Datasets:

CShorten

/

Last-Week-on-ML-ArXiv

like 0

Enabling Capsule Networks at the Edge through Approximate Softmax and Squash Operations

Complex Deep Neural Networks such as Capsule Networks (CapsNets) exhibit high learning capabilities at the cost of compute-intensive operations. To enable their deployment on edge devices, we propose to leverage approximate computing for designing approximate variants of the complex operations like softmax and squash. In our experiments, we evaluate tradeoffs between area, power consumption, and critical path delay of the designs implemented with the ASIC design flow, and the accuracy of the quantized CapsNets, compared to the exact functions.

A General Theory for Federated Optimization with Asynchronous and Heterogeneous Clients Updates

We propose a novel framework to study asynchronous federated learning optimization with delays in gradient updates. Our theoretical framework extends the standard FedAvg aggregation scheme by introducing stochastic aggregation weights to represent the variability of the clients update time, due for example to heterogeneous hardware capabilities. Our formalism applies to the general federated setting where clients have heterogeneous datasets and perform at least one step of stochastic gradient descent (SGD). We demonstrate convergence for such a scheme and provide sufficient conditions for the related minimum to be the optimum of the federated problem. We show that our general framework applies to existing optimization schemes including centralized learning, FedAvg, asynchronous FedAvg, and FedBuff. The theory here provided allows drawing meaningful guidelines for designing a federated learning experiment in heterogeneous conditions. In particular, we develop in this work FedFix, a novel extension of FedAvg enabling efficient asynchronous federated training while preserving the convergence stability of synchronous aggregation. We empirically demonstrate our theory on a series of experiments showing that asynchronous FedAvg leads to fast convergence at the expense of stability, and we finally demonstrate the improvements of FedFix over synchronous and asynchronous FedAvg.

Analysis of Self-Supervised Learning and Dimensionality Reduction Methods in Clustering-Based Active Learning for Speech Emotion Recognition

When domain experts are needed to perform data annotation for complex machine-learning tasks, reducing annotation effort is crucial in order to cut down time and expenses. For cases when there are no annotations available, one approach is to utilize the structure of the feature space for clustering-based active learning (AL) methods. However, these methods are heavily dependent on how the samples are organized in the feature space and what distance metric is used. Unsupervised methods such as contrastive predictive coding (CPC) can potentially be used to learn organized feature spaces, but these methods typically create high-dimensional features which might be challenging for estimating data density. In this paper, we combine CPC and multiple dimensionality reduction methods in search of functioning practices for clustering-based AL. Our experiments for simulating speech emotion recognition system deployment show that both the local and global topology of the feature space can be successfully used for AL, and that CPC can be used to improve clustering-based AL performance over traditional signal features. Additionally, we observe that compressing data dimensionality does not harm AL performance substantially, and that 2-D feature representations achieved similar AL performance as higher-dimensional representations when the number of annotations is not very low.

Federated Reinforcement Learning: Linear Speedup Under Markovian Sampling

Since reinforcement learning algorithms are notoriously data-intensive, the task of sampling observations from the environment is usually split across multiple agents. However, transferring these observations from the agents to a central location can be prohibitively expensive in terms of the communication cost, and it can also compromise the privacy of each agent's local behavior policy. In this paper, we consider a federated reinforcement learning framework where multiple agents collaboratively learn a global model, without sharing their individual data and policies. Each agent maintains a local copy of the model and updates it using locally sampled data. Although having N agents enables the sampling of N times more data, it is not clear if it leads to proportional convergence speedup. We propose federated versions of on-policy TD, off-policy TD and Q-learning, and analyze their convergence. For all these algorithms, to the best of our knowledge, we are the first to consider Markovian noise and multiple local updates, and prove a linear convergence speedup with respect to the number of agents. To obtain these results, we show that federated TD and Q-learning are special cases of a general framework for federated stochastic approximation with Markovian noise, and we leverage this framework to provide a unified convergence analysis that applies to all the algorithms.

Knowledge Graph Fusion for Language Model Fine-tuning

Language Models such as BERT have grown in popularity due to their ability to be pre-trained and perform robustly on a wide range of Natural Language Processing tasks. Often seen as an evolution over traditional word embedding techniques, they can produce semantic representations of text, useful for tasks such as semantic similarity. However, state-of-the-art models often have high computational requirements and lack global context or domain knowledge which is required for complete language understanding. To address these limitations, we investigate the benefits of knowledge incorporation into the fine-tuning stages of BERT. An existing K-BERT model, which enriches sentences with triplets from a Knowledge Graph, is adapted for the English language and extended to inject contextually relevant information into sentences. As a side-effect, changes made to K-BERT for accommodating the English language also extend to other word-based languages. Experiments conducted indicate that injected knowledge introduces noise. We see statistically significant improvements for knowledge-driven tasks when this noise is minimised. We show evidence that, given the appropriate task, modest injection with relevant, high-quality knowledge is most performant.

Predicting Parking Lot Availability by Graph-to-Sequence Model: A Case Study with SmartSantander

Nowadays, so as to improve services and urban areas livability, multiple smart city initiatives are being carried out throughout the world. SmartSantander is a smart city project in Santander, Spain, which has relied on wireless sensor network technologies to deploy heterogeneous sensors within the city to measure multiple parameters, including outdoor parking information. In this paper, we study the prediction of parking lot availability using historical data from more than 300 outdoor parking sensors with SmartSantander. We design a graph-to-sequence model to capture the periodical fluctuation and geographical proximity of parking lots. For developing and evaluating our model, we use a 3-year dataset of parking lot availability in the city of Santander. Our model achieves a high accuracy compared with existing sequence-to-sequence models, which is accurate enough to provide a parking information service in the city. We apply our model to a smartphone application to be widely used by citizens and tourists.

Certifiably Robust Policy Learning against Adversarial Communication in Multi-agent Systems

Communication is important in many multi-agent reinforcement learning (MARL) problems for agents to share information and make good decisions. However, when deploying trained communicative agents in a real-world application where noise and potential attackers exist, the safety of communication-based policies becomes a severe issue that is underexplored. Specifically, if communication messages are manipulated by malicious attackers, agents relying on untrustworthy communication may take unsafe actions that lead to catastrophic consequences. Therefore, it is crucial to ensure that agents will not be misled by corrupted communication, while still benefiting from benign communication. In this work, we consider an environment with $N$ agents, where the attacker may arbitrarily change the communication from any $C<\frac{N-1}{2}$ agents to a victim agent. For this strong threat model, we propose a certifiable defense by constructing a message-ensemble policy that aggregates multiple randomly ablated message sets. Theoretical analysis shows that this message-ensemble policy can utilize benign communication while being certifiably robust to adversarial communication, regardless of the attacking algorithm. Experiments in multiple environments verify that our defense significantly improves the robustness of trained policies against various types of attacks.

Deep Reinforcement Learning for Turbulence Modeling in Large Eddy Simulations

Over the last years, supervised learning (SL) has established itself as the state-of-the-art for data-driven turbulence modeling. In the SL paradigm, models are trained based on a dataset, which is typically computed a priori from a high-fidelity solution by applying the respective filter function, which separates the resolved and the unresolved flow scales. For implicitly filtered large eddy simulation (LES), this approach is infeasible, since here, the employed discretization itself acts as an implicit filter function. As a consequence, the exact filter form is generally not known and thus, the corresponding closure terms cannot be computed even if the full solution is available. The reinforcement learning (RL) paradigm can be used to avoid this inconsistency by training not on a previously obtained training dataset, but instead by interacting directly with the dynamical LES environment itself. This allows to incorporate the potentially complex implicit LES filter into the training process by design. In this work, we apply a reinforcement learning framework to find an optimal eddy-viscosity for implicitly filtered large eddy simulations of forced homogeneous isotropic turbulence. For this, we formulate the task of turbulence modeling as an RL task with a policy network based on convolutional neural networks that adapts the eddy-viscosity in LES dynamically in space and time based on the local flow state only. We demonstrate that the trained models can provide long-term stable simulations and that they outperform established analytical models in terms of accuracy. In addition, the models generalize well to other resolutions and discretizations. We thus demonstrate that RL can provide a framework for consistent, accurate and stable turbulence modeling especially for implicitly filtered LES.

Open-Source Framework for Encrypted Internet and Malicious Traffic Classification

Internet traffic classification plays a key role in network visibility, Quality of Services (QoS), intrusion detection, Quality of Experience (QoE) and traffic-trend analyses. In order to improve privacy, integrity, confidentiality, and protocol obfuscation, the current traffic is based on encryption protocols, e.g., SSL/TLS. With the increased use of Machine-Learning (ML) and Deep-Learning (DL) models in the literature, comparison between different models and methods has become cumbersome and difficult due to a lack of a standardized framework. In this paper, we propose an open-source framework, named OSF-EIMTC, which can provide the full pipeline of the learning process. From the well-known datasets to extracting new and well-known features, it provides implementations of well-known ML and DL models (from the traffic classification literature) as well as evaluations. Such a framework can facilitate research in traffic classification domains, so that it will be more repeatable, reproducible, easier to execute, and will allow a more accurate comparison of well-known and novel features and models. As part of our framework evaluation, we demonstrate a variety of cases where the framework can be of use, utilizing multiple datasets, models, and feature sets. We show analyses of publicly available datasets and invite the community to participate in our open challenges using the OSF-EIMTC.

A Contrastive Approach to Online Change Point Detection

We suggest a novel procedure for online change point detection. Our approach expands an idea of maximizing a discrepancy measure between points from pre-change and post-change distributions. This leads to a flexible procedure suitable for both parametric and nonparametric scenarios. We prove non-asymptotic bounds on the average running length of the procedure and its expected detection delay. The efficiency of the algorithm is illustrated with numerical experiments on synthetic and real-world data sets.

Propagation with Adaptive Mask then Training for Node Classification on Attributed Networks

Node classification on attributed networks is a semi-supervised task that is crucial for network analysis. By decoupling two critical operations in Graph Convolutional Networks (GCNs), namely feature transformation and neighborhood aggregation, some recent works of decoupled GCNs could support the information to propagate deeper and achieve advanced performance. However, they follow the traditional structure-aware propagation strategy of GCNs, making it hard to capture the attribute correlation of nodes and sensitive to the structure noise described by edges whose two endpoints belong to different categories. To address these issues, we propose a new method called the itshape Propagation with Adaptive Mask then Training (PAMT). The key idea is to integrate the attribute similarity mask into the structure-aware propagation process. In this way, PAMT could preserve the attribute correlation of adjacent nodes during the propagation and effectively reduce the influence of structure noise. Moreover, we develop an iterative refinement mechanism to update the similarity mask during the training process for improving the training performance. Extensive experiments on four real-world datasets demonstrate the superior performance and robustness of PAMT.

Comprehensive Analysis of Negative Sampling in Knowledge Graph Representation Learning

Negative sampling (NS) loss plays an important role in learning knowledge graph embedding (KGE) to handle a huge number of entities. However, the performance of KGE degrades without hyperparameters such as the margin term and number of negative samples in NS loss being appropriately selected. Currently, empirical hyperparameter tuning addresses this problem at the cost of computational time. To solve this problem, we theoretically analyzed NS loss to assist hyperparameter tuning and understand the better use of the NS loss in KGE learning. Our theoretical analysis showed that scoring methods with restricted value ranges, such as TransE and RotatE, require appropriate adjustment of the margin term or the number of negative samples different from those without restricted value ranges, such as RESCAL, ComplEx, and DistMult. We also propose subsampling methods specialized for the NS loss in KGE studied from a theoretical aspect. Our empirical analysis on the FB15k-237, WN18RR, and YAGO3-10 datasets showed that the results of actually trained models agree with our theoretical findings.

Insights into Pre-training via Simpler Synthetic Tasks

Pre-training produces representations that are effective for a wide range of downstream tasks, but it is still unclear what properties of pre-training are necessary for effective gains. Notably, recent work shows that even pre-training on synthetic tasks can achieve significant gains in downstream tasks. In this work, we perform three experiments that iteratively simplify pre-training and show that the simplifications still retain much of its gains. First, building on prior work, we perform a systematic evaluation of three existing synthetic pre-training methods on six downstream tasks. We find the best synthetic pre-training method, LIME, attains an average of $67\%$ of the benefits of natural pre-training. Second, to our surprise, we find that pre-training on a simple and generic synthetic task defined by the Set function achieves $65\%$ of the benefits, almost matching LIME. Third, we find that $39\%$ of the benefits can be attained by using merely the parameter statistics of synthetic pre-training. We release the source code at https://github.com/felixzli/synthetic_pretraining.

Few-Max: Few-Shot Domain Adaptation for Unsupervised Contrastive Representation Learning

Contrastive self-supervised learning methods learn to map data points such as images into non-parametric representation space without requiring labels. While highly successful, current methods require a large amount of data in the training phase. In situations where the target training set is limited in size, generalization is known to be poor. Pretraining on a large source data set and fine-tuning on the target samples is prone to overfitting in the few-shot regime, where only a small number of target samples are available. Motivated by this, we propose a domain adaption method for self-supervised contrastive learning, termed Few-Max, to address the issue of adaptation to a target distribution under few-shot learning. To quantify the representation quality, we evaluate Few-Max on a range of source and target datasets, including ImageNet, VisDA, and fastMRI, on which Few-Max consistently outperforms other approaches.

Transferable Graph Backdoor Attack

Graph Neural Networks (GNNs) have achieved tremendous success in many graph mining tasks benefitting from the message passing strategy that fuses the local structure and node features for better graph representation learning. Despite the success of GNNs, and similar to other types of deep neural networks, GNNs are found to be vulnerable to unnoticeable perturbations on both graph structure and node features. Many adversarial attacks have been proposed to disclose the fragility of GNNs under different perturbation strategies to create adversarial examples. However, vulnerability of GNNs to successful backdoor attacks was only shown recently. In this paper, we disclose the TRAP attack, a Transferable GRAPh backdoor attack. The core attack principle is to poison the training dataset with perturbation-based triggers that can lead to an effective and transferable backdoor attack. The perturbation trigger for a graph is generated by performing the perturbation actions on the graph structure via a gradient based score matrix from a surrogate model. Compared with prior works, TRAP attack is different in several ways: i) it exploits a surrogate Graph Convolutional Network (GCN) model to generate perturbation triggers for a blackbox based backdoor attack; ii) it generates sample-specific perturbation triggers which do not have a fixed pattern; and iii) the attack transfers, for the first time in the context of GNNs, to different GNN models when trained with the forged poisoned training dataset. Through extensive evaluations on four real-world datasets, we demonstrate the effectiveness of the TRAP attack to build transferable backdoors in four different popular GNNs using four real-world datasets.

Automatic Concept Extraction for Concept Bottleneck-based Video Classification

Recent efforts in interpretable deep learning models have shown that concept-based explanation methods achieve competitive accuracy with standard end-to-end models and enable reasoning and intervention about extracted high-level visual concepts from images, e.g., identifying the wing color and beak length for bird-species classification. However, these concept bottleneck models rely on a necessary and sufficient set of predefined concepts-which is intractable for complex tasks such as video classification. For complex tasks, the labels and the relationship between visual elements span many frames, e.g., identifying a bird flying or catching prey-necessitating concepts with various levels of abstraction. To this end, we present CoDEx, an automatic Concept Discovery and Extraction module that rigorously composes a necessary and sufficient set of concept abstractions for concept-based video classification. CoDEx identifies a rich set of complex concept abstractions from natural language explanations of videos-obviating the need to predefine the amorphous set of concepts. To demonstrate our method's viability, we construct two new public datasets that combine existing complex video classification datasets with short, crowd-sourced natural language explanations for their labels. Our method elicits inherent complex concept abstractions in natural language to generalize concept-bottleneck methods to complex tasks.

Safe and Psychologically Pleasant Traffic Signal Control with Reinforcement Learning using Action Masking

Reinforcement learning (RL) for traffic signal control (TSC) has shown better performance in simulation for controlling the traffic flow of intersections than conventional approaches. However, due to several challenges, no RL-based TSC has been deployed in the field yet. One major challenge for real-world deployment is to ensure that all safety requirements are met at all times during operation. We present an approach to ensure safety in a real-world intersection by using an action space that is safe by design. The action space encompasses traffic phases, which represent the combination of non-conflicting signal colors of the intersection. Additionally, an action masking mechanism makes sure that only appropriate phase transitions are carried out. Another challenge for real-world deployment is to ensure a control behavior that avoids stress for road users. We demonstrate how to achieve this by incorporating domain knowledge through extending the action masking mechanism. We test and verify our approach in a realistic simulation scenario. By ensuring safety and psychologically pleasant control behavior, our approach drives development towards real-world deployment of RL for TSC.

Finite Expression Method for Solving High-Dimensional Partial Differential Equations

Designing efficient and accurate numerical solvers for high-dimensional partial differential equations (PDEs) remains a challenging and important topic in computational science and engineering, mainly due to the ``curse of dimensionality" in designing numerical schemes that scale in dimension. This paper introduces a new methodology that seeks an approximate PDE solution in the space of functions with finitely many analytic expressions and, hence, this methodology is named the finite expression method (FEX). It is proved in approximation theory that FEX can avoid the curse of dimensionality. As a proof of concept, a deep reinforcement learning method is proposed to implement FEX for various high-dimensional PDEs in different dimensions, achieving high and even machine accuracy with a memory complexity polynomial in dimension and an amenable time complexity. An approximate solution with finite analytic expressions also provides interpretable insights into the ground truth PDE solution, which can further help to advance the understanding of physical systems and design postprocessing techniques for a refined solution.

Automatic Controllable Product Copywriting for E-Commerce

Automatic product description generation for e-commerce has witnessed significant advancement in the past decade. Product copywriting aims to attract users' interest and improve user experience by highlighting product characteristics with textual descriptions. As the services provided by e-commerce platforms become diverse, it is necessary to adapt the patterns of automatically-generated descriptions dynamically. In this paper, we report our experience in deploying an E-commerce Prefix-based Controllable Copywriting Generation (EPCCG) system into the JD.com e-commerce product recommendation platform. The development of the system contains two main components: 1) copywriting aspect extraction; 2) weakly supervised aspect labeling; 3) text generation with a prefix-based language model; 4) copywriting quality control. We conduct experiments to validate the effectiveness of the proposed EPCCG. In addition, we introduce the deployed architecture which cooperates with the EPCCG into the real-time JD.com e-commerce recommendation platform and the significant payoff since deployment.

Model-Based Imitation Learning Using Entropy Regularization of Model and Policy

Approaches based on generative adversarial networks for imitation learning are promising because they are sample efficient in terms of expert demonstrations. However, training a generator requires many interactions with the actual environment because model-free reinforcement learning is adopted to update a policy. To improve the sample efficiency using model-based reinforcement learning, we propose model-based Entropy-Regularized Imitation Learning (MB-ERIL) under the entropy-regularized Markov decision process to reduce the number of interactions with the actual environment. MB-ERIL uses two discriminators. A policy discriminator distinguishes the actions generated by a robot from expert ones, and a model discriminator distinguishes the counterfactual state transitions generated by the model from the actual ones. We derive the structured discriminators so that the learning of the policy and the model is efficient. Computer simulations and real robot experiments show that MB-ERIL achieves a competitive performance and significantly improves the sample efficiency compared to baseline methods.

Reconstruct from Top View: A 3D Lane Detection Approach based on Geometry Structure Prior

In this paper, we propose an advanced approach in targeting the problem of monocular 3D lane detection by leveraging geometry structure underneath the process of 2D to 3D lane reconstruction. Inspired by previous methods, we first analyze the geometry heuristic between the 3D lane and its 2D representation on the ground and propose to impose explicit supervision based on the structure prior, which makes it achievable to build inter-lane and intra-lane relationships to facilitate the reconstruction of 3D lanes from local to global. Second, to reduce the structure loss in 2D lane representation, we directly extract top view lane information from front view images, which tremendously eases the confusion of distant lane features in previous methods. Furthermore, we propose a novel task-specific data augmentation method by synthesizing new training data for both segmentation and reconstruction tasks in our pipeline, to counter the imbalanced data distribution of camera pose and ground slope to improve generalization on unseen data. Our work marks the first attempt to employ the geometry prior information into DNN-based 3D lane detection and makes it achievable for detecting lanes in an extra-long distance, doubling the original detection range. The proposed method can be smoothly adopted by other frameworks without extra costs. Experimental results show that our work outperforms state-of-the-art approaches by 3.8% F-Score on Apollo 3D synthetic dataset at real-time speed of 82 FPS without introducing extra parameters.

DeePKS+ABACUS as a Bridge between Expensive Quantum Mechanical Models and Machine Learning Potentials

Recently, the development of machine learning (ML) potentials has made it possible to perform large-scale and long-time molecular simulations with the accuracy of quantum mechanical (QM) models. However, for high-level QM methods, such as density functional theory (DFT) at the meta-GGA level and/or with exact exchange, quantum Monte Carlo, etc., generating a sufficient amount of data for training a ML potential has remained computationally challenging due to their high cost. In this work, we demonstrate that this issue can be largely alleviated with Deep Kohn-Sham (DeePKS), a ML-based DFT model. DeePKS employs a computationally efficient neural network-based functional model to construct a correction term added upon a cheap DFT model. Upon training, DeePKS offers closely-matched energies and forces compared with high-level QM method, but the number of training data required is orders of magnitude less than that required for training a reliable ML potential. As such, DeePKS can serve as a bridge between expensive QM models and ML potentials: one can generate a decent amount of high-accuracy QM data to train a DeePKS model, and then use the DeePKS model to label a much larger amount of configurations to train a ML potential. This scheme for periodic systems is implemented in a DFT package ABACUS, which is open-source and ready for use in various applications.

Renormalized Sparse Neural Network Pruning

Large neural networks are heavily over-parameterized. This is done because it improves training to optimality. However once the network is trained, this means many parameters can be zeroed, or pruned, leaving an equivalent sparse neural network. We propose renormalizing sparse neural networks in order to improve accuracy. We prove that our method's error converges to 0 as network parameters cluster or concentrate. We prove that without renormalizing, the error does not converge to zero in general. We experiment with our method on real world datasets MNIST, Fashion MNIST, and CIFAR-10 and confirm a large improvement in accuracy with renormalization versus standard pruning.

A Simple Approach for Visual Rearrangement: 3D Mapping and Semantic Search

Physically rearranging objects is an important capability for embodied agents. Visual room rearrangement evaluates an agent's ability to rearrange objects in a room to a desired goal based solely on visual input. We propose a simple yet effective method for this problem: (1) search for and map which objects need to be rearranged, and (2) rearrange each object until the task is complete. Our approach consists of an off-the-shelf semantic segmentation model, voxel-based semantic map, and semantic search policy to efficiently find objects that need to be rearranged. On the AI2-THOR Rearrangement Challenge, our method improves on current state-of-the-art end-to-end reinforcement learning-based methods that learn visual rearrangement policies from 0.53% correct rearrangement to 16.56%, using only 2.7% as many samples from the environment.

The Manifold Scattering Transform for High-Dimensional Point Cloud Data

The manifold scattering transform is a deep feature extractor for data defined on a Riemannian manifold. It is one of the first examples of extending convolutional neural network-like operators to general manifolds. The initial work on this model focused primarily on its theoretical stability and invariance properties but did not provide methods for its numerical implementation except in the case of two-dimensional surfaces with predefined meshes. In this work, we present practical schemes, based on the theory of diffusion maps, for implementing the manifold scattering transform to datasets arising in naturalistic systems, such as single cell genetics, where the data is a high-dimensional point cloud modeled as lying on a low-dimensional manifold. We show that our methods are effective for signal classification and manifold classification tasks.

Finding Optimal Policy for Queueing Models: New Parameterization

Queueing systems appear in many important real-life applications including communication networks, transportation and manufacturing systems. Reinforcement learning (RL) framework is a suitable model for the queueing control problem where the underlying dynamics are usually unknown and the agent receives little information from the environment to navigate. In this work, we investigate the optimization aspects of the queueing model as a RL environment and provide insight to learn the optimal policy efficiently. We propose a new parameterization of the policy by using the intrinsic properties of queueing network systems. Experiments show good performance of our methods with various load conditions from light to heavy traffic.

Benchmarking Node Outlier Detection on Graphs

Graph outlier detection is an emerging but crucial machine learning task with numerous applications. Despite the proliferation of algorithms developed in recent years, the lack of a standard and unified setting for performance evaluation limits their advancement and usage in real-world applications. To tap the gap, we present, (to our best knowledge) the first comprehensive unsupervised node outlier detection benchmark for graphs called UNOD, with the following highlights: (1) evaluating fourteen methods with backbone spanning from classical matrix factorization to the latest graph neural networks; (2) benchmarking the method performance with different types of injected outliers and organic outliers on real-world datasets; (3) comparing the efficiency and scalability of the algorithms by runtime and GPU memory usage on synthetic graphs at different scales. Based on the analyses of extensive experimental results, we discuss the pros and cons of current UNOD methods, and point out multiple crucial and promising future research directions.

Robust Deep Reinforcement Learning through Bootstrapped Opportunistic Curriculum

Despite considerable advances in deep reinforcement learning, it has been shown to be highly vulnerable to adversarial perturbations to state observations. Recent efforts that have attempted to improve adversarial robustness of reinforcement learning can nevertheless tolerate only very small perturbations, and remain fragile as perturbation size increases. We propose Bootstrapped Opportunistic Adversarial Curriculum Learning (BCL), a novel flexible adversarial curriculum learning framework for robust reinforcement learning. Our framework combines two ideas: conservatively bootstrapping each curriculum phase with highest quality solutions obtained from multiple runs of the previous phase, and opportunistically skipping forward in the curriculum. In our experiments we show that the proposed BCL framework enables dramatic improvements in robustness of learned policies to adversarial perturbations. The greatest improvement is for Pong, where our framework yields robustness to perturbations of up to 25/255; in contrast, the best existing approach can only tolerate adversarial noise up to 5/255. Our code is available at: https://github.com/jlwu002/BCL.

A Novel Three-Dimensional Navigation Method for the Visually Impaired

According to the World Health Organization, visual impairment is estimated to affect approximately 2.2 billion people worldwide. The visually impaired must currently rely on navigational aids to replace their sense of sight, like a white cane or GPS (Global Positioning System) based navigation, both of which fail to work well indoors. The white cane cannot be used to determine a user's position within a room, while GPS can often lose connection indoors and does not provide orientation information, making both approaches unsuitable for indoor use. Therefore, this research seeks to develop a 3D-imaging solution that enables contactless navigation through a complex indoor environment. The device can pinpoint a user's position and orientation with 31% less error compared to previous approaches while requiring only 53.1% of the memory, and processing 125% faster. The device can also detect obstacles with 60.2% more accuracy than the previous state-of-the-art models while requiring only 41% of the memory and processing 260% faster. When testing with human participants, the device allows for a 94.5% reduction in collisions with obstacles in the environment and allows for a 48.3% increase in walking speed, showing that my device enables safer and more rapid navigation for the visually impaired. All in all, this research demonstrates a 3D-based navigation system for the visually impaired. The approach can be used by a wide variety of mobile low-power devices, like cell phones, ensuring this research remains accessible to all.

Decentralized Distributed Learning with Privacy-Preserving Data Synthesis

In the medical field, multi-center collaborations are often sought to yield more generalizable findings by leveraging the heterogeneity of patient and clinical data. However, recent privacy regulations hinder the possibility to share data, and consequently, to come up with machine learning-based solutions that support diagnosis and prognosis. Federated learning (FL) aims at sidestepping this limitation by bringing AI-based solutions to data owners and only sharing local AI models, or parts thereof, that need then to be aggregated. However, most of the existing federated learning solutions are still at their infancy and show several shortcomings, from the lack of a reliable and effective aggregation scheme able to retain the knowledge learned locally to weak privacy preservation as real data may be reconstructed from model updates. Furthermore, the majority of these approaches, especially those dealing with medical data, relies on a centralized distributed learning strategy that poses robustness, scalability and trust issues. In this paper we present a decentralized distributed method that, exploiting concepts from experience replay and generative adversarial research, effectively integrates features from local nodes, providing models able to generalize across multiple datasets while maintaining privacy. The proposed approach is tested on two tasks - tuberculosis and melanoma classification - using multiple datasets in order to simulate realistic non-i.i.d. data scenarios. Results show that our approach achieves performance comparable to both standard (non-federated) learning and federated methods in their centralized (thus, more favourable) formulation.

Identifiability of deep generative models under mixture priors without auxiliary information

We prove identifiability of a broad class of deep latent variable models that (a) have universal approximation capabilities and (b) are the decoders of variational autoencoders that are commonly used in practice. Unlike existing work, our analysis does not require weak supervision, auxiliary information, or conditioning in the latent space. Recently, there has been a surge of works studying identifiability of such models. In these works, the main assumption is that along with the data, an auxiliary variable $u$ (also known as side information) is observed as well. At the same time, several works have empirically observed that this doesn't seem to be necessary in practice. In this work, we explain this behavior by showing that for a broad class of generative (i.e. unsupervised) models with universal approximation capabilities, the side information $u$ is not necessary: We prove identifiability of the entire generative model where we do not observe $u$ and only observe the data $x$. The models we consider are tightly connected with autoencoder architectures used in practice that leverage mixture priors in the latent space and ReLU/leaky-ReLU activations in the encoder. Our main result is an identifiability hierarchy that significantly generalizes previous work and exposes how different assumptions lead to different "strengths" of identifiability. For example, our weakest result establishes (unsupervised) identifiability up to an affine transformation, which already improves existing work. It's well known that these models have universal approximation capabilities and moreover, they have been extensively used in practice to learn representations of data.

Achieving Utility, Fairness, and Compactness via Tunable Information Bottleneck Measures

Designing machine learning algorithms that are accurate yet fair, not discriminating based on any sensitive attribute, is of paramount importance for society to accept AI for critical applications. In this article, we propose a novel fair representation learning method termed the R\'enyi Fair Information Bottleneck Method (RFIB) which incorporates constraints for utility, fairness, and compactness of representation, and apply it to image classification. A key attribute of our approach is that we consider - in contrast to most prior work - both demographic parity and equalized odds as fairness constraints, allowing for a more nuanced satisfaction of both criteria. Leveraging a variational approach, we show that our objectives yield a loss function involving classical Information Bottleneck (IB) measures and establish an upper bound in terms of the R\'enyi divergence of order $\alpha$ on the mutual information IB term measuring compactness between the input and its encoded embedding. Experimenting on three different image datasets (EyePACS, CelebA, and FairFace), we study the influence of the $\alpha$ parameter as well as two other tunable IB parameters on achieving utility/fairness trade-off goals, and show that the $\alpha$ parameter gives an additional degree of freedom that can be used to control the compactness of the representation. We evaluate the performance of our method using various utility, fairness, and compound utility/fairness metrics, showing that RFIB outperforms current state-of-the-art approaches.

Deep Learning Models on CPUs: A Methodology for Efficient Training

GPUs have been favored for training deep learning models due to their highly parallelized architecture. As a result, most studies on training optimization focus on GPUs. There is often a trade-off, however, between cost and efficiency when deciding on how to choose the proper hardware for training. In particular, CPU servers can be beneficial if training on CPUs was more efficient, as they incur fewer hardware update costs and better utilizing existing infrastructure. This paper makes several contributions to research on training deep learning models using CPUs. First, it presents a method for optimizing the training of deep learning models on Intel CPUs and a toolkit called ProfileDNN, which we developed to improve performance profiling. Second, we describe a generic training optimization method that guides our workflow and explores several case studies where we identified performance issues and then optimized the Intel Extension for PyTorch, resulting in an overall 2x training performance increase for the RetinaNet-ResNext50 model. Third, we show how to leverage the visualization capabilities of ProfileDNN, which enabled us to pinpoint bottlenecks and create a custom focal loss kernel that was two times faster than the official reference PyTorch implementation.

QuAFL: Federated Averaging Can Be Both Asynchronous and Communication-Efficient

Federated Learning (FL) is an emerging paradigm to enable the large-scale distributed training of machine learning models, while still providing privacy guarantees. In this work, we jointly address two of the main practical challenges when scaling federated optimization to large node counts: the need for tight synchronization between the central authority and individual computing nodes, and the large communication cost of transmissions between the central server and clients. Specifically, we present a new variant of the classic federated averaging (FedAvg) algorithm, which supports both asynchronous communication and communication compression. We provide a new analysis technique showing that, in spite of these system relaxations, our algorithm essentially matches the best known bounds for FedAvg, under reasonable parameter settings. On the experimental side, we show that our algorithm ensures fast practical convergence for standard federated tasks.

DNA: Proximal Policy Optimization with a Dual Network Architecture

This paper explores the problem of simultaneously learning a value function and policy in deep actor-critic reinforcement learning models. We find that the common practice of learning these functions jointly is sub-optimal, due to an order-of-magnitude difference in noise levels between these two tasks. Instead, we show that learning these tasks independently, but with a constrained distillation phase, significantly improves performance. Furthermore, we find that the policy gradient noise levels can be decreased by using a lower \textit{variance} return estimate. Whereas, the value learning noise level decreases with a lower \textit{bias} estimate. Together these insights inform an extension to Proximal Policy Optimization we call \textit{Dual Network Architecture} (DNA), which significantly outperforms its predecessor. DNA also exceeds the performance of the popular Rainbow DQN algorithm on four of the five environments tested, even under more difficult stochastic control settings.

Stochastic Online Learning with Feedback Graphs: Finite-Time and Asymptotic Optimality

We revisit the problem of stochastic online learning with feedback graphs, with the goal of devising algorithms that are optimal, up to constants, both asymptotically and in finite time. We show that, surprisingly, the notion of optimal finite-time regret is not a uniquely defined property in this context and that, in general, it is decoupled from the asymptotic rate. We discuss alternative choices and propose a notion of finite-time optimality that we argue is \emph{meaningful}. For that notion, we give an algorithm that admits quasi-optimal regret both in finite-time and asymptotically.

flow-based clustering and spectral clustering: a comparison

We propose and study a novel graph clustering method for data with an intrinsic network structure. Similar to spectral clustering, we exploit an intrinsic network structure of data to construct Euclidean feature vectors. These feature vectors can then be fed into basic clustering methods such as k-means or Gaussian mixture model (GMM) based soft clustering. What sets our approach apart from spectral clustering is that we do not use the eigenvectors of a graph Laplacian to construct the feature vectors. Instead, we use the solutions of total variation minimization problems to construct feature vectors that reflect connectivity between data points. Our motivation is that the solutions of total variation minimization are piece-wise constant around a given set of seed nodes. These seed nodes can be obtained from domain knowledge or by simple heuristics that are based on the network structure of data. Our results indicate that our clustering methods can cope with certain graph structures that are challenging for spectral clustering methods.

Deep Partial Least Squares for Empirical Asset Pricing

We use deep partial least squares (DPLS) to estimate an asset pricing model for individual stock returns that exploits conditioning information in a flexible and dynamic way while attributing excess returns to a small set of statistical risk factors. The novel contribution is to resolve the non-linear factor structure, thus advancing the current paradigm of deep learning in empirical asset pricing which uses linear stochastic discount factors under an assumption of Gaussian asset returns and factors. This non-linear factor structure is extracted by using projected least squares to jointly project firm characteristics and asset returns on to a subspace of latent factors and using deep learning to learn the non-linear map from the factor loadings to the asset returns. The result of capturing this non-linear risk factor structure is to characterize anomalies in asset returns by both linear risk factor exposure and interaction effects. Thus the well known ability of deep learning to capture outliers, shed lights on the role of convexity and higher order terms in the latent factor structure on the factor risk premia. On the empirical side, we implement our DPLS factor models and exhibit superior performance to LASSO and plain vanilla deep learning models. Furthermore, our network training times are significantly reduced due to the more parsimonious architecture of DPLS. Specifically, using 3290 assets in the Russell 1000 index over a period of December 1989 to January 2018, we assess our DPLS factor model and generate information ratios that are approximately 1.2x greater than deep learning. DPLS explains variation and pricing errors and identifies the most prominent latent factors and firm characteristics.

Measuring the Effect of Training Data on Deep Learning Predictions via Randomized Experiments

We develop a new, principled algorithm for estimating the contribution of training data points to the behavior of a deep learning model, such as a specific prediction it makes. Our algorithm estimates the AME, a quantity that measures the expected (average) marginal effect of adding a data point to a subset of the training data, sampled from a given distribution. When subsets are sampled from the uniform distribution, the AME reduces to the well-known Shapley value. Our approach is inspired by causal inference and randomized experiments: we sample different subsets of the training data to train multiple submodels, and evaluate each submodel's behavior. We then use a LASSO regression to jointly estimate the AME of each data point, based on the subset compositions. Under sparsity assumptions ($k \ll N$ datapoints have large AME), our estimator requires only $O(k\log N)$ randomized submodel trainings, improving upon the best prior Shapley value estimators.

Limitations of the NTK for Understanding Generalization in Deep Learning

The ``Neural Tangent Kernel'' (NTK) (Jacot et al 2018), and its empirical variants have been proposed as a proxy to capture certain behaviors of real neural networks. In this work, we study NTKs through the lens of scaling laws, and demonstrate that they fall short of explaining important aspects of neural network generalization. In particular, we demonstrate realistic settings where finite-width neural networks have significantly better data scaling exponents as compared to their corresponding empirical and infinite NTKs at initialization. This reveals a more fundamental difference between the real networks and NTKs, beyond just a few percentage points of test accuracy. Further, we show that even if the empirical NTK is allowed to be pre-trained on a constant number of samples, the kernel scaling does not catch up to the neural network scaling. Finally, we show that the empirical NTK continues to evolve throughout most of the training, in contrast with prior work which suggests that it stabilizes after a few epochs of training. Altogether, our work establishes concrete limitations of the NTK approach in understanding generalization of real networks on natural datasets.

When Does Re-initialization Work?

Re-initializing a neural network during training has been observed to improve generalization in recent works. Yet it is neither widely adopted in deep learning practice nor is it often used in state-of-the-art training protocols. This raises the question of when re-initialization works, and whether it should be used together with regularization techniques such as data augmentation, weight decay and learning rate schedules. In this work, we conduct an extensive empirical comparison of standard training with a selection of re-initialization methods to answer this question, training over 15,000 models on a variety of image classification benchmarks. We first establish that such methods are consistently beneficial for generalization in the absence of any other regularization. However, when deployed alongside other carefully tuned regularization techniques, re-initialization methods offer little to no added benefit for generalization, although optimal generalization performance becomes less sensitive to the choice of learning rate and weight decay hyperparameters. To investigate the impact of re-initialization methods on noisy data, we also consider learning under label noise. Surprisingly, in this case, re-initialization significantly improves upon standard training, even in the presence of other carefully tuned regularization techniques.

Open Set Classification of Untranscribed Handwritten Documents

Huge amounts of digital page images of important manuscripts are preserved in archives worldwide. The amounts are so large that it is generally unfeasible for archivists to adequately tag most of the documents with the required metadata so as to low proper organization of the archives and effective exploration by scholars and the general public. The class or ``typology'' of a document is perhaps the most important tag to be included in the metadata. The technical problem is one of automatic classification of documents, each consisting of a set of untranscribed handwritten text images, by the textual contents of the images. The approach considered is based on ``probabilistic indexing'', a relatively novel technology which allows to effectively represent the intrinsic word-level uncertainty exhibited by handwritten text images. We assess the performance of this approach on a large collection of complex notarial manuscripts from the Spanish Archivo Host\'orico Provincial de C\'adiz, with promising results.

Hyperparameter Importance of Quantum Neural Networks Across Small Datasets

As restricted quantum computers are slowly becoming a reality, the search for meaningful first applications intensifies. In this domain, one of the more investigated approaches is the use of a special type of quantum circuit - a so-called quantum neural network -- to serve as a basis for a machine learning model. Roughly speaking, as the name suggests, a quantum neural network can play a similar role to a neural network. However, specifically for applications in machine learning contexts, very little is known about suitable circuit architectures, or model hyperparameters one should use to achieve good learning performance. In this work, we apply the functional ANOVA framework to quantum neural networks to analyze which of the hyperparameters were most influential for their predictive performance. We analyze one of the most typically used quantum neural network architectures. We then apply this to $7$ open-source datasets from the OpenML-CC18 classification benchmark whose number of features is small enough to fit on quantum hardware with less than $20$ qubits. Three main levels of importance were detected from the ranking of hyperparameters obtained with functional ANOVA. Our experiment both confirmed expected patterns and revealed new insights. For instance, setting well the learning rate is deemed the most critical hyperparameter in terms of marginal contribution on all datasets, whereas the particular choice of entangling gates used is considered the least important except on one dataset. This work introduces new methodologies to study quantum machine learning models and provides new insights toward quantum model selection.

Model Optimization in Imbalanced Regression

Imbalanced domain learning aims to produce accurate models in predicting instances that, though underrepresented, are of utmost importance for the domain. Research in this field has been mainly focused on classification tasks. Comparatively, the number of studies carried out in the context of regression tasks is negligible. One of the main reasons for this is the lack of loss functions capable of focusing on minimizing the errors of extreme (rare) values. Recently, an evaluation metric was introduced: Squared Error Relevance Area (SERA). This metric posits a bigger emphasis on the errors committed at extreme values while also accounting for the performance in the overall target variable domain, thus preventing severe bias. However, its effectiveness as an optimization metric is unknown. In this paper, our goal is to study the impacts of using SERA as an optimization criterion in imbalanced regression tasks. Using gradient boosting algorithms as proof of concept, we perform an experimental study with 36 data sets of different domains and sizes. Results show that models that used SERA as an objective function are practically better than the models produced by their respective standard boosting algorithms at the prediction of extreme values. This confirms that SERA can be embedded as a loss function into optimization-based learning algorithms for imbalanced regression scenarios.

Measuring Class-Imbalance Sensitivity of Deterministic Performance Evaluation Metrics

The class-imbalance issue is intrinsic to many real-world machine learning tasks, particularly to the rare-event classification problems. Although the impact and treatment of imbalanced data is widely known, the magnitude of a metric's sensitivity to class imbalance has attracted little attention. As a result, often the sensitive metrics are dismissed while their sensitivity may only be marginal. In this paper, we introduce an intuitive evaluation framework that quantifies metrics' sensitivity to the class imbalance. Moreover, we reveal an interesting fact that there is a logarithmic behavior in metrics' sensitivity meaning that the higher imbalance ratios are associated with the lower sensitivity of metrics. Our framework builds an intuitive understanding of the class-imbalance impact on metrics. We believe this can help avoid many common mistakes, specially the less-emphasized and incorrect assumption that all metrics' quantities are comparable under different class-imbalance ratios.

Mitigating Data Heterogeneity in Federated Learning with Data Augmentation

Federated Learning (FL) is a prominent framework that enables training a centralized model while securing user privacy by fusing local, decentralized models. In this setting, one major obstacle is data heterogeneity, i.e., each client having non-identically and independently distributed (non-IID) data. This is analogous to the context of Domain Generalization (DG), where each client can be treated as a different domain. However, while many approaches in DG tackle data heterogeneity from the algorithmic perspective, recent evidence suggests that data augmentation can induce equal or greater performance. Motivated by this connection, we present federated versions of popular DG algorithms, and show that by applying appropriate data augmentation, we can mitigate data heterogeneity in the federated setting, and obtain higher accuracy on unseen clients. Equipped with data augmentation, we can achieve state-of-the-art performance using even the most basic Federated Averaging algorithm, with much sparser communication.

Thompson Sampling Efficiently Learns to Control Diffusion Processes

Diffusion processes that evolve according to linear stochastic differential equations are an important family of continuous-time dynamic decision-making models. Optimal policies are well-studied for them, under full certainty about the drift matrices. However, little is known about data-driven control of diffusion processes with uncertain drift matrices as conventional discrete-time analysis techniques are not applicable. In addition, while the task can be viewed as a reinforcement learning problem involving exploration and exploitation trade-off, ensuring system stability is a fundamental component of designing optimal policies. We establish that the popular Thompson sampling algorithm learns optimal actions fast, incurring only a square-root of time regret, and also stabilizes the system in a short time period. To the best of our knowledge, this is the first such result for Thompson sampling in a diffusion process control problem. We validate our theoretical results through empirical simulations with real parameter matrices from two settings of airplane and blood glucose control. Moreover, we observe that Thompson sampling significantly improves (worst-case) regret, compared to the state-of-the-art algorithms, suggesting Thompson sampling explores in a more guarded fashion. Our theoretical analysis involves characterization of a certain optimality manifold that ties the local geometry of the drift parameters to the optimal control of the diffusion process. We expect this technique to be of broader interest.

Noise Estimation in Gaussian Process Regression

We develop a computational procedure to estimate the covariance hyperparameters for semiparametric Gaussian process regression models with additive noise. Namely, the presented method can be used to efficiently estimate the variance of the correlated error, and the variance of the noise based on maximizing a marginal likelihood function. Our method involves suitably reducing the dimensionality of the hyperparameter space to simplify the estimation procedure to a univariate root-finding problem. Moreover, we derive bounds and asymptotes of the marginal likelihood function and its derivatives, which are useful to narrowing the initial range of the hyperparameter search. Using numerical examples, we demonstrate the computational advantages and robustness of the presented approach compared to traditional parameter optimization.

Critical Investigation of Failure Modes in Physics-informed Neural Networks

Several recent works in scientific machine learning have revived interest in the application of neural networks to partial differential equations (PDEs). A popular approach is to aggregate the residual form of the governing PDE and its boundary conditions as soft penalties into a composite objective/loss function for training neural networks, which is commonly referred to as physics-informed neural networks (PINNs). In the present study, we visualize the loss landscapes and distributions of learned parameters and explain the ways this particular formulation of the objective function may hinder or even prevent convergence when dealing with challenging target solutions. We construct a purely data-driven loss function composed of both the boundary loss and the domain loss. Using this data-driven loss function and, separately, a physics-informed loss function, we then train two neural network models with the same architecture. We show that incomparable scales between boundary and domain loss terms are the culprit behind the poor performance. Additionally, we assess the performance of both approaches on two elliptic problems with increasingly complex target solutions. Based on our analysis of their loss landscapes and learned parameter distributions, we observe that a physics-informed neural network with a composite objective function formulation produces highly non-convex loss surfaces that are difficult to optimize and are more prone to the problem of vanishing gradients.

Global Context Vision Transformers

We propose global context vision transformer (GC ViT), a novel architecture that enhances parameter and compute utilization. Our method leverages global context self-attention modules, joint with local self-attention, to effectively yet efficiently model both long and short-range spatial interactions, without the need for expensive operations such as computing attention masks or shifting local windows. In addition, we address the issue of lack of the inductive bias in ViTs via proposing to use a modified fused inverted residual blocks in our architecture. Our proposed GC ViT achieves state-of-the-art results across image classification, object detection and semantic segmentation tasks. On ImageNet-1K dataset for classification, the base, small and tiny variants of GC ViT with $28$M, $51$M and $90$M parameters achieve $\textbf{83.2\%}$, $\textbf{83.9\%}$ and $\textbf{84.4\%}$ Top-1 accuracy, respectively, surpassing comparably-sized prior art such as CNN-based ConvNeXt and ViT-based Swin Transformer by a large margin. Pre-trained GC ViT backbones in downstream tasks of object detection, instance segmentation, and semantic segmentation using MS COCO and ADE20K datasets outperform prior work consistently, sometimes by large margins. Code available at https://github.com/NVlabs/GCViT.

Quantum machine learning channel discrimination

In the problem of quantum channel discrimination, one distinguishes between a given number of quantum channels, which is done by sending an input state through a channel and measuring the output state. This work studies applications of variational quantum circuits and machine learning techniques for discriminating such channels. In particular, we explore (i) the practical implementation of embedding this task into the framework of variational quantum computing, (ii) training a quantum classifier based on variational quantum circuits, and (iii) applying the quantum kernel estimation technique. For testing these three channel discrimination approaches, we considered a pair of entanglement-breaking channels and the depolarizing channel with two different depolarization factors. For the approach (i), we address solving the quantum channel discrimination problem using widely discussed parallel and sequential strategies. We show the advantage of the latter in terms of better convergence with less quantum resources. Quantum channel discrimination with a variational quantum classifier (ii) allows one to operate even with random and mixed input states and simple variational circuits. The kernel-based classification approach (iii) is also found effective as it allows one to discriminate depolarizing channels associated not with just fixed values of the depolarization factor, but with ranges of it. Additionally, we discovered that a simple modification of one of the commonly used kernels significantly increases the efficiency of this approach. Finally, our numerical findings reveal that the performance of variational methods of channel discrimination depends on the trace of the product of the output states. These findings demonstrate that quantum machine learning can be used to discriminate channels, such as those representing physical noise processes.

Inference-Based Quantum Sensing

In a standard Quantum Sensing (QS) task one aims at estimating an unknown parameter $\theta$, encoded into an $n$-qubit probe state, via measurements of the system. The success of this task hinges on the ability to correlate changes in the parameter to changes in the system response $\mathcal{R}(\theta)$ (i.e., changes in the measurement outcomes). For simple cases the form of $\mathcal{R}(\theta)$ is known, but the same cannot be said for realistic scenarios, as no general closed-form expression exists. In this work we present an inference-based scheme for QS. We show that, for a general class of unitary families of encoding, $\mathcal{R}(\theta)$ can be fully characterized by only measuring the system response at $2n+1$ parameters. In turn, this allows us to infer the value of an unknown parameter given the measured response, as well as to determine the sensitivity of the sensing scheme, which characterizes its overall performance. We show that inference error is, with high probability, smaller than $\delta$, if one measures the system response with a number of shots that scales only as $\Omega(\log^3(n)/\delta^2)$. Furthermore, the framework presented can be broadly applied as it remains valid for arbitrary probe states and measurement schemes, and, even holds in the presence of quantum noise. We also discuss how to extend our results beyond unitary families. Finally, to showcase our method we implement it for a QS task on real quantum hardware, and in numerical simulations.

A Langevin-like Sampler for Discrete Distributions

We propose discrete Langevin proposal (DLP), a simple and scalable gradient-based proposal for sampling complex high-dimensional discrete distributions. In contrast to Gibbs sampling-based methods, DLP is able to update all coordinates in parallel in a single step and the magnitude of changes is controlled by a stepsize. This allows a cheap and efficient exploration in the space of high-dimensional and strongly correlated variables. We prove the efficiency of DLP by showing that the asymptotic bias of its stationary distribution is zero for log-quadratic distributions, and is small for distributions that are close to being log-quadratic. With DLP, we develop several variants of sampling algorithms, including unadjusted, Metropolis-adjusted, stochastic and preconditioned versions. DLP outperforms many popular alternatives on a wide variety of tasks, including Ising models, restricted Boltzmann machines, deep energy-based models, binary neural networks and language generation.

Low-Precision Stochastic Gradient Langevin Dynamics

While low-precision optimization has been widely used to accelerate deep learning, low-precision sampling remains largely unexplored. As a consequence, sampling is simply infeasible in many large-scale scenarios, despite providing remarkable benefits to generalization and uncertainty estimation for neural networks. In this paper, we provide the first study of low-precision Stochastic Gradient Langevin Dynamics (SGLD), showing that its costs can be significantly reduced without sacrificing performance, due to its intrinsic ability to handle system noise. We prove that the convergence of low-precision SGLD with full-precision gradient accumulators is less affected by the quantization error than its SGD counterpart in the strongly convex setting. To further enable low-precision gradient accumulators, we develop a new quantization function for SGLD that preserves the variance in each update step. We demonstrate that low-precision SGLD achieves comparable performance to full-precision SGLD with only 8 bits on a variety of deep learning tasks.

Only Tails Matter: Average-Case Universality and Robustness in the Convex Regime

The recently developed average-case analysis of optimization methods allows a more fine-grained and representative convergence analysis than usual worst-case results. In exchange, this analysis requires a more precise hypothesis over the data generating process, namely assuming knowledge of the expected spectral distribution (ESD) of the random matrix associated with the problem. This work shows that the concentration of eigenvalues near the edges of the ESD determines a problem's asymptotic average complexity. This a priori information on this concentration is a more grounded assumption than complete knowledge of the ESD. This approximate concentration is effectively a middle ground between the coarseness of the worst-case scenario convergence and the restrictive previous average-case analysis. We also introduce the Generalized Chebyshev method, asymptotically optimal under a hypothesis on this concentration and globally optimal when the ESD follows a Beta distribution. We compare its performance to classical optimization algorithms, such as gradient descent or Nesterov's scheme, and we show that, in the average-case context, Nesterov's method is universally nearly optimal asymptotically.

Multiple Fairness and Cardinality constraints for Students-Topics Grouping Problem

Group work is a prevalent activity in educational settings, where students are often divided into topic-specific groups based on their preferences. The grouping should reflect the students' aspirations as much as possible. Usually, the resulting groups should also be balanced in terms of protected attributes like gender or race since studies indicate that students might learn better in a diverse group. Moreover, balancing the group cardinalities is also an essential requirement for fair workload distribution across the groups. In this paper, we introduce the multi-fair capacitated (MFC) grouping problem that fairly partitions students into non-overlapping groups while ensuring balanced group cardinalities (with a lower bound and an upper bound), and maximizing the diversity of members in terms of protected attributes. We propose two approaches: a heuristic method and a knapsack-based method to obtain the MFC grouping. The experiments on a real dataset and a semi-synthetic dataset show that our proposed methods can satisfy students' preferences well and deliver balanced and diverse groups regarding cardinality and the protected attribute, respectively.

Latent Variable Modelling Using Variational Autoencoders: A survey

A probability distribution allows practitioners to uncover hidden structure in the data and build models to solve supervised learning problems using limited data. The focus of this report is on Variational autoencoders, a method to learn the probability distribution of large complex datasets. The report provides a theoretical understanding of variational autoencoders and consolidates the current research in the field. The report is divided into multiple chapters, the first chapter introduces the problem, describes variational autoencoders and identifies key research directions in the field. Chapters 2, 3, 4 and 5 dive into the details of each of the key research areas. Chapter 6 concludes the report and suggests directions for future work. A reader who has a basic idea of machine learning but wants to learn about general themes in machine learning research can benefit from the report. The report explains central ideas on learning probability distributions, what people did to make this tractable and goes into details around how deep learning is currently applied. The report also serves a gentle introduction for someone looking to contribute to this sub-field.

On the Impossibility of Learning to Cooperate with Adaptive Partner Strategies in Repeated Games

Learning to cooperate with other agents is challenging when those agents also possess the ability to adapt to our own behavior. Practical and theoretical approaches to learning in cooperative settings typically assume that other agents' behaviors are stationary, or else make very specific assumptions about other agents' learning processes. The goal of this work is to understand whether we can reliably learn to cooperate with other agents without such restrictive assumptions, which are unlikely to hold in real-world applications. Our main contribution is a set of impossibility results, which show that no learning algorithm can reliably learn to cooperate with all possible adaptive partners in a repeated matrix game, even if that partner is guaranteed to cooperate with some stationary strategy. Motivated by these results, we then discuss potential alternative assumptions which capture the idea that an adaptive partner will only adapt rationally to our behavior.

Nocturne: a scalable driving benchmark for bringing multi-agent learning one step closer to the real world

We introduce \textit{Nocturne}, a new 2D driving simulator for investigating multi-agent coordination under partial observability. The focus of Nocturne is to enable research into inference and theory of mind in real-world multi-agent settings without the computational overhead of computer vision and feature extraction from images. Agents in this simulator only observe an obstructed view of the scene, mimicking human visual sensing constraints. Unlike existing benchmarks that are bottlenecked by rendering human-like observations directly using a camera input, Nocturne uses efficient intersection methods to compute a vectorized set of visible features in a C++ back-end, allowing the simulator to run at $2000+$ steps-per-second. Using open-source trajectory and map data, we construct a simulator to load and replay arbitrary trajectories and scenes from real-world driving data. Using this environment, we benchmark reinforcement-learning and imitation-learning agents and demonstrate that the agents are quite far from human-level coordination ability and deviate significantly from the expert trajectories.

SoteriaFL: A Unified Framework for Private Federated Learning with Communication Compression

To enable large-scale machine learning in bandwidth-hungry environments such as wireless networks, significant progress has been made recently in designing communication-efficient federated learning algorithms with the aid of communication compression. On the other end, privacy-preserving, especially at the client level, is another important desideratum that has not been addressed simultaneously in the presence of advanced communication compression techniques yet. In this paper, we propose a unified framework that enhances the communication efficiency of private federated learning with communication compression. Exploiting both general compression operators and local differential privacy, we first examine a simple algorithm that applies compression directly to differentially-private stochastic gradient descent, and identify its limitations. We then propose a unified framework SoteriaFL for private federated learning, which accommodates a general family of local gradient estimators including popular stochastic variance-reduced gradient methods and the state-of-the-art shifted compression scheme. We provide a comprehensive characterization of its performance trade-offs in terms of privacy, utility, and communication complexity, where SoteraFL is shown to achieve better communication complexity without sacrificing privacy nor utility than other private federated learning algorithms without communication compression.

Deep Learning-Based Defect Classification and Detection in SEM Images

This proposes a novel ensemble deep learning-based model to accurately classify, detect and localize different defect categories for aggressive pitches and thin resists (High NA applications).In particular, we train RetinaNet models using different ResNet, VGGNet architectures as backbone and present the comparison between the accuracies of these models and their performance analysis on SEM images with different types of defect patterns such as bridge, break and line collapses. Finally, we propose a preference-based ensemble strategy to combine the output predictions from different models in order to achieve better performance on classification and detection of defects. As CDSEM images inherently contain a significant level of noise, detailed feature information is often shadowed by noise. For certain resist profiles, the challenge is also to differentiate between a microbridge, footing, break, and zones of probable breaks. Therefore, we have applied an unsupervised machine learning model to denoise the SEM images to remove the False-Positive defects and optimize the effect of stochastic noise on structured pixels for better metrology and enhanced defect inspection. We repeated the defect inspection step with the same trained model and performed a comparative analysis for "robustness" and "accuracy" metric with conventional approach for both noisy/denoised image pair. The proposed ensemble method demonstrates improvement of the average precision metric (mAP) of the most difficult defect classes. In this work we have developed a novel robust supervised deep learning training scheme to accurately classify as well as localize different defect types in SEM images with high degree of accuracy. Our proposed approach demonstrates its effectiveness both quantitatively and qualitatively.

Breaking Down Out-of-Distribution Detection: Many Methods Based on OOD Training Data Estimate a Combination of the Same Core Quantities

It is an important problem in trustworthy machine learning to recognize out-of-distribution (OOD) inputs which are inputs unrelated to the in-distribution task. Many out-of-distribution detection methods have been suggested in recent years. The goal of this paper is to recognize common objectives as well as to identify the implicit scoring functions of different OOD detection methods. We focus on the sub-class of methods that use surrogate OOD data during training in order to learn an OOD detection score that generalizes to new unseen out-distributions at test time. We show that binary discrimination between in- and (different) out-distributions is equivalent to several distinct formulations of the OOD detection problem. When trained in a shared fashion with a standard classifier, this binary discriminator reaches an OOD detection performance similar to that of Outlier Exposure. Moreover, we show that the confidence loss which is used by Outlier Exposure has an implicit scoring function which differs in a non-trivial fashion from the theoretically optimal scoring function in the case where training and test out-distribution are the same, which again is similar to the one used when training an Energy-Based OOD detector or when adding a background class. In practice, when trained in exactly the same way, all these methods perform similarly.

Algorithmic Fairness and Vertical Equity: Income Fairness with IRS Tax Audit Models

This study examines issues of algorithmic fairness in the context of systems that inform tax audit selection by the United States Internal Revenue Service (IRS). While the field of algorithmic fairness has developed primarily around notions of treating like individuals alike, we instead explore the concept of vertical equity -- appropriately accounting for relevant differences across individuals -- which is a central component of fairness in many public policy settings. Applied to the design of the U.S. individual income tax system, vertical equity relates to the fair allocation of tax and enforcement burdens across taxpayers of different income levels. Through a unique collaboration with the Treasury Department and IRS, we use access to anonymized individual taxpayer microdata, risk-selected audits, and random audits from 2010-14 to study vertical equity in tax administration. In particular, we assess how the use of modern machine learning methods for selecting audits may affect vertical equity. First, we show how the use of more flexible machine learning (classification) methods -- as opposed to simpler models -- shifts audit burdens from high to middle-income taxpayers. Second, we show that while existing algorithmic fairness techniques can mitigate some disparities across income, they can incur a steep cost to performance. Third, we show that the choice of whether to treat risk of underreporting as a classification or regression problem is highly consequential. Moving from classification to regression models to predict underreporting shifts audit burden substantially toward high income individuals, while increasing revenue. Last, we explore the role of differential audit cost in shaping the audit distribution. We show that a narrow focus on return-on-investment can undermine vertical equity. Our results have implications for the design of algorithmic tools across the public sector.

COVYT: Introducing the Coronavirus YouTube and TikTok speech dataset featuring the same speakers with and without infection

More than two years after its outbreak, the COVID-19 pandemic continues to plague medical systems around the world, putting a strain on scarce resources, and claiming human lives. From the very beginning, various AI-based COVID-19 detection and monitoring tools have been pursued in an attempt to stem the tide of infections through timely diagnosis. In particular, computer audition has been suggested as a non-invasive, cost-efficient, and eco-friendly alternative for detecting COVID-19 infections through vocal sounds. However, like all AI methods, also computer audition is heavily dependent on the quantity and quality of available data, and large-scale COVID-19 sound datasets are difficult to acquire -- amongst other reasons -- due to the sensitive nature of such data. To that end, we introduce the COVYT dataset -- a novel COVID-19 dataset collected from public sources containing more than 8 hours of speech from 65 speakers. As compared to other existing COVID-19 sound datasets, the unique feature of the COVYT dataset is that it comprises both COVID-19 positive and negative samples from all 65 speakers. We analyse the acoustic manifestation of COVID-19 on the basis of these perfectly speaker characteristic balanced `in-the-wild' data using interpretable audio descriptors, and investigate several classification scenarios that shed light into proper partitioning strategies for a fair speech-based COVID-19 detection.

Regression of high dimensional angular momentum states of light

The Orbital Angular Momentum (OAM) of light is an infinite-dimensional degree of freedom of light with several applications in both classical and quantum optics. However, to fully take advantage of the potential of OAM states, reliable detection platforms to characterize generated states in experimental conditions are needed. Here, we present an approach to reconstruct input OAM states from measurements of the spatial intensity distributions they produce. To obviate issues arising from intrinsic symmetry of Laguerre-Gauss modes, we employ a pair of intensity profiles per state projecting it only on two distinct bases, showing how this allows to uniquely recover input states from the collected data. Our approach is based on a combined application of dimensionality reduction via principal component analysis, and linear regression, and thus has a low computational cost during both training and testing stages. We showcase our approach in a real photonic setup, generating up-to-four-dimensional OAM states through a quantum walk dynamics. The high performances and versatility of the demonstrated approach make it an ideal tool to characterize high dimensional states in quantum information protocols.

A Neural Network Based Method with Transfer Learning for Genetic Data Analysis

Transfer learning has emerged as a powerful technique in many application problems, such as computer vision and natural language processing. However, this technique is largely ignored in application to genetic data analysis. In this paper, we combine transfer learning technique with a neural network based method(expectile neural networks). With transfer learning, instead of starting the learning process from scratch, we start from one task that have been learned when solving a different task. We leverage previous learnings and avoid starting from scratch to improve the model performance by passing information gained in different but related task. To demonstrate the performance, we run two real data sets. By using transfer learning algorithm, the performance of expectile neural networks is improved compared to expectile neural network without using transfer learning technique.

Automatic Autism Spectrum Disorder Detection Using Artificial Intelligence Methods with MRI Neuroimaging: A Review

Autism spectrum disorder (ASD) is a brain condition characterized by diverse signs and symptoms that appear in early childhood. ASD is also associated with communication deficits and repetitive behavior in affected individuals. Various ASD detection methods have been developed, including neuroimaging modalities and psychological tests. Among these methods, magnetic resonance imaging (MRI) imaging modalities are of paramount importance to physicians. Clinicians rely on MRI modalities to diagnose ASD accurately. The MRI modalities are non-invasive methods that include functional (fMRI) and structural (sMRI) neuroimaging methods. However, the process of diagnosing ASD with fMRI and sMRI for specialists is often laborious and time-consuming; therefore, several computer-aided design systems (CADS) based on artificial intelligence (AI) have been developed to assist the specialist physicians. Conventional machine learning (ML) and deep learning (DL) are the most popular schemes of AI used for diagnosing ASD. This study aims to review the automated detection of ASD using AI. We review several CADS that have been developed using ML techniques for the automated diagnosis of ASD using MRI modalities. There has been very limited work on the use of DL techniques to develop automated diagnostic models for ASD. A summary of the studies developed using DL is provided in the appendix. Then, the challenges encountered during the automated diagnosis of ASD using MRI and AI techniques are described in detail. Additionally, a graphical comparison of studies using ML and DL to diagnose ASD automatically is discussed. We conclude by suggesting future approaches to detecting ASDs using AI techniques and MRI neuroimaging.

Understanding Robust Learning through the Lens of Representation Similarities

Representation learning, i.e. the generation of representations useful for downstream applications, is a task of fundamental importance that underlies much of the success of deep neural networks (DNNs). Recently, robustness to adversarial examples has emerged as a desirable property for DNNs, spurring the development of robust training methods that account for adversarial examples. In this paper, we aim to understand how the properties of representations learned by robust training differ from those obtained from standard, non-robust training. This is critical to diagnosing numerous salient pitfalls in robust networks, such as, degradation of performance on benign inputs, poor generalization of robustness, and increase in over-fitting. We utilize a powerful set of tools known as representation similarity metrics, across three vision datasets, to obtain layer-wise comparisons between robust and non-robust DNNs with different architectures, training procedures and adversarial constraints. Our experiments highlight hitherto unseen properties of robust representations that we posit underlie the behavioral differences of robust networks. We discover a lack of specialization in robust networks' representations along with a disappearance of `block structure'. We also find overfitting during robust training largely impacts deeper layers. These, along with other findings, suggest ways forward for the design and training of better robust networks.

Additive Gaussian Processes Revisited

Gaussian Process (GP) models are a class of flexible non-parametric models that have rich representational power. By using a Gaussian process with additive structure, complex responses can be modelled whilst retaining interpretability. Previous work showed that additive Gaussian process models require high-dimensional interaction terms. We propose the orthogonal additive kernel (OAK), which imposes an orthogonality constraint on the additive functions, enabling an identifiable, low-dimensional representation of the functional relationship. We connect the OAK kernel to functional ANOVA decomposition, and show improved convergence rates for sparse computation methods. With only a small number of additive low-dimensional terms, we demonstrate the OAK model achieves similar or better predictive performance compared to black-box models, while retaining interpretability.

Contextual Squeeze-and-Excitation for Efficient Few-Shot Image Classification

Recent years have seen a growth in user-centric applications that require effective knowledge transfer across tasks in the low-data regime. An example is personalization, where a pretrained system is adapted by learning on small amounts of labeled data belonging to a specific user. This setting requires high accuracy under low computational complexity, therefore the Pareto frontier of accuracy vs.~adaptation cost plays a crucial role. In this paper we push this Pareto frontier in the few-shot image classification setting with two key contributions: (i) a new adaptive block called Contextual Squeeze-and-Excitation (CaSE) that adjusts a pretrained neural network on a new task to significantly improve performance with a single forward pass of the user data (context), and (ii) a hybrid training protocol based on Coordinate-Descent called UpperCaSE that exploits meta-trained CaSE blocks and fine-tuning routines for efficient adaptation. UpperCaSE achieves a new state-of-the-art accuracy relative to meta-learners on the 26 datasets of VTAB+MD and on a challenging real-world personalization benchmark (ORBIT), narrowing the gap with leading fine-tuning methods with the benefit of orders of magnitude lower adaptation cost.

Label noise (stochastic) gradient descent implicitly solves the Lasso for quadratic parametrisation

Understanding the implicit bias of training algorithms is of crucial importance in order to explain the success of overparametrised neural networks. In this paper, we study the role of the label noise in the training dynamics of a quadratically parametrised model through its continuous time version. We explicitly characterise the solution chosen by the stochastic flow and prove that it implicitly solves a Lasso program. To fully complete our analysis, we provide nonasymptotic convergence guarantees for the dynamics as well as conditions for support recovery. We also give experimental results which support our theoretical claims. Our findings highlight the fact that structured noise can induce better generalisation and help explain the greater performances of stochastic dynamics as observed in practice.

Business Document Information Extraction: Towards Practical Benchmarks

Information extraction from semi-structured documents is crucial for frictionless business-to-business (B2B) communication. While machine learning problems related to Document Information Extraction (IE) have been studied for decades, many common problem definitions and benchmarks do not reflect domain-specific aspects and practical needs for automating B2B document communication. We review the landscape of Document IE problems, datasets and benchmarks. We highlight the practical aspects missing in the common definitions and define the Key Information Localization and Extraction (KILE) and Line Item Recognition (LIR) problems. There is a lack of relevant datasets and benchmarks for Document IE on semi-structured business documents as their content is typically legally protected or sensitive. We discuss potential sources of available documents including synthetic data.

The Right Tool for the Job: Open-Source Auditing Tools in Machine Learning

In recent years, discussions about fairness in machine learning, AI ethics and algorithm audits have increased. Many entities have developed framework guidance to establish a baseline rubric for fairness and accountability. However, in spite of increased discussions and multiple frameworks, algorithm and data auditing still remain difficult to execute in practice. Many open-source auditing tools are available, but users aren't always aware of the tools, what they are useful for, or how to access them. Model auditing and evaluation are not frequently emphasized skills in machine learning. There are also legal reasons for the proactive adoption of these tools that extend beyond the desire for greater fairness in machine learning. There are positive social issues of public perception and goodwill that matter in our highly connected global society. Greater awareness of these tools and the reasons for actively utilizing them may be helpful to the entire continuum of programmers, data scientists, engineers, researchers, users and consumers of AI and machine learning products. It is important for everyone to better understand the input and output differentials, how they are occurring, and what can be done to promote FATE (fairness, accountability, transparency, and ethics) in machine- and deep learning. The ability to freely access open-source auditing tools removes barriers to fairness assessment at the most basic levels of machine learning. This paper aims to reinforce the urgent need to actually use these tools and provides motivations for doing so. The exemplary tools highlighted herein are open-source with software or code-base repositories available that can be used immediately by anyone worldwide.

CS$^2$: A Controllable and Simultaneous Synthesizer of Images and Annotations with Minimal Human Intervention

The destitution of image data and corresponding expert annotations limit the training capacities of AI diagnostic models and potentially inhibit their performance. To address such a problem of data and label scarcity, generative models have been developed to augment the training datasets. Previously proposed generative models usually require manually adjusted annotations (e.g., segmentation masks) or need pre-labeling. However, studies have found that these pre-labeling based methods can induce hallucinating artifacts, which might mislead the downstream clinical tasks, while manual adjustment could be onerous and subjective. To avoid manual adjustment and pre-labeling, we propose a novel controllable and simultaneous synthesizer (dubbed CS$^2$) in this study to generate both realistic images and corresponding annotations at the same time. Our CS$^2$ model is trained and validated using high resolution CT (HRCT) data collected from COVID-19 patients to realize an efficient infections segmentation with minimal human intervention. Our contributions include 1) a conditional image synthesis network that receives both style information from reference CT images and structural information from unsupervised segmentation masks, and 2) a corresponding segmentation mask synthesis network to automatically segment these synthesized images simultaneously. Our experimental studies on HRCT scans collected from COVID-19 patients demonstrate that our CS$^2$ model can lead to realistic synthesized datasets and promising segmentation results of COVID infections compared to the state-of-the-art nnUNet trained and fine-tuned in a fully supervised manner.

A Distributional Approach for Soft Clustering Comparison and Evaluation

The development of external evaluation criteria for soft clustering (SC) has received limited attention: existing methods do not provide a general approach to extend comparison measures to SC, and are unable to account for the uncertainty represented in the results of SC algorithms. In this article, we propose a general method to address these limitations, grounding on a novel interpretation of SC as distributions over hard clusterings, which we call \emph{distributional measures}. We provide an in-depth study of complexity- and metric-theoretic properties of the proposed approach, and we describe approximation techniques that can make the calculations tractable. Finally, we illustrate our approach through a simple but illustrative experiment.

Exceedance Probability Forecasting via Regression for Significant Wave Height Forecasting

Significant wave height forecasting is a key problem in ocean data analytics. Predicting the significant wave height is crucial for estimating the energy production from waves. Moreover, the timely prediction of large waves is important to ensure the safety of maritime operations, e.g. passage of vessels. We frame the task of predicting extreme values of significant wave height as an exceedance probability forecasting problem. Accordingly, we aim at estimating the probability that the significant wave height will exceed a predefined threshold. This task is usually solved using a probabilistic binary classification model. Instead, we propose a novel approach based on a forecasting model. The method leverages the forecasts for the upcoming observations to estimate the exceedance probability according to the cumulative distribution function. We carried out experiments using data from a buoy placed in the coast of Halifax, Canada. The results suggest that the proposed methodology is better than state-of-the-art approaches for exceedance probability forecasting.

SMT-DTA: Improving Drug-Target Affinity Prediction with Semi-supervised Multi-task Training

Drug-Target Affinity (DTA) prediction is an essential task for drug discovery and pharmaceutical research. Accurate predictions of DTA can greatly benefit the design of new drug. As wet experiments are costly and time consuming, the supervised data for DTA prediction is extremely limited. This seriously hinders the application of deep learning based methods, which require a large scale of supervised data. To address this challenge and improve the DTA prediction accuracy, we propose a framework with several simple yet effective strategies in this work: (1) a multi-task training strategy, which takes the DTA prediction and the masked language modeling (MLM) task on the paired drug-target dataset; (2) a semi-supervised training method to empower the drug and target representation learning by leveraging large-scale unpaired molecules and proteins in training, which differs from previous pre-training and fine-tuning methods that only utilize molecules or proteins in pre-training; and (3) a cross-attention module to enhance the interaction between drug and target representation. Extensive experiments are conducted on three real-world benchmark datasets: BindingDB, DAVIS and KIBA. The results show that our framework significantly outperforms existing methods and achieves state-of-the-art performances, e.g., $0.712$ RMSE on BindingDB IC$_{50}$ measurement with more than $5\%$ improvement than previous best work. In addition, case studies on specific drug-target binding activities, drug feature visualizations, and real-world applications demonstrate the great potential of our work. The code and data are released at https://github.com/QizhiPei/SMT-DTA

Convex space learning improves deep-generative oversampling for tabular imbalanced classification on smaller datasets

Data is commonly stored in tabular format. Several fields of research (e.g., biomedical, fault/fraud detection), are prone to small imbalanced tabular data. Supervised Machine Learning on such data is often difficult due to class imbalance, adding further to the challenge. Synthetic data generation i.e. oversampling is a common remedy used to improve classifier performance. State-of-the-art linear interpolation approaches, such as LoRAS and ProWRAS can be used to generate synthetic samples from the convex space of the minority class to improve classifier performance in such cases. Generative Adversarial Networks (GANs) are common deep learning approaches for synthetic sample generation. Although GANs are widely used for synthetic image generation, their scope on tabular data in the context of imbalanced classification is not adequately explored. In this article, we show that existing deep generative models perform poorly compared to linear interpolation approaches generating synthetic samples from the convex space of the minority class, for imbalanced classification problems on tabular datasets of small size. We propose a deep generative model, ConvGeN combining the idea of convex space learning and deep generative models. ConVGeN learns the coefficients for the convex combinations of the minority class samples, such that the synthetic data is distinct enough from the majority class. We demonstrate that our proposed model ConvGeN improves imbalanced classification on such small datasets, as compared to existing deep generative models while being at par with the existing linear interpolation approaches. Moreover, we discuss how our model can be used for synthetic tabular data generation in general, even outside the scope of data imbalance, and thus, improves the overall applicability of convex space learning.

Shapley-NAS: Discovering Operation Contribution for Neural Architecture Search

In this paper, we propose a Shapley value based method to evaluate operation contribution (Shapley-NAS) for neural architecture search. Differentiable architecture search (DARTS) acquires the optimal architectures by optimizing the architecture parameters with gradient descent, which significantly reduces the search cost. However, the magnitude of architecture parameters updated by gradient descent fails to reveal the actual operation importance to the task performance and therefore harms the effectiveness of obtained architectures. By contrast, we propose to evaluate the direct influence of operations on validation accuracy. To deal with the complex relationships between supernet components, we leverage Shapley value to quantify their marginal contributions by considering all possible combinations. Specifically, we iteratively optimize the supernet weights and update the architecture parameters by evaluating operation contributions via Shapley value, so that the optimal architectures are derived by selecting the operations that contribute significantly to the tasks. Since the exact computation of Shapley value is NP-hard, the Monte-Carlo sampling based algorithm with early truncation is employed for efficient approximation, and the momentum update mechanism is adopted to alleviate fluctuation of the sampling process. Extensive experiments on various datasets and various search spaces show that our Shapley-NAS outperforms the state-of-the-art methods by a considerable margin with light search cost. The code is available at https://github.com/Euphoria16/Shapley-NAS.git

Actively Learning Deep Neural Networks with Uncertainty Sampling Based on Sum-Product Networks

Active learning is popular approach for reducing the amount of data in training deep neural network model. Its success hinges on the choice of an effective acquisition function, which ranks not yet labeled data points according to their expected informativeness. In uncertainty sampling, the uncertainty that the current model has about a point's class label is the main criterion for this type of ranking. This paper proposes a new approach to uncertainty sampling in training a Convolutional Neural Network (CNN). The main idea is to use feature representation extracted extracted by the CNN as data for training a Sum-Product Network (SPN). Since SPNs are typically used for estimating the distribution of a dataset, they are well suited to the task of estimating class probabilities that can be used directly by standard acquisition functions such as max entropy and variational ratio. Moreover, we enhance these acquisition functions by weights calculated with the help of the SPN model; these weights make the acquisition function more sensitive to the diversity of conceivable class labels for data points. The effectiveness of our method is demonstrated in an experimental study on the MNIST, Fashion-MNIST and CIFAR-10 datasets, where we compare it to the state-of-the-art methods MC Dropout and Bayesian Batch.

Actively learning to learn causal relationships

How do people actively learn to learn? That is, how and when do people choose actions that facilitate long-term learning and choosing future actions that are more informative? We explore these questions in the domain of active causal learning. We propose a hierarchical Bayesian model that goes beyond past models by predicting that people pursue information not only about the causal relationship at hand but also about causal overhypotheses$\unicode{x2014}$abstract beliefs about causal relationships that span multiple situations and constrain how we learn the specifics in each situation. In two active "blicket detector" experiments with 14 between-subjects manipulations, our model was supported by both qualitative trends in participant behavior and an individual-differences-based model comparison. Our results suggest when there are abstract similarities across active causal learning problems, people readily learn and transfer overhypotheses about these similarities. Moreover, people exploit these overhypotheses to facilitate long-term active learning.

Quantitative CT texture-based method to predict diagnosis and prognosis of fibrosing interstitial lung disease patterns

Purpose: To utilize high-resolution quantitative CT (QCT) imaging features for prediction of diagnosis and prognosis in fibrosing interstitial lung diseases (ILD). Approach: 40 ILD patients (20 usual interstitial pneumonia (UIP), 20 non-UIP pattern ILD) were classified by expert consensus of 2 radiologists and followed for 7 years. Clinical variables were recorded. Following segmentation of the lung field, a total of 26 texture features were extracted using a lattice-based approach (TM model). The TM model was compared with previously histogram-based model (HM) for their abilities to classify UIP vs non-UIP. For prognostic assessment, survival analysis was performed comparing the expert diagnostic labels versus TM metrics. Results: In the classification analysis, the TM model outperformed the HM method with AUC of 0.70. While survival curves of UIP vs non-UIP expert labels in Cox regression analysis were not statistically different, TM QCT features allowed statistically significant partition of the cohort. Conclusions: TM model outperformed HM model in distinguishing UIP from non-UIP patterns. Most importantly, TM allows for partitioning of the cohort into distinct survival groups, whereas expert UIP vs non-UIP labeling does not. QCT TM models may improve diagnosis of ILD and offer more accurate prognostication, better guiding patient management.

Towards Perspective-Based Specification of Machine Learning-Enabled Systems

Machine learning (ML) teams often work on a project just to realize the performance of the model is not good enough. Indeed, the success of ML-enabled systems involves aligning data with business problems, translating them into ML tasks, experimenting with algorithms, evaluating models, capturing data from users, among others. Literature has shown that ML-enabled systems are rarely built based on precise specifications for such concerns, leading ML teams to become misaligned due to incorrect assumptions, which may affect the quality of such systems and overall project success. In order to help addressing this issue, this paper describes our work towards a perspective-based approach for specifying ML-enabled systems. The approach involves analyzing a set of 45 ML concerns grouped into five perspectives: objectives, user experience, infrastructure, model, and data. The main contribution of this paper is to provide two new artifacts that can be used to help specifying ML-enabled systems: (i) the perspective-based ML task and concern diagram and (ii) the perspective-based ML specification template.

Time Gated Convolutional Neural Networks for Crop Classification

This paper presented a state-of-the-art framework, Time Gated Convolutional Neural Network (TGCNN) that takes advantage of temporal information and gating mechanisms for the crop classification problem. Besides, several vegetation indices were constructed to expand dimensions of input data to take advantage of spectral information. Both spatial (channel-wise) and temporal (step-wise) correlation are considered in TGCNN. Specifically, our preliminary analysis indicates that step-wise information is of greater importance in this data set. Lastly, the gating mechanism helps capture high-order relationship. Our TGCNN solution achieves $0.973$ F1 score, $0.977$ AUC ROC and $0.948$ IoU, respectively. In addition, it outperforms three other benchmarks in different local tasks (Kenya, Brazil and Togo). Overall, our experiments demonstrate that TGCNN is advantageous in this earth observation time series classification task.

Square One Bias in NLP: Towards a Multi-Dimensional Exploration of the Research Manifold

The prototypical NLP experiment trains a standard architecture on labeled English data and optimizes for accuracy, without accounting for other dimensions such as fairness, interpretability, or computational efficiency. We show through a manual classification of recent NLP research papers that this is indeed the case and refer to it as the square one experimental setup. We observe that NLP research often goes beyond the square one setup, e.g, focusing not only on accuracy, but also on fairness or interpretability, but typically only along a single dimension. Most work targeting multilinguality, for example, considers only accuracy; most work on fairness or interpretability considers only English; and so on. We show this through manual classification of recent NLP research papers and ACL Test-of-Time award recipients. Such one-dimensionality of most research means we are only exploring a fraction of the NLP research search space. We provide historical and recent examples of how the square one bias has led researchers to draw false conclusions or make unwise choices, point to promising yet unexplored directions on the research manifold, and make practical recommendations to enable more multi-dimensional research. We open-source the results of our annotations to enable further analysis at https://github.com/google-research/url-nlp

A Comparative Study on Application of Class-Imbalance Learning for Severity Prediction of Adverse Events Following Immunization

In collaboration with the Liaoning CDC, China, we propose a prediction system to predict the subsequent hospitalization of children with adverse reactions based on data on adverse events following immunization. We extracted multiple features from the data, and selected "hospitalization or not" as the target for classification. Since the data are imbalanced, we used various class-imbalance learning methods for training and improved the RUSBoost algorithm. Experimental results show that the improved RUSBoost has the highest Area Under the ROC Curve on the target among these algorithms. Additionally, we compared these class-imbalance learning methods with some common machine learning algorithms. We combined the improved RUSBoost with dynamic web resource development techniques to build an evaluation system with information entry and vaccination response prediction capabilities for relevant medical practitioners.

Guided Safe Shooting: model based reinforcement learning with safety constraints

In the last decade, reinforcement learning successfully solved complex control tasks and decision-making problems, like the Go board game. Yet, there are few success stories when it comes to deploying those algorithms to real-world scenarios. One of the reasons is the lack of guarantees when dealing with and avoiding unsafe states, a fundamental requirement in critical control engineering systems. In this paper, we introduce Guided Safe Shooting (GuSS), a model-based RL approach that can learn to control systems with minimal violations of the safety constraints. The model is learned on the data collected during the operation of the system in an iterated batch fashion, and is then used to plan for the best action to perform at each time step. We propose three different safe planners, one based on a simple random shooting strategy and two based on MAP-Elites, a more advanced divergent-search algorithm. Experiments show that these planners help the learning agent avoid unsafe situations while maximally exploring the state space, a necessary aspect when learning an accurate model of the system. Furthermore, compared to model-free approaches, learning a model allows GuSS reducing the number of interactions with the real-system while still reaching high rewards, a fundamental requirement when handling engineering systems.

Metareview-informed Explainable Cytokine Storm Detection during CAR-T cell Therapy

Cytokine release syndrome (CRS), also known as cytokine storm, is one of the most consequential adverse effects of chimeric antigen receptor therapies that have shown promising results in cancer treatment. When emerging, CRS could be identified by the analysis of specific cytokine and chemokine profiles that tend to exhibit similarities across patients. In this paper, we exploit these similarities using machine learning algorithms and set out to pioneer a meta--review informed method for the identification of CRS based on specific cytokine peak concentrations and evidence from previous clinical studies. We argue that such methods could support clinicians in analyzing suspect cytokine profiles by matching them against CRS knowledge from past clinical studies, with the ultimate aim of swift CRS diagnosis. During evaluation with real--world CRS clinical data, we emphasize the potential of our proposed method of producing interpretable results, in addition to being effective in identifying the onset of cytokine storm.

Cross-Modal Transformer GAN: A Brain Structure-Function Deep Fusing Framework for Alzheimer's Disease

Cross-modal fusion of different types of neuroimaging data has shown great promise for predicting the progression of Alzheimer's Disease(AD). However, most existing methods applied in neuroimaging can not efficiently fuse the functional and structural information from multi-modal neuroimages. In this work, a novel cross-modal transformer generative adversarial network(CT-GAN) is proposed to fuse functional information contained in resting-state functional magnetic resonance imaging (rs-fMRI) and structural information contained in Diffusion Tensor Imaging (DTI). The developed bi-attention mechanism can match functional information to structural information efficiently and maximize the capability of extracting complementary information from rs-fMRI and DTI. By capturing the deep complementary information between structural features and functional features, the proposed CT-GAN can detect the AD-related brain connectivity, which could be used as a bio-marker of AD. Experimental results show that the proposed model can not only improve classification performance but also detect the AD-related brain connectivity effectively.

A Note on the Convergence of Mirrored Stein Variational Gradient Descent under $(L_0,L_1)-$Smoothness Condition

In this note, we establish a descent lemma for the population limit Mirrored Stein Variational Gradient Method~(MSVGD). This descent lemma does not rely on the path information of MSVGD but rather on a simple assumption for the mirrored distribution $\nabla\Psi_{\#}\pi\propto\exp(-V)$. Our analysis demonstrates that MSVGD can be applied to a broader class of constrained sampling problems with non-smooth $V$. We also investigate the complexity of the population limit MSVGD in terms of dimension $d$.

Great Expectations: Unsupervised Inference of Suspense, Surprise and Salience in Storytelling

Stories interest us not because they are a sequence of mundane and predictable events but because they have drama and tension. Crucial to creating dramatic and exciting stories are surprise and suspense. The thesis trains a series of deep learning models via only reading stories, a self-supervised (or unsupervised) system. Narrative theory methods (rules and procedures) are applied to the knowledge built into deep learning models to directly infer salience, surprise, and salience in stories. Extensions add memory and external knowledge from story plots and from Wikipedia to infer salience on novels such as Great Expectations and plays such as Macbeth. Other work adapts the models as a planning system for generating original stories. The thesis finds that applying the narrative theory to deep learning models can align with the typical reader. In follow-up work, the insights could help improve computer models for tasks such as automatic story writing and assistance for writing, summarising or editing stories. Moreover, the approach of applying narrative theory to the inherent qualities built in a system that learns itself (self-supervised) from reading from books, watching videos, and listening to audio is much cheaper and more adaptable to other domains and tasks. Progress is swift in improving self-supervised systems. As such, the thesis's relevance is that applying domain expertise with these systems may be a more productive approach for applying machine learning in many areas of interest.

The Role of Machine Learning in Cybersecurity

Machine Learning (ML) represents a pivotal technology for current and future information systems, and many domains already leverage the capabilities of ML. However, deployment of ML in cybersecurity is still at an early stage, revealing a significant discrepancy between research and practice. Such discrepancy has its root cause in the current state-of-the-art, which does not allow to identify the role of ML in cybersecurity. The full potential of ML will never be unleashed unless its pros and cons are understood by a broad audience. This paper is the first attempt to provide a holistic understanding of the role of ML in the entire cybersecurity domain -- to any potential reader with an interest in this topic. We highlight the advantages of ML with respect to human-driven detection methods, as well as the additional tasks that can be addressed by ML in cybersecurity. Moreover, we elucidate various intrinsic problems affecting real ML deployments in cybersecurity. Finally, we present how various stakeholders can contribute to future developments of ML in cybersecurity, which is essential for further progress in this field. Our contributions are complemented with two real case studies describing industrial applications of ML as defense against cyber-threats.

Technical Report: Combining knowledge from Transfer Learning during training and Wide Resnets

In this report, we combine the idea of Wide ResNets and transfer learning to optimize the architecture of deep neural networks. The first improvement of the architecture is the use of all layers as information source for the last layer. This idea comes from transfer learning, which uses networks pre-trained on other data and extracts different levels of the network as input for the new task. The second improvement is the use of deeper layers instead of deeper sequences of blocks. This idea comes from Wide ResNets. Using both optimizations, both high data augmentation and standard data augmentation can produce better results for different models. Link: https://github.com/wolfgangfuhl/PublicationStuff/tree/master/TechnicalReport1/Supp

GiDR-DUN; Gradient Dimensionality Reduction -- Differences and Unification

TSNE and UMAP are two of the most popular dimensionality reduction algorithms due to their speed and interpretable low-dimensional embeddings. However, while attempts have been made to improve on TSNE's computational complexity, no existing method can obtain TSNE embeddings at the speed of UMAP. In this work, we show that this is indeed possible by combining the two approaches into a single method. We theoretically and experimentally evaluate the full space of parameters in the TSNE and UMAP algorithms and observe that a single parameter, the normalization, is responsible for switching between them. This, in turn, implies that a majority of the algorithmic differences can be toggled without affecting the embeddings. We discuss the implications this has on several theoretic claims underpinning the UMAP framework, as well as how to reconcile them with existing TSNE interpretations. Based on our analysis, we propose a new dimensionality reduction algorithm, GDR, that combines previously incompatible techniques from TSNE and UMAP and can replicate the results of either algorithm by changing the normalization. As a further advantage, GDR performs the optimization faster than available UMAP methods and thus an order of magnitude faster than available TSNE methods. Our implementation is plug-and-play with the traditional UMAP and TSNE libraries and can be found at github.com/Andrew-Draganov/GiDR-DUN.

Remote Sensing Image Classification using Transfer Learning and Attention Based Deep Neural Network

The task of remote sensing image scene classification (RSISC), which aims at classifying remote sensing images into groups of semantic categories based on their contents, has taken the important role in a wide range of applications such as urban planning, natural hazards detection, environment monitoring,vegetation mapping, or geospatial object detection. During the past years, research community focusing on RSISC task has shown significant effort to publish diverse datasets as well as propose different approaches to deal with the RSISC challenges. Recently, almost proposed RSISC systems base on deep learning models which prove powerful and outperform traditional approaches using image processing and machine learning. In this paper, we also leverage the power of deep learning technology, evaluate a variety of deep neural network architectures, indicate main factors affecting the performance of a RSISC system. Given the comprehensive analysis, we propose a deep learning based framework for RSISC, which makes use of the transfer learning technique and multihead attention scheme. The proposed deep learning framework is evaluated on the benchmark NWPU-RESISC45 dataset and achieves the best classification accuracy of 94.7% which shows competitive to the state-of-the-art systems and potential for real-life applications.

GraphFramEx: Towards Systematic Evaluation of Explainability Methods for Graph Neural Networks

As one of the most popular machine learning models today, graph neural networks (GNNs) have attracted intense interest recently, and so does their explainability. Users are increasingly interested in a better understanding of GNN models and their outcomes. Unfortunately, today's evaluation frameworks for GNN explainability often rely on synthetic datasets, leading to conclusions of limited scope due to a lack of complexity in the problem instances. As GNN models are deployed to more mission-critical applications, we are in dire need for a common evaluation protocol of explainability methods of GNNs. In this paper, we propose, to our best knowledge, the first systematic evaluation framework for GNN explainability, considering explainability on three different "user needs:" explanation focus, mask nature, and mask transformation. We propose a unique metric that combines the fidelity measures and classify explanations based on their quality of being sufficient or necessary. We scope ourselves to node classification tasks and compare the most representative techniques in the field of input-level explainability for GNNs. For the widely used synthetic benchmarks, surprisingly shallow techniques such as personalized PageRank have the best performance for a minimum computation time. But when the graph structure is more complex and nodes have meaningful features, gradient-based methods, in particular Saliency, are the best according to our evaluation criteria. However, none dominates the others on all evaluation dimensions and there is always a trade-off. We further apply our evaluation protocol in a case study on eBay graphs to reflect the production environment.

Deep reinforced active learning for multi-class image classification

High accuracy medical image classification can be limited by the costs of acquiring more data as well as the time and expertise needed to label existing images. In this paper, we apply active learning to medical image classification, a method which aims to maximise model performance on a minimal subset from a larger pool of data. We present a new active learning framework, based on deep reinforcement learning, to learn an active learning query strategy to label images based on predictions from a convolutional neural network. Our framework modifies the deep-Q network formulation, allowing us to pick data based additionally on geometric arguments in the latent space of the classifier, allowing for high accuracy multi-class classification in a batch-based active learning setting, enabling the agent to label datapoints that are both diverse and about which it is most uncertain. We apply our framework to two medical imaging datasets and compare with standard query strategies as well as the most recent reinforcement learning based active learning approach for image classification.

EAGER: Asking and Answering Questions for Automatic Reward Shaping in Language-guided RL

Reinforcement learning (RL) in long horizon and sparse reward tasks is notoriously difficult and requires a lot of training steps. A standard solution to speed up the process is to leverage additional reward signals, shaping it to better guide the learning process. In the context of language-conditioned RL, the abstraction and generalisation properties of the language input provide opportunities for more efficient ways of shaping the reward. In this paper, we leverage this idea and propose an automated reward shaping method where the agent extracts auxiliary objectives from the general language goal. These auxiliary objectives use a question generation (QG) and question answering (QA) system: they consist of questions leading the agent to try to reconstruct partial information about the global goal using its own trajectory. When it succeeds, it receives an intrinsic reward proportional to its confidence in its answer. This incentivizes the agent to generate trajectories which unambiguously explain various aspects of the general language goal. Our experimental study shows that this approach, which does not require engineer intervention to design the auxiliary objectives, improves sample efficiency by effectively directing exploration.

Benchmarking Constraint Inference in Inverse Reinforcement Learning

When deploying Reinforcement Learning (RL) agents into a physical system, we must ensure that these agents are well aware of the underlying constraints. In many real-world problems, however, the constraints followed by expert agents (e.g., humans) are often hard to specify mathematically and unknown to the RL agents. To tackle these issues, Constraint Inverse Reinforcement Learning (CIRL) considers the formalism of Constrained Markov Decision Processes (CMDPs) and estimates constraints from expert demonstrations by learning a constraint function. As an emerging research topic, CIRL does not have common benchmarks, and previous works tested their algorithms with hand-crafted environments (e.g., grid worlds). In this paper, we construct a CIRL benchmark in the context of two major application domains: robot control and autonomous driving. We design relevant constraints for each environment and empirically study the ability of different algorithms to recover those constraints based on expert trajectories that respect those constraints. To handle stochastic dynamics, we propose a variational approach that infers constraint distributions, and we demonstrate its performance by comparing it with other CIRL baselines on our benchmark. The benchmark, including the information for reproducing the performance of CIRL algorithms, is publicly available at https://github.com/Guiliang/CIRL-benchmarks-public

What Can be Seen is What You Get: Structure Aware Point Cloud Augmentation

To train a well performing neural network for semantic segmentation, it is crucial to have a large dataset with available ground truth for the network to generalize on unseen data. In this paper we present novel point cloud augmentation methods to artificially diversify a dataset. Our sensor-centric methods keep the data structure consistent with the lidar sensor capabilities. Due to these new methods, we are able to enrich low-value data with high-value instances, as well as create entirely new scenes. We validate our methods on multiple neural networks with the public SemanticKITTI dataset and demonstrate that all networks improve compared to their respective baseline. In addition, we show that our methods enable the use of very small datasets, saving annotation time, training time and the associated costs.