to be evaluated, requiring well-adjusted confidence in one’s evaluation and prompting abilities, and metacognitive flexibility to iterate on the prompting strategy as necessary. We posit that these metacognitive demands are induced by GenAI's model flexibility, generality, and originality. [The perspective of metacognition is equally relevant for understanding the usability of search engines and similar technologies. However, we focus our scope to GenAI systems as they present features that are relevantly distinct from that of search engines . Firstly, they are flexible in their responsiveness to user prompts (moreso than search engines), in the wide range of implicit and explicit parameters available to users in their prompts (moreso than search engines), and relatedly, in the multi-modality of their input/output space (unlike search engines). Secondly, they function as general-purpose tools, able to perform content generation, discrimination, and editing, among other functions (unlike search engines). Thirdly, they are able to generate original content (rather than merely retrieve existing content). Finally, unlike search engines, current GenAI systems are non-deterministic in their responses. As we aim to demonstrate here, all of these features place unique demands on users' metacognition and inform the design space of solutions to address these demands, a design space which necessarily extends beyond that of current search engines. Relatedly, we also note that Dan Russell proposed a connected idea of `meta-literacy’ for search engine usability (see also ); however, Russell does not delve into the psychological and cognitive science of metacognition that is central to the current work.] Alongside the local interactions with GenAI systems, the generality of GenAI poses another, higher-level metacognitive demand: the challenge of knowing whether and how to incorporate GenAI into workflows (i.e., one’s `automation strategy'; see also ). This demands self-awareness of GenAI’s applicability to, and impact on, one’s workflow; well-adjusted confidence in manual versus GenAI-supported task completion; and metacognitive flexibility to adapt one’s workflows as needed. In Section , we draw on metacognition research and recent user studies of GenAI to illustrate these metacognitive demands and offer new research directions to probe them further. The metacognitive demands posed by GenAI can be addressed in at least two complementary ways. Firstly, given that metacognitive abilities can be taught , we can improve users’ metacognition via metacognitive support strategies that can be integrated into GenAI systems. Evidence-based metacognitive support strategies include those that help users in their planning, self-evaluation, and self-management . Recent HCI work has begun to pursue this direction , albeit without explicitly grounding it in metacognition; we suggest that a metacognitive lens offers new research and design directions for augmenting GenAI system usability. Secondly, we can reduce the metacognitive demand of generative AI systems by designing task-appropriate approaches to GenAI explainability and customizability. We suggest that explainability can help offload metacognitive processing from the user to the system, and that prior explainability approaches can be augmented by considering metacognition. Likewise, we suggest that a metacognitive perspective can provide insights on approaching the end-user customizability of GenAI systems. In Section , we draw on intervention studies to improve metacognition and recent studies of GenAI prototypes and human-AI interaction to explore research and design directions that can address the metacognitive demands of GenAI. Critically, we also highlight how GenAI’s model flexibility, generality, and originality can serve as a design solution to these demands. Finally, we briefly discuss the relationship between cognitive load and addressing metacognitive demands, offering ways to manage their balance. In summary, our work makes three distinct contributions: * We conceptualize and ground the usability challenges of GenAI in an understanding of human metacognition, drawing on research from psychological and cognitive science and recent GenAI user studies. * We draw from research on metacognitive interventions, GenAI prototypes, and human-AI interaction to propose two directions for addressing the metacognitive demands of GenAI: improving users’ metacognition, and reducing the metacognitive demands of GenAI. * Beyond re-framing the usability challenges of GenAI, we use the metacognition lens to identify the need—and concrete directions—for further research into the metacognitive demands of GenAI. Future research directions suggested by the presented work are discussed here. In the next sections we define the concept of metacognition, summarizing key research findings (Section ); illustrate the metacognitive demands of GenAI, focusing on prompting, evaluating and relying on outputs, and deciding on one’s automation strategy (Section ); and propose ways to address these metacognitive demands (Section ).

What future research directions are proposed in this work related to GenAI and metacognition?

The work identifies the need for further research into the metacognitive demands of GenAI and the potential to leverage GenAI to augment system usability.

future

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure is described here. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. It is of great significance that the map Φ_0 has a genuinely geometric origin—namely, the Procesi bundle—which exhibits Macdonald positivity. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. While this point of view is not strong enough to establish Macdonald positivity, it does provide a way to uniquely characterize the map Φ_0, up to a scalar. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. Here _n,n-k is the parabolic flag Hilbert scheme, also introduced in , which parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. We propose an explicit answer for all k involving a new family of modified partially-symmetric Macdonald functions _(|). [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The partially-symmetric integral form Macdonald functions _(|)^w_0(X y_1,…,y_k)∈Λ⊗[y_1,…,y_k] are defined in () by partially-symmetrizing the (stable) nonsymmetric Macdonald polynomials, multiplying by explicit scalars introduced in and , and then performing a twist by the long element of S_k. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. In this paper, we make substantial progress toward a proof of Conjecture by showing that it is compatible with the action of certain -generators; see Theorem . We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. In light of our conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials, it natural to hope for a geometric realization of Φ exhibiting positivity.

What is the key structure in Haiman's geometric realization of Macdonald symmetric functions?

A key structure in Haiman's geometric realization of Macdonald symmetric functions is a vector space isomorphism between Λ and the localized equivariant K-groups of the Hilbert schemes of points in the plane.

summary

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure in Haiman's geometric realization of Macdonald symmetric functions (see also ) is a vector space isomorphism ⊕_n≥ 0 K_(_n)_locΦ_0⟶Λ between Λ and the localized equivariant K-groups of the Hilbert schemes _n of points in the plane ^2, where is the natural two-dimensional torus acting on the plane. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. The origin and significant property of the map Φ_0 are mentioned here. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. While this point of view is not strong enough to establish Macdonald positivity, it does provide a way to uniquely characterize the map Φ_0, up to a scalar. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. Here _n,n-k is the parabolic flag Hilbert scheme, also introduced in , which parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. We propose an explicit answer for all k involving a new family of modified partially-symmetric Macdonald functions _(|). [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The partially-symmetric integral form Macdonald functions _(|)^w_0(X y_1,…,y_k)∈Λ⊗[y_1,…,y_k] are defined in () by partially-symmetrizing the (stable) nonsymmetric Macdonald polynomials, multiplying by explicit scalars introduced in and , and then performing a twist by the long element of S_k. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. In this paper, we make substantial progress toward a proof of Conjecture by showing that it is compatible with the action of certain -generators; see Theorem . We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. In light of our conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials, it natural to hope for a geometric realization of Φ exhibiting positivity.

What is the geometric origin of the map Φ_0 and what significant property does it exhibit?

The map Φ_0 has a geometric origin in the Procesi bundle, which exhibits Macdonald positivity.

purpose

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure in Haiman's geometric realization of Macdonald symmetric functions (see also ) is a vector space isomorphism ⊕_n≥ 0 K_(_n)_locΦ_0⟶Λ between Λ and the localized equivariant K-groups of the Hilbert schemes _n of points in the plane ^2, where is the natural two-dimensional torus acting on the plane. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. It is of great significance that the map Φ_0 has a genuinely geometric origin—namely, the Procesi bundle—which exhibits Macdonald positivity. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. Schiffmann and Vasserot's perspective on the map Φ_0 is summarized here. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. Here _n,n-k is the parabolic flag Hilbert scheme, also introduced in , which parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. We propose an explicit answer for all k involving a new family of modified partially-symmetric Macdonald functions _(|). [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The partially-symmetric integral form Macdonald functions _(|)^w_0(X y_1,…,y_k)∈Λ⊗[y_1,…,y_k] are defined in () by partially-symmetrizing the (stable) nonsymmetric Macdonald polynomials, multiplying by explicit scalars introduced in and , and then performing a twist by the long element of S_k. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. In this paper, we make substantial progress toward a proof of Conjecture by showing that it is compatible with the action of certain -generators; see Theorem . We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. In light of our conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials, it natural to hope for a geometric realization of Φ exhibiting positivity.

How do Schiffmann and Vasserot view the map Φ_0?

The point of view of Schiffmann and Vasserot regarding the map Φ_0 is that it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ.

summary

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure in Haiman's geometric realization of Macdonald symmetric functions (see also ) is a vector space isomorphism ⊕_n≥ 0 K_(_n)_locΦ_0⟶Λ between Λ and the localized equivariant K-groups of the Hilbert schemes _n of points in the plane ^2, where is the natural two-dimensional torus acting on the plane. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. It is of great significance that the map Φ_0 has a genuinely geometric origin—namely, the Procesi bundle—which exhibits Macdonald positivity. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. While this point of view is not strong enough to establish Macdonald positivity, it does provide a way to uniquely characterize the map Φ_0, up to a scalar. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. The function of the parabolic flag Hilbert scheme _n,n-k is explained here. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. We propose an explicit answer for all k involving a new family of modified partially-symmetric Macdonald functions _(|). [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The partially-symmetric integral form Macdonald functions _(|)^w_0(X y_1,…,y_k)∈Λ⊗[y_1,…,y_k] are defined in () by partially-symmetrizing the (stable) nonsymmetric Macdonald polynomials, multiplying by explicit scalars introduced in and , and then performing a twist by the long element of S_k. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. In this paper, we make substantial progress toward a proof of Conjecture by showing that it is compatible with the action of certain -generators; see Theorem . We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. In light of our conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials, it natural to hope for a geometric realization of Φ exhibiting positivity.

What does the parabolic flag Hilbert scheme _n,n-k parametrize?

The parabolic flag Hilbert scheme, _n,n-k, parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n.

terminology

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure in Haiman's geometric realization of Macdonald symmetric functions (see also ) is a vector space isomorphism ⊕_n≥ 0 K_(_n)_locΦ_0⟶Λ between Λ and the localized equivariant K-groups of the Hilbert schemes _n of points in the plane ^2, where is the natural two-dimensional torus acting on the plane. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. It is of great significance that the map Φ_0 has a genuinely geometric origin—namely, the Procesi bundle—which exhibits Macdonald positivity. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. While this point of view is not strong enough to establish Macdonald positivity, it does provide a way to uniquely characterize the map Φ_0, up to a scalar. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. Here _n,n-k is the parabolic flag Hilbert scheme, also introduced in , which parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. The conjecture about the images of fixed-point classes is presented here. [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The partially-symmetric integral form Macdonald functions _(|)^w_0(X y_1,…,y_k)∈Λ⊗[y_1,…,y_k] are defined in () by partially-symmetrizing the (stable) nonsymmetric Macdonald polynomials, multiplying by explicit scalars introduced in and , and then performing a twist by the long element of S_k. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. In this paper, we make substantial progress toward a proof of Conjecture by showing that it is compatible with the action of certain -generators; see Theorem . We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. In light of our conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials, it natural to hope for a geometric realization of Φ exhibiting positivity.

What does the conjecture propose about the images of fixed-point classes under the -module isomorphism?

The conjecture proposes that under the -module isomorphism, the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials.

future

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure in Haiman's geometric realization of Macdonald symmetric functions (see also ) is a vector space isomorphism ⊕_n≥ 0 K_(_n)_locΦ_0⟶Λ between Λ and the localized equivariant K-groups of the Hilbert schemes _n of points in the plane ^2, where is the natural two-dimensional torus acting on the plane. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. It is of great significance that the map Φ_0 has a genuinely geometric origin—namely, the Procesi bundle—which exhibits Macdonald positivity. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. While this point of view is not strong enough to establish Macdonald positivity, it does provide a way to uniquely characterize the map Φ_0, up to a scalar. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. Here _n,n-k is the parabolic flag Hilbert scheme, also introduced in , which parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. We propose an explicit answer for all k involving a new family of modified partially-symmetric Macdonald functions _(|). [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The definition of the partially-symmetric integral form Macdonald functions is provided here. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. In this paper, we make substantial progress toward a proof of Conjecture by showing that it is compatible with the action of certain -generators; see Theorem . We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. In light of our conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials, it natural to hope for a geometric realization of Φ exhibiting positivity.

How are the partially-symmetric integral form Macdonald functions defined?

The partially-symmetric integral form Macdonald functions are defined by partially-symmetrizing the stable nonsymmetric Macdonald polynomials, multiplying by explicit scalars, and then performing a twist by the long element of S_k.

terminology

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure in Haiman's geometric realization of Macdonald symmetric functions (see also ) is a vector space isomorphism ⊕_n≥ 0 K_(_n)_locΦ_0⟶Λ between Λ and the localized equivariant K-groups of the Hilbert schemes _n of points in the plane ^2, where is the natural two-dimensional torus acting on the plane. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. It is of great significance that the map Φ_0 has a genuinely geometric origin—namely, the Procesi bundle—which exhibits Macdonald positivity. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. While this point of view is not strong enough to establish Macdonald positivity, it does provide a way to uniquely characterize the map Φ_0, up to a scalar. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. Here _n,n-k is the parabolic flag Hilbert scheme, also introduced in , which parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. We propose an explicit answer for all k involving a new family of modified partially-symmetric Macdonald functions _(|). [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The partially-symmetric integral form Macdonald functions _(|)^w_0(X y_1,…,y_k)∈Λ⊗[y_1,…,y_k] are defined in () by partially-symmetrizing the (stable) nonsymmetric Macdonald polynomials, multiplying by explicit scalars introduced in and , and then performing a twist by the long element of S_k. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. Progress made toward proving the conjecture is detailed here. We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. In light of our conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials, it natural to hope for a geometric realization of Φ exhibiting positivity.

What progress has been made towards proving the conjecture?

Substantial progress toward a proof of the conjecture has been made by showing that it is compatible with the action of certain -generators.

summary

arxiv

2312.11657v1

INTRODUCTION Let Λ be the algebra of symmetric functions in x_1,x_2,… over =(q,t). A key structure in Haiman's geometric realization of Macdonald symmetric functions (see also ) is a vector space isomorphism ⊕_n≥ 0 K_(_n)_locΦ_0⟶Λ between Λ and the localized equivariant K-groups of the Hilbert schemes _n of points in the plane ^2, where is the natural two-dimensional torus acting on the plane. Under this isomorphism, the modified Macdonald symmetric functions _μ are realized as the images Φ_0([I_μ]) of the K-theory classes corresponding to torus-fixed points I_μ in ⊔_n≥ 0_n. It is of great significance that the map Φ_0 has a genuinely geometric origin—namely, the Procesi bundle—which exhibits Macdonald positivity. Work of Schiffmann and Vasserot offers another point of view on the map Φ_0: it is an intertwining map between two irreducible representation of the elliptic Hall algebra ℰ. While this point of view is not strong enough to establish Macdonald positivity, it does provide a way to uniquely characterize the map Φ_0, up to a scalar. By methods akin to those of , Carlsson, Gorsky, and Mellit costruct an extension of the map Φ_0, with Carlsson and Mellit's algebra in place of the elliptic Hall algebra ℰ. The extension Φ maps between the following larger spaces ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k] both of which afford an irreducible action of the algebra with a canonical generator. Here _n,n-k is the parabolic flag Hilbert scheme, also introduced in , which parametrizes flags of ideals I_n⊂ I_n-1⊂…⊂ I_n-k in [x,y] with the property that [x,y]/I_s = s and yI_n-k⊂ I_n. The action of on the direct sum of K-groups of these spaces is constructed in by geometrically-defined operators, while the right-hand side of () affords the polynomial representation of . Of course, the requirement that Φ intertwine these two actions is sufficient to uniquely characterize it, up to a scalar. Our work is motivated by the following question: what are the images of torus fixed-point classes in _n,n-k under the isomorphism Φ? One knows that Φ restricts to () on the k=0 summands (up to a minor twist by a sign and a monomial in q,t), where the torus fixed point-classes are sent to modified Macdonald symmetric functions _μ. We propose an explicit answer for all k involving a new family of modified partially-symmetric Macdonald functions _(|). [see Conjecture ] Under the -module isomorphism ⊕_n≥ 0⊕_0≤ k≤ n K_(_n,n-k)_locΦ⟶⊕_k≥ 0Λ⊗[y_1,…,y_k], the fixed-point classes are sent to modified partially-symmetric Macdonald polynomials, as follows: H_μ,w↦ t^n(sort(|))_(|)^w_0(X/1-t | y)^*=:_(|). For the sake of orientation, let us describe a few pieces of the notation used above, leaving the rest for the main body of the paper. The fixed-points I_μ,w∈_n,n-k are indexed by pairs (μ,w) such that μ is a partition of size n and w is an ordered horizontal strip of size k in μ; the classes H_μ,w are explicit scalar multiples of the [I_μ,w]. The partially-symmetric integral form Macdonald functions _(|)^w_0(X y_1,…,y_k)∈Λ⊗[y_1,…,y_k] are defined in () by partially-symmetrizing the (stable) nonsymmetric Macdonald polynomials, multiplying by explicit scalars introduced in and , and then performing a twist by the long element of S_k. The _(|)^w_0(X y)—and their cousins _(|)—are indexed by pairs consisting of a partition λ and a composition γ∈(_≥ 0)^k such that the total size |λ|+|γ| is n-k. An explicit bijection ϕ between these indexing sets, so that ϕ(μ,w)=, is recalled in Section following . Finally, X↦ X/(1-t) denotes the familiar plethystic transformation on Λ from Macdonald theory, and f(X y)^*=f(X y)|_t↦ t^-1. Thus our conjecture closely resembles—and clearly extends—the construction of modified Macdonald functions _μ from the usual integral form Macdonald functions. In this paper, we make substantial progress toward a proof of Conjecture by showing that it is compatible with the action of certain -generators; see Theorem . We will complete the proof in forthcoming work with M. Bechtloff Weising by means of the limit Cherednik operators . It is important to mention that the map Φ of , while uniquely characterized as an -intertwiner, has no genuinely geometric construction at present. The future hope regarding the geometric realization of Φ is discussed here.

What is the future hope regarding the geometric realization of Φ and its properties?

It is natural to hope for a geometric realization of Φ exhibiting positivity in light of the conjecture and Lapointe's positivity conjectures for partially-symmetric Macdonald polynomials.

future

arxiv

2312.11106v1

INTRODUCTION. The concept of a (mod p) homology decomposition of BG is defined here. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

What is a (mod p) homology decomposition of BG?

A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG.

terminology

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. The main property of the spectral sequence in question is summarized here. It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

What is the main property of the spectral sequence derived from the homology decomposition as proved by Dwyer?

The spectral sequence deriving from such homology decomposition is in fact sharp, meaning that [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1).

summary

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). The consequence of the spectral sequence being sharp is discussed here. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

What consequence follows from the spectral sequence being sharp according to the paper?

It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds.

purpose

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. The contribution of Aschbacher, Kessar, and Oliver to the discussion is highlighted here. Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

How did Aschbacher, Kessar, and Oliver contribute to the topic discussed in the passage?

The parallelism between the cases of finite groups and fusion systems naturally leads to the question formulated by Aschbacher, Kessar and Oliver.

example

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . The result of Theorem A is presented here. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

What does Theorem A prove in the paper?

The paper proves that if certain conditions are satisfied, then Conjecture 11 is satisfied for every Mackey functor over 𝒪 on S.

summary

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The discussion regarding the applicability of Theorem A's conditions is provided here. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

Why might the conditions for applying Theorem A appear too restrictive, and how does the paper address this concern?

The conditions needed to apply Theorem A might seem too restrictive, but for any given fusion system there exist several families of fusion subsystems satisfying these conditions, hence the only real problem is proving another conjecture.

purpose

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). The specific fusion systems for which Conjecture 11 is proven by Theorem B are mentioned here. More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

For which specific fusion systems does Theorem B validate Conjecture 11?

Theorem B proves Conjecture 11 for all Benson-Solomon fusion systems.

example

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. The conditions under which Conjecture 11 is satisfied for a fusion system are provided here. We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

Under what conditions is Conjecture 11 satisfied for a fusion system 𝒪 according to the paper?

Conjecture 11 is satisfied for 𝒪 if it is satisfied for fusion systems 𝒪_1 and 𝒪_2 which generate 𝒪.

summary

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . The purpose of listing common notation in the conclusion section is explained here. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The author would like to thank Nadia Mazza for the guidance she provided during the author's PhD and the seemingly limitless amount of resources she's been able to provide. The author would also like to thank Lancaster University for its financial support.

What is the purpose of the list of common notation in the conclusion section of the paper?

The conclusion section provides a list of common notation used throughout the paper for clarity and reference.

purpose

arxiv

2312.11106v1

INTRODUCTION. Let G be a finite group, let p be a prime number, let C be a small category and let F:C→Top be a functor sending every object in C to the classifying space BH of a subgroup H of G. A (mod p) homology decomposition of BG is a mod-p homology equivalence of the form _C(F)p→̃BG. Due to a result of Bousfield and Kan (see ) we know that any such homology decomposition leads to a first quadrant cohomology spectral sequence [i]_CH^j(F(-),)⇒ H^i+j(G,). In Dwyer proves that an homology decomposition of G can be obtained by taking C to be the p-centric orbit category 𝒪_p^c(G) of G and defining F(H):=BH for every subgroup H≤ G in 𝒪_p^c(G). Building on this Dwyer proves in that the spectral sequence of Equation () deriving from such homology decomposition is in fact sharp (i.e. [i]_𝒪_p^c(G)(H^j(-,))=0 for every i≥1). It follows that for every n≥0 the isomorphism of abelian groups lim_𝒪_p^c(G)(H^n(-,))≅ H^n(G,) holds. On the other hand, work of Broto, Levi and Oliver (see ) and of Chermack (see ) leads to the description existence and uniqueness of classifying space of a fusion system B≃_(B(-)). Moreover, from we know that, analogously to the case of finite groups, the isomorphism lim_(H^n(-,))≅ H^n(B,) holds. This parallelism naturally leads to the following question which was first formulated by Aschbacher, Kessar and Oliver in and later generalized by Daz and Park in . Let S be a finite p-group, let be a fusion system over S and let M=(M_*,M^*) be a Mackey functor over on (see Definition ). Then _(M^*↓_^)=0 for every n≥1. Although still unresolved there have recently been several developments regarding this conjecture (see ) which add to the results in . In this paper we aim to add to these efforts by developing some tools that can be used to approach Conjecture . More precisely we prove the following. [Theorem A] Let S be a finite p-group, let be a fusion system over S, let I be a finite set, let F={_i} _i∈ I be a collection of fusion subsystems of and for each i∈ I let S_i≤ S be the finite p-group such that _i is a fusion system over S_i. If the following 4 conditions are satisfied then Conjecture is satisfied for every Mackey functor over on : * θ_F (see Definition ) is an epimorphism. * For every i∈ I all _i-centric-radical subgroups of S_i are -centric (see Definition ()). * For every with i∈ I Conjecture is satisfied for _i. * The family F satisfies the lifting property (see Definition ). The conditions needed to apply Theorem [thm:A]A might seem too restrictive. However, as we show in Section , for any given fusion system there exist several families of fusion subsystems satisfying Conditions ()-() and, therefore, the only real problem is given by Condition () (see Conjecture ). As an application of Theorem [thm:A]A we complete the work started in by proving [Theorem B] Conjecture is satisfied for all Benson-Solomon fusion systems (see ). More in general we prove that [Theorem C] Let S be a finite p-group, let ,_1 and _2 be fusion systems over S such that: * _1,_2⊆ and is generated by _1 and _2 (i.e. all morphisms of can be written as a composition of finitely many morphisms in _1 and _2). * For every i=1,2 all _i-centric-radical subgroups of S are -centric. * For every i=1,2 Conjecture is satisfied for _i. Then Conjecture is satisfied for . We conclude this section with a list of common notation that we use throughout the paper. . * The letter p denotes a prime number. * Every module is understood to be a right module. All arguments can however be adapted to left modules. * All rings are understood to have a unit. * For any ring we denote by the category of right -modules and by the category of finitely generated right -modules. * Let C be a small category, let A,B and C be objects in C and let f∈_C(A,B). We denote by f^*:_C(B,M)→_C(A,M) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(B,M) to the composition gf∈_C(A,M). * With notation as above we denote by f_*:_C(M,A)→_C(M,B) the set map (or morphism of abelian groups if C is abelian) sending every morphism g∈_C(M,A) to the composition fg∈_C(M,B). The reason for the author's gratitude towards Nadia Mazza is expressed here. The author would also like to thank Lancaster University for its financial support.

For what reason does the author thank Nadia Mazza?

The author thanks Nadia Mazza for the guidance and resources provided during the author's PhD.

purpose

arxiv

2312.11006v1

The introduction of the concept of quantum battery (QB) is discussed here. Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

What is the concept of a quantum battery as introduced by Alicki and Fannes?

Quantum battery (QB) is a notion introduced by Alicki and Fannes to temporarily store energy, suggesting that entangling unitary operations can extract more work than local ones.

terminology

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . A summary of the development state of quantum batteries (QBs) is provided here. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

Can you summarize the current state of the development of quantum batteries?

The development of quantum batteries (QBs) is still in their infancy with many challenges to be solved before practical implementation.

summary

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. The purpose of addressing the self-discharging process of quantum batteries is mentioned here. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

Why is it important to address the self-discharging process in quantum batteries?

It is important to address the self-discharging process of quantum batteries which is characterized by the decay of energy when no consumption hubs are connected.

purpose

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . Examples of protocols for energy preservation in QBs are listed here. Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

Could you provide examples of protocols designed to preserve the energy stored in quantum batteries?

Some protocols to preserve the energy in QBs for longer periods include spectrum engineering, non-Markovian effects, local field fluctuation with a disorder term in the Hamiltonian, and manipulating coherence and geometry of intracell connections.

example

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. Details of the performance of the first realization of a quantum battery are given here. Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

What was the performance outcome of the first realization of a quantum battery using an organic semiconductor?

The first realization of quantum battery using an organic semiconductor coupled to a microcavity showed a power density of 67 kW/kg, surpassing lithium-ion batteries.

example

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . The reasons for the importance of superconducting circuits for the future of QBs are explained here. In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

Why are superconducting circuits particularly important for the future development of quantum batteries?

Superconducting circuits are of particular interest for the future of quantum batteries due to their flexibility, adjustability, scalability, and strong coupling with external fields.

future

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. The reason for exploring many-body quantum batteries on superconducting circuits is addressed here. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

What is the purpose behind exploring many-body quantum batteries on superconducting circuits?

The exploration of many-body quantum batteries with efficient charging, stable energy storage, and slow self-discharging processes on superconducting circuits platform is necessary and meaningful.

purpose

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. The definitions of three types of quantum batteries in superconducting circuits are described here. We investigate the performance of these QBs in the presence of the resonator decay and the qutrit decoherence, then discuss the roles of quantum entanglement and quantum coherence during their charging and self-discharging processes. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

What are the three types of quantum batteries defined in the context of superconducting circuits?

Three types of quantum batteries (QBs) are defined: one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB.

terminology

arxiv

2312.11006v1

INTRODUCTION Motivated by the ongoing miniaturization of electronic equipments and the rapid development of quantum thermodynamics in micro- and mesoscopic quantum systems, Alicki and Fannes first introduced the notion of quantum battery (QB) to temporarily store energy and suggested that entangling unitary operations can extract more work than local ones . Subsequently, it was shown that there exists some protocols for optimal work extraction actually without creating any entanglement, at the cost of generically requiring more operations . Further researches uncovered that entanglement may have a destructive effect on the performance of quantum batteries (QBs) , whereas coherence appears to be a potential resource . In recent years, how to exploit quantum resources like entanglement and coherence to outperform its classical counterpart has been a major focus in the quantum thermodynamics realm, but the couplings among different quantum resources lead to the problem intractable and complicated . Nowadays, the development of QBs is still in their infancy and there are many challenges that need to be adequately solved before such quantum devices can be implemented in practice. Besides the widely concerned efficient charging and stable energy-storage processes , it is also important to highlight the self-discharging process characterized by the decay of energy when no consumption hubs are connected to QBs. How to against the undesired effects and boost the ability of QBs in storing energy during the self-discharging process is a key task . At present, some protocols have been purposed to preserve the energy stored in QBs for long times, e.g., considering spectrum engineering and non-Markovian effects , introducing local field fluctuation described by disorder term in the system Hamiltonian , manipulating coherence of the initial states and geometry of the intracell connections . Experimentally, several progresses have exhibited some advantages of QBs. The first realization of QB, using an organic semiconductor as an ensemble of two-level systems coupled to a microcavity, shows that the power density of Dicke QB is up to 67 kW/kg which far exceeds the power densities of existing lithium-ion batteries . Up to now, QBs can be promisingly implemented on many physical platforms , ranging from organic semiconductor and semiconductor quantum dot coupled to optical microcavities, nuclear spin systems with a star-topology configuration , photonic systems using polarizations of single photons , to superconducting circuits systems operating at milli-Kelvin temperatures . Compared to other platforms, superconducting circuits are widely concerned since its flexibility, adjustability, scalability and strong coupling with external fields . In 2019, Santos et al. have proposed that superconducting circuits is an effective platform to realize QB and the viewpoint has been confirmed in subsequent experiments which generally consist of one superconducting qutrit driven by classical fields . Remarkably, the semi-classical QB model designed on superconducting circuits can ensure a stable charged state by utilizing the stimulated Raman adiabatic passage, but has limited charging energy and fast self-discharging speed. Therefore, the exploration of many-body QB with efficient charging, stable energy-storage and slow self-discharging processes on superconducting circuits platform is necessary and meaningful. In this paper, we focus on a superconducting circuit system composed by a one-dimensional transmission line resonator and N coupled transmons. Under the condition that the anharmonicity of transmon is smaller than the transmon-resonator detuning, we derive the general Hamiltonian of this superconducting circuit system and define three QBs, including the one-mode or two-modes resonator-qutrit QBs and the resonator-qutrits QB. The main focus of the paper's investigation on the performance of quantum batteries is discussed here. The paper is organized as follows. In Sec. we derive the Hamiltonian of a superconducting circuit system and define a general QB model. In Sec. we investigate the performance of the QB in the presence of the resonator decay and the qutrit decoherence. The roles of quantum entanglement and quantum coherence during their charging and self-discharging processes are discussed in Sec. . Finally, a brief conclusion is given in Sec. .

What does the paper focus on in terms of the performance investigation of quantum batteries?

The paper investigates the performance of quantum batteries in the presence of resonator decay and qutrit decoherence, discussing the roles of quantum entanglement and coherence during charging and self-discharging.

purpose

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. The main conclusion about the origin of most FXTs is summarized here. Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

What is the main conclusion drawn about the origin of most FXTs discussed in this passage?

Based on limited photometric and spectroscopic information, most of these FXTs are suggested to be due to a supernova shock breakout (SN SBO).

summary

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An explanation of the first electromagnetic emission from a supernova explosion is provided here. This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

What is predicted to be the initial electromagnetic emission from a supernova explosion?

A supernova shock breakout (SBO) is predicted to be the first electromagnetic emission from a supernova explosion.

terminology

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. An example of an event leading to a longer-duration transient during a supernova shock breakout is discussed here. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

Can you provide an example of what might cause a longer-duration transient event during a supernova shock breakout?

If the circumstellar medium (CSM) is dense enough, the breakout occurs at larger radii within the CSM, resulting in a longer-duration transient event.

example

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Examples of physical mechanisms that might explain the origin of FXTs are mentioned here. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

What are some of the physical mechanisms suggested to be the origin of FXTs?

Physical mechanisms suggested as the origin of FXTs include a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger.

example

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Future sources of Fast X-ray Transients are hypothesized here. For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

What future events are expected to produce Fast X-ray Transients according to the passage?

Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs.

future

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). The reason for using a large beaming correction factor in this context is explained here. While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

Why is a large beaming correction factor necessary when comparing FXTs to SGRBs according to the passage?

A large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets.

purpose

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. The significance of conducting host galaxy studies for FXTs is articulated here. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

What is the importance of studying host galaxies in relation to FXTs?

Host studies provide vital information to decipher the origin of the FXTs.

purpose

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. An example of a host galaxy linked to an FXT is given here. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

Can you name a specific host galaxy associated with an FXT as mentioned in the text?

A dwarf galaxy host at ∼ z_ phot=0.4-3.2 is found to be associated with CDF-S XT1.

example

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). The cosmological model and parameters used in this study are defined here. Throughout, magnitudes are quoted in the AB system. We use the term "optical counterpart" for optical transient light associated with the FXT, and "candidate host" or "candidate host galaxy" for the host galaxy of the FXT. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

What cosmological model and parameters are assumed in the study discussed in the passage?

A flat Λ-CDM cosmology is assumed throughout the paper, with specific values for the Hubble constant (H_0) and matter density parameter (Ω_m).

terminology

arxiv

2312.10786v1

INTRODUCTION Fast X-ray Transients (FXTs) are singular bursts of soft X-rays that last from minutes to hours, whose origin is still unknown. The majority (22 out of the 36) of the known FXTs have been discovered in Chandra archival data (see ; ; ; ; ; ; ; ; ; ), while 12 have been found in archival XMM-Newton data (). [Note that out of the 12 XMM-Newton-discovered FXTs, three are suspected to be due to flares by late-type stars ().] In addition, one FXT was discovered in eROSITA data and one in Swift/XRT data (XRT 080109/SN2008D; ). Earlier observatories are presumed to have detected some FXTs, but lacked sufficient spatial resolution to pinpoint them as extragalactic phenomena. Among the 12 XMM-Newton-discovered FXTs, nine have candidate host galaxies, while for three no host has been identified to date. estimate the redshifts of those host galaxies to be between 0.1 and 1. Based on relatively limited photometric and spectroscopic information, they suggest that most of these FXTs are due to a supernova shock breakout (SN SBO). Specifically, these SN SBOs are consistent with the expectations for blue supergiant core-collapse supernovae (CC-SNe). An SBO is predicted to be the first electromagnetic emission from a supernova explosion (). This happens when the radiation-mediated shock crosses the surface of the star, thereby producing a bright X-ray/UV flash on a timescale of seconds to a fraction of an hour . This is followed by months-long SN emission, which is predominantly powered by radioactive decay and detectable at UV/optical/near-infrared (IR) wavelengths. The SBO emission carries information about the progenitor star, potentially enabling the inference of physical properties of the progenitor such as the stellar radius and mass-loss history. If the circumstellar medium (CSM) is dense enough, the breakout will occur at larger radii within the CSM, resulting in a longer-duration transient event. The only confirmed SN SBO observed to date is XRT 080109, associated with SN 2008D, serendipitously discovered in Swift observations of SN 2007uy (). XRT 080109 was detected in one of the arms of the spiral galaxy NGC 2770 at a distance of 27 Mpc. The X-ray light curve is characterised by a fast rise followed by an exponential decay with a peak X-ray luminosity of ∼ 6× 10^43 . Based on the multi-wavelength dataset, shows that the observed properties are consistent with those of Wolf-Rayet stars (R_∗ ∼ 10^11cm ), the favoured progenitors for type Ibc supernovae. More luminous SBOs have also been posited as an origin for some of the low luminosity long duration gamma-ray bursts (LGRBs) seen in the local Universe (e.g., GRB 060218; ). Several other physical mechanisms have been suggested as the origin of FXTs; specifically a white dwarf (WD) - intermediate-mass black hole (IMBH) tidal disruption event (TDE) and a binary neutron star (BNS) merger. In general, the X-ray light curves of TDEs involving a main sequence star and a supermassive black hole have a time scale ranging from months to years (). Instead, the relatively low mass of an IMBH compared to that of a supermassive black hole and the compactness of a WD compared to that of a main-sequence star implies that the time scale associated with the X-ray light curve of IMBH-WD TDEs is much shorter than that of an SMBH – main-sequence TDE (; ). A WD TDE remains a viable interpretation of at least some FXTs detected so far (e.g., XRT 000519; ). In order to be consistent with the observed volumetric rates of FXTs , one needs to assume that those IMBH-WD TDEs producing the X-ray emission are capable of launching jetted outflows. A beamed TDE emission would also naturally account for fairly high X-ray luminosities (), similar to the ones observed in some FXTs. Indeed, there are also apparent similarities between these events and the sub-set of ultra-long duration gamma-ray bursts (). A BNS merger is another possible progenitor for FXTs. Some BNS mergers could leave behind supramassive or even stable magnetars, which are expected to produce FXTs (; ; ; ; ). For some FXTs like CDF-S XT2 () and XRT 210423 (), the X-ray light curve shows a plateau and decay, similar to those seen in short gamma-ray bursts (SGRBs). This has been used as an argument in favour of a BNS origin. Since GW170817 it is known that SGRBs are caused by BNS mergers (). Converting the observed SGRB rate to an intrinsic rate that can be compared with the (nearly isotropic) FXT rate, a large beaming correction factor is required to reconcile the current observed FXTs volumetric density rate with that of SGRBs, providing potentially strong constraints on the opening angle of SGRB jets . While transient counterparts have been hard to find for FXTs (), we still expect progenitors of FXTs to be associated with a host. Host studies provide vital information to decipher the origin of the FXTs. For instance, finding the distance to the host galaxy helps constrain the energetics of the FXT (e.g., ; ). Understanding host environments can also play an important role in constraining the progenitor mechanism. Host studies have been undertaken for a few of the Chandra FXTs. finds a dwarf galaxy host (m_r=27.5) at ∼ z_ phot=0.4-3.2 that lies 0.13^'' offset from the center of the X-ray uncertainty position of CDF-S XT1. reports that CDF-XT2 () lies at a projected distance of 3.3 ± 1.9 kpc from a z_ spec=0.74 star-forming galaxy. Deep Very Large Telescope (VLT) and Gran Telescopio Canarias (GTC) observations of XRT 000519 () reveal a candidate host galaxy to the northwest of the FXT position in the g-band image at m_g=26.29±0.09 mag () which, if the host galaxy – FXT association is real, rules out an SN SBO origin for this FXT. For XRT 170901, an elongated, possibly late-type, star-forming galaxy is reported by to fall within the 2σ X-ray position. No clear host galaxy was found for XRT 030511 (). For XRT 210423, report a candidate host that lies within the 3σ uncertainty region with a magnitude of 25.9±0.1 in the GTC/HiPERCAM g_s-filter. In addition, they reported two other candidate host galaxies; one with z_ spec=1.5082 ± 0.0001 and offset 4.2±1 (37±9 kpc) from the FXT position and another one with z_ phot=1.04^+0.22_-0.14, offset by 3.6±1 (30±8 kpc). Irrespective of which of these three candidate host galaxies is the real host galaxy, if one of them is the host, the properties are consistent with a BNS merger origin for XRT 210423 (). In this paper, we look in detail at the host properties of 7 of the XMM-Newton-discovered FXTs reported by , namely , , , , , and . We also search for contemporaneous optical counterparts for those FXTs where such observations exist in either Pan-STARRS (), DECam () or ATLAS (). Throughout the paper, we assume a flat Λ-CDM cosmology, with Hubble constant H_0= 67.4 ± 0.5 km s^-1Mpc^-1 and matter density parameter Ω_m=0.315 ± 0.007 (). Throughout, magnitudes are quoted in the AB system. The definitions of 'optical counterpart' and 'candidate host galaxy' are presented here. The uncertainties mentioned in the paper are at the 1σ confidence level, unless mentioned otherwise.

How does the passage define the terms 'optical counterpart' and 'candidate host galaxy'?

The term 'optical counterpart' is used for optical transient light associated with the FXT, and 'candidate host' or 'candidate host galaxy' for the host galaxy of the FXT.

terminology

arxiv

2312.12037v1

The match between a startup founder's skills and their business idea is known as the 'founder-idea' fit. Venture capitalists often rely on platforms like LinkedIn to judge whether a founder has what it takes to succeed, but this method can be biased and misses out on patterns from past successes. To address this, we've developed a new method using language models that applies a range of techniques: embeddings to capture the nuances of a founder's profile, self-play for generating diverse perspectives, tree-of-thought for structured reasoning, critique-based refinement to improve accuracy, chain of thought to build reasoning steps, and least to most prompting to prioritize information effectively. This approach aims to give a clearer, more objective view of a founder's potential, helping investors make better decisions and founders build stronger teams.

What does the term 'founder-idea' fit refer to?

The 'founder-idea' fit refers to the match between a startup founder's skills and their business idea.

terminology

arxiv

2312.12037v1

INTRODUCTION This research tackles the challenge of assessing the match between a startup founder's skills and their business idea, known as the "founder-idea" fit, which is crucial for predicting startup success. Venture capitalists often rely on platforms like LinkedIn to judge whether a founder has what it takes to succeed, but this method can be biased and misses out on patterns from past successes. Here, a summary of the new method developed to assess the 'founder-idea' fit is provided. This approach aims to give a clearer, more objective view of a founder's potential, helping investors make better decisions and founders build stronger teams.

What has been developed to assess the 'founder-idea' fit and what does it involve?

A new method using language models has been developed that applies a range of techniques to assess 'founder-idea' fit.

summary

arxiv

2312.12037v1

INTRODUCTION This research tackles the challenge of assessing the match between a startup founder's skills and their business idea, known as the "founder-idea" fit, which is crucial for predicting startup success. Venture capitalists often rely on platforms like LinkedIn to judge whether a founder has what it takes to succeed, but this method can be biased and misses out on patterns from past successes. The techniques used in the new method for assessing the 'founder-idea' fit are listed here. This approach aims to give a clearer, more objective view of a founder's potential, helping investors make better decisions and founders build stronger teams.

Which range of techniques are used in the new method for assessing the 'founder-idea' fit?

Embeddings, self-play, tree-of-thought, critique-based refinement, chain of thought, and least to most prompting are the range of techniques used in the new method.

terminology

arxiv

2312.12037v1

INTRODUCTION This research tackles the challenge of assessing the match between a startup founder's skills and their business idea, known as the "founder-idea" fit, which is crucial for predicting startup success. Venture capitalists often rely on platforms like LinkedIn to judge whether a founder has what it takes to succeed, but this method can be biased and misses out on patterns from past successes. To address this, we've developed a new method using language models that applies a range of techniques: embeddings to capture the nuances of a founder's profile, self-play for generating diverse perspectives, tree-of-thought for structured reasoning, critique-based refinement to improve accuracy, chain of thought to build reasoning steps, and least to most prompting to prioritize information effectively. The intended purpose of the approach in providing insight into a founder's potential is indicated here.

What is the aim of the new method using language models in evaluating a founder's potential?

The approach aims to give a clearer, more objective view of a founder's potential.

purpose

arxiv

2312.12137v1

INTRODUCTION We study the problem of best-arm identification in stochastic bandits in the fixed budget setting. In this problem, abbreviated by BAI-FB, a learner faces certain challenges... She sequentially pulls arms and observes samples of the corresponding distributions. More precisely, in round t≥ 1, she pulls an arm A_t=k selected depending on previous observations and observes X_k(t) a sample of a ν_k-distributed random variable. (X_k(t),t≥ 1, k∈ [K]) are assumed to be independent over rounds and arms. After T arm draws, the learner returns , an estimate of the best arm 1() :=max_kμ_k. We assume that the best arm is unique, and denote by Λ the set of parameters such that this assumption holds. The objective is to devise an adaptive sampling algorithm minimizing the error probability ℙ_[≠ 1()]. This learning task is one of the most important problems in stochastic bandits, and despite recent research efforts, it remains largely open . In particular, researchers have so far failed at characterizing the minimal instance-specific error probability. This contrasts with other basic learning tasks in stochastic bandits such as regret minimization and BAI with fixed confidence , for which indeed, asymptotic instance-specific performance limits and matching algorithms have been derived. In BAI-FB, the error probability typically decreases exponentially with the sample budget T, i.e., it scales as exp(-R()T) where the instance-specific rate R() depends on the sampling algorithm. Maximizing this rate over the set of adaptive algorithms is an open problem. Instance-specific error probability lower bound. To guess the maximal rate at which the error probability decays, one may apply the same strategy as that used in regret minimization or BAI in the fixed confidence setting: (i) derive instance-specific lower bound for the error probability for some notion of uniformly good algorithms; (ii) devise a sampling strategy mimicking the optimal proportions of arm draws identified in the lower bound. Here the notion of uniformly good algorithms is that of consistent algorithms. Under such an algorithm, for any ∈Λ, _[= 1() ] → 1 as T→∞. conjectures the following asymptotic lower bound satisfied by any consistent algorithm (refer to Appendix for details): as T→∞, 1/Tlog1/_[ ≠ 1() ] ≤max_∈Σinf_∈ () Ψ(,) , where Σ is the (K-1)-dimensional simplex, Ψ(,)= ∑_k=1^Kω_k d(λ_k ,μ_k), ()={∈Λ: 1()≠ 1()} is the set of confusing parameters (those for which 1() is not the best arm), and d(x,y) denotes the KL divergence between two distributions of parameters x and y. Interestingly, the solution ^⋆∈Σ of the optimization problem max_∈Σinf_∈ () Ψ(,) provides the best static proportions of arm draws. More precisely, an algorithm selecting arms according to the allocation ^⋆, i.e., selecting arm k ω_k^⋆ T times and returning the best empirical arm after T samples, has an error rate matching the lower bound (). This is a direct consequence of the fact that, under a static algorithm with allocation , the empirical reward process {(t)}_t≥ 1 satisfies a LDP with rate function ↦Ψ(,), see and refer to Section for more details. Adaptive sampling algorithms and their analysis. The optimal allocation ^⋆ depends on the instance and is initially unknown. We may devise an adaptive sampling algorithm that (i) estimates ^⋆ and (ii) tracks this estimated optimal allocation. In the BAI with fixed confidence, such tracking scheme exhibits asymptotically optimal performance . Here however, the error made estimating ^⋆ would inevitably impact the overall error probability of the algorithm. To quantify this impact or more generally to analyze the performance of adaptive algorithms, one would need to understand the connection between the statistical properties of the arm selection process and the asymptotic statistics of the estimated expected rewards. To be more specific, any adaptive algorithm generates a stochastic process {Z(t)}_t≥ 1={((t), (t))}_t≥ 1. (t)=(ω_1(t),…,ω_K(t)) represents the allocation realized by the algorithm up to round t (ω_k(t)=N_k(t)/t and N_k(t) denotes the number of times arm k has been selected up to round t). (t)=(μ̂_1(t),…,μ̂_K(t)) denotes the empirical average rewards of the various arms up to round t. Now assuming that at the end of round T, the algorithm returns the arm with the highest empirical reward, the error probability is ℙ_[≠ 1()]=ℙ_[(T)∈ ()]. Assessing the error probability at least asymptotically requires understanding the asymptotic behavior of (t) as t grows large. Ideally, one would wish to establish the Large Deviation properties of the process {Z(t)}_t≥ 1. This task is easy for algorithms using static allocations , but becomes challenging and open for adaptive algorithms. Addressing this challenge is the main objective of this paper. Contributions. In this paper, we develop and leverage tools towards the analysis of adaptive sampling algorithms for the BAI-FB problem. More precisely, our contributions are as follows. (a) We establish a connection between the LDP satisfied by the empirical proportions of arm draws {(t)}_t≥ 1 and that satisfied by the empirical arm rewards. This connection holds for any adaptive algorithm. Specifically, we show that if the rate function of {(t)}_t≥ 1 is lower bounded by ↦ I(), then that of ((t))_t≥ 1 is also lower bounded by ↦min_∈Σmax{Ψ(,),I()}. This result has interesting interpretations and implies the following asymptotic upper bound on the error probability of the algorithm considered: as T→∞, 1/Tlog1/_[ ≠ 1() ] ≥inf_∈Σ, ∈ ()max{Ψ(,),I()}. The above formula, when compared to the lower bound (), quantifies the price of not knowing ^⋆ initially, and relates the error probability to the asymptotic statistics of the sampling process used by the algorithm. (b) We show that by simply applying our generic Large Deviation result, we may improve the error probability upper bounds of some existing algorithms, such as the celebrated algorithm . Our result further opens up opportunities to devise and analyze new algorithms with a higher level of adaptiveness. In particular, we present (Continuous Rejects), an algorithm that, unlike , can eliminate arms in each round. This sequential elimination process is performed by comparing the empirical rewards of the various candidate arms using continuously updated thresholds. Leveraging the LDP tools developed in (a), we establish that enjoys better performance guarantees than . Hence becomes the algorithm with the lowest instance-specific and guaranteed error probability. We illustrate our results via numerical experiments, and compare to other BAI algorithms.

What does BAI-FB stand for in the context of stochastic bandits?

BAI-FB stands for the problem of best-arm identification in stochastic bandits in the fixed budget setting

terminology

arxiv

2312.12137v1

INTRODUCTION We study the problem of best-arm identification in stochastic bandits in the fixed budget setting. In this problem, abbreviated by BAI-FB, a learner faces K distributions or arms ν_1,…,ν_K characterized by their unknown means =(μ_1,…,μ_K) (we restrict our attention to distributions taken from a one-parameter exponential family). She sequentially pulls arms and observes samples of the corresponding distributions. More precisely, in round t≥ 1, she pulls an arm A_t=k selected depending on previous observations and observes X_k(t) a sample of a ν_k-distributed random variable. (X_k(t),t≥ 1, k∈ [K]) are assumed to be independent over rounds and arms. After T arm draws, the learner returns ..., an estimate of the best arm... We assume that the best arm is unique, and denote by Λ the set of parameters such that this assumption holds. The objective is to devise an adaptive sampling algorithm minimizing the error probability ℙ_[≠ 1()]. This learning task is one of the most important problems in stochastic bandits, and despite recent research efforts, it remains largely open . In particular, researchers have so far failed at characterizing the minimal instance-specific error probability. This contrasts with other basic learning tasks in stochastic bandits such as regret minimization and BAI with fixed confidence , for which indeed, asymptotic instance-specific performance limits and matching algorithms have been derived. In BAI-FB, the error probability typically decreases exponentially with the sample budget T, i.e., it scales as exp(-R()T) where the instance-specific rate R() depends on the sampling algorithm. Maximizing this rate over the set of adaptive algorithms is an open problem. Instance-specific error probability lower bound. To guess the maximal rate at which the error probability decays, one may apply the same strategy as that used in regret minimization or BAI in the fixed confidence setting: (i) derive instance-specific lower bound for the error probability for some notion of uniformly good algorithms; (ii) devise a sampling strategy mimicking the optimal proportions of arm draws identified in the lower bound. Here the notion of uniformly good algorithms is that of consistent algorithms. Under such an algorithm, for any ∈Λ, _[= 1() ] → 1 as T→∞. conjectures the following asymptotic lower bound satisfied by any consistent algorithm (refer to Appendix for details): as T→∞, 1/Tlog1/_[ ≠ 1() ] ≤max_∈Σinf_∈ () Ψ(,) , where Σ is the (K-1)-dimensional simplex, Ψ(,)= ∑_k=1^Kω_k d(λ_k ,μ_k), ()={∈Λ: 1()≠ 1()} is the set of confusing parameters (those for which 1() is not the best arm), and d(x,y) denotes the KL divergence between two distributions of parameters x and y. Interestingly, the solution ^⋆∈Σ of the optimization problem max_∈Σinf_∈ () Ψ(,) provides the best static proportions of arm draws. More precisely, an algorithm selecting arms according to the allocation ^⋆, i.e., selecting arm k ω_k^⋆ T times and returning the best empirical arm after T samples, has an error rate matching the lower bound (). This is a direct consequence of the fact that, under a static algorithm with allocation , the empirical reward process {(t)}_t≥ 1 satisfies a LDP with rate function ↦Ψ(,), see and refer to Section for more details. Adaptive sampling algorithms and their analysis. The optimal allocation ^⋆ depends on the instance and is initially unknown. We may devise an adaptive sampling algorithm that (i) estimates ^⋆ and (ii) tracks this estimated optimal allocation. In the BAI with fixed confidence, such tracking scheme exhibits asymptotically optimal performance . Here however, the error made estimating ^⋆ would inevitably impact the overall error probability of the algorithm. To quantify this impact or more generally to analyze the performance of adaptive algorithms, one would need to understand the connection between the statistical properties of the arm selection process and the asymptotic statistics of the estimated expected rewards. To be more specific, any adaptive algorithm generates a stochastic process {Z(t)}_t≥ 1={((t), (t))}_t≥ 1. (t)=(ω_1(t),…,ω_K(t)) represents the allocation realized by the algorithm up to round t (ω_k(t)=N_k(t)/t and N_k(t) denotes the number of times arm k has been selected up to round t). (t)=(μ̂_1(t),…,μ̂_K(t)) denotes the empirical average rewards of the various arms up to round t. Now assuming that at the end of round T, the algorithm returns the arm with the highest empirical reward, the error probability is ℙ_[≠ 1()]=ℙ_[(T)∈ ()]. Assessing the error probability at least asymptotically requires understanding the asymptotic behavior of (t) as t grows large. Ideally, one would wish to establish the Large Deviation properties of the process {Z(t)}_t≥ 1. This task is easy for algorithms using static allocations , but becomes challenging and open for adaptive algorithms. Addressing this challenge is the main objective of this paper. Contributions. In this paper, we develop and leverage tools towards the analysis of adaptive sampling algorithms for the BAI-FB problem. More precisely, our contributions are as follows. (a) We establish a connection between the LDP satisfied by the empirical proportions of arm draws {(t)}_t≥ 1 and that satisfied by the empirical arm rewards. This connection holds for any adaptive algorithm. Specifically, we show that if the rate function of {(t)}_t≥ 1 is lower bounded by ↦ I(), then that of ((t))_t≥ 1 is also lower bounded by ↦min_∈Σmax{Ψ(,),I()}. This result has interesting interpretations and implies the following asymptotic upper bound on the error probability of the algorithm considered: as T→∞, 1/Tlog1/_[ ≠ 1() ] ≥inf_∈Σ, ∈ ()max{Ψ(,),I()}. The above formula, when compared to the lower bound (), quantifies the price of not knowing ^⋆ initially, and relates the error probability to the asymptotic statistics of the sampling process used by the algorithm. (b) We show that by simply applying our generic Large Deviation result, we may improve the error probability upper bounds of some existing algorithms, such as the celebrated algorithm . Our result further opens up opportunities to devise and analyze new algorithms with a higher level of adaptiveness. In particular, we present (Continuous Rejects), an algorithm that, unlike , can eliminate arms in each round. This sequential elimination process is performed by comparing the empirical rewards of the various candidate arms using continuously updated thresholds. Leveraging the LDP tools developed in (a), we establish that enjoys better performance guarantees than . Hence becomes the algorithm with the lowest instance-specific and guaranteed error probability. We illustrate our results via numerical experiments, and compare to other BAI algorithms.

How is an estimate of the best arm represented in the given passage?

An estimate of the best arm is denoted as ̂*

terminology

arxiv

2312.12137v1

INTRODUCTION We study the problem of best-arm identification in stochastic bandits in the fixed budget setting. In this problem, abbreviated by BAI-FB, a learner faces K distributions or arms ν_1,…,ν_K characterized by their unknown means =(μ_1,…,μ_K) (we restrict our attention to distributions taken from a one-parameter exponential family). She sequentially pulls arms and observes samples of the corresponding distributions. More precisely, in round t≥ 1, she pulls an arm A_t=k selected depending on previous observations and observes X_k(t) a sample of a ν_k-distributed random variable. (X_k(t),t≥ 1, k∈ [K]) are assumed to be independent over rounds and arms. After T arm draws, the learner returns , an estimate of the best arm 1() :=max_kμ_k. We assume that the best arm is unique, and denote by Λ the set of parameters such that this assumption holds. The objective is to devise an adaptive sampling algorithm minimizing... This learning task is one of the most important problems in stochastic bandits, and despite recent research efforts, it remains largely open . In particular, researchers have so far failed at characterizing the minimal instance-specific error probability. This contrasts with other basic learning tasks in stochastic bandits such as regret minimization and BAI with fixed confidence , for which indeed, asymptotic instance-specific performance limits and matching algorithms have been derived. In BAI-FB, the error probability typically decreases exponentially with the sample budget T, i.e., it scales as exp(-R()T) where the instance-specific rate R() depends on the sampling algorithm. Maximizing this rate over the set of adaptive algorithms is an open problem. Instance-specific error probability lower bound. To guess the maximal rate at which the error probability decays, one may apply the same strategy as that used in regret minimization or BAI in the fixed confidence setting: (i) derive instance-specific lower bound for the error probability for some notion of uniformly good algorithms; (ii) devise a sampling strategy mimicking the optimal proportions of arm draws identified in the lower bound. Here the notion of uniformly good algorithms is that of consistent algorithms. Under such an algorithm, for any ∈Λ, _[= 1() ] → 1 as T→∞. conjectures the following asymptotic lower bound satisfied by any consistent algorithm (refer to Appendix for details): as T→∞, 1/Tlog1/_[ ≠ 1() ] ≤max_∈Σinf_∈ () Ψ(,) , where Σ is the (K-1)-dimensional simplex, Ψ(,)= ∑_k=1^Kω_k d(λ_k ,μ_k), ()={∈Λ: 1()≠ 1()} is the set of confusing parameters (those for which 1() is not the best arm), and d(x,y) denotes the KL divergence between two distributions of parameters x and y. Interestingly, the solution ^⋆∈Σ of the optimization problem max_∈Σinf_∈ () Ψ(,) provides the best static proportions of arm draws. More precisely, an algorithm selecting arms according to the allocation ^⋆, i.e., selecting arm k ω_k^⋆ T times and returning the best empirical arm after T samples, has an error rate matching the lower bound (). This is a direct consequence of the fact that, under a static algorithm with allocation , the empirical reward process {(t)}_t≥ 1 satisfies a LDP with rate function ↦Ψ(,), see and refer to Section for more details. Adaptive sampling algorithms and their analysis. The optimal allocation ^⋆ depends on the instance and is initially unknown. We may devise an adaptive sampling algorithm that (i) estimates ^⋆ and (ii) tracks this estimated optimal allocation. In the BAI with fixed confidence, such tracking scheme exhibits asymptotically optimal performance . Here however, the error made estimating ^⋆ would inevitably impact the overall error probability of the algorithm. To quantify this impact or more generally to analyze the performance of adaptive algorithms, one would need to understand the connection between the statistical properties of the arm selection process and the asymptotic statistics of the estimated expected rewards. To be more specific, any adaptive algorithm generates a stochastic process {Z(t)}_t≥ 1={((t), (t))}_t≥ 1. (t)=(ω_1(t),…,ω_K(t)) represents the allocation realized by the algorithm up to round t (ω_k(t)=N_k(t)/t and N_k(t) denotes the number of times arm k has been selected up to round t). (t)=(μ̂_1(t),…,μ̂_K(t)) denotes the empirical average rewards of the various arms up to round t. Now assuming that at the end of round T, the algorithm returns the arm with the highest empirical reward, the error probability is ℙ_[≠ 1()]=ℙ_[(T)∈ ()]. Assessing the error probability at least asymptotically requires understanding the asymptotic behavior of (t) as t grows large. Ideally, one would wish to establish the Large Deviation properties of the process {Z(t)}_t≥ 1. This task is easy for algorithms using static allocations , but becomes challenging and open for adaptive algorithms. Addressing this challenge is the main objective of this paper. Contributions. In this paper, we develop and leverage tools towards the analysis of adaptive sampling algorithms for the BAI-FB problem. More precisely, our contributions are as follows. (a) We establish a connection between the LDP satisfied by the empirical proportions of arm draws {(t)}_t≥ 1 and that satisfied by the empirical arm rewards. This connection holds for any adaptive algorithm. Specifically, we show that if the rate function of {(t)}_t≥ 1 is lower bounded by ↦ I(), then that of ((t))_t≥ 1 is also lower bounded by ↦min_∈Σmax{Ψ(,),I()}. This result has interesting interpretations and implies the following asymptotic upper bound on the error probability of the algorithm considered: as T→∞, 1/Tlog1/_[ ≠ 1() ] ≥inf_∈Σ, ∈ ()max{Ψ(,),I()}. The above formula, when compared to the lower bound (), quantifies the price of not knowing ^⋆ initially, and relates the error probability to the asymptotic statistics of the sampling process used by the algorithm. (b) We show that by simply applying our generic Large Deviation result, we may improve the error probability upper bounds of some existing algorithms, such as the celebrated algorithm . Our result further opens up opportunities to devise and analyze new algorithms with a higher level of adaptiveness. In particular, we present (Continuous Rejects), an algorithm that, unlike , can eliminate arms in each round. This sequential elimination process is performed by comparing the empirical rewards of the various candidate arms using continuously updated thresholds. Leveraging the LDP tools developed in (a), we establish that enjoys better performance guarantees than . Hence becomes the algorithm with the lowest instance-specific and guaranteed error probability. We illustrate our results via numerical experiments, and compare to other BAI algorithms.

What is the main objective for the learner in the BAI-FB problem?

The objective of the learner in BAI-FB is to devise an adaptive sampling algorithm minimizing the error probability P_{θ̂* ≠ μ*}

purpose