Datasets:

feradauto

/

NLP4SGPapers

like 2

kameswara-sarma-2018-learning

https://aclanthology.org/N18-4007

Learning Word Embeddings for Data Sparse and Sentiment Rich Data Sets

This research proposal describes two algorithms that are aimed at learning word embeddings for data sparse and sentiment rich data sets. The goal is to use word embeddings adapted for domain specific data sets in downstream applications such as sentiment classification. The first approach learns word embeddings in a supervised fashion via SWESA (Supervised Word Embeddings for Sentiment Analysis), an algorithm for sentiment analysis on data sets that are of modest size. SWESA leverages document labels to jointly learn polarity-aware word embeddings and a classifier to classify unseen documents. In the second approach domain adapted (DA) word embeddings are learned by exploiting the specificity of domain specific data sets and the breadth of generic word embeddings. The new embeddings are formed by aligning corresponding word vectors using Canonical Correlation Analysis (CCA) or the related nonlinear Kernel CCA. Experimental results on binary sentiment classification tasks using both approaches for standard data sets are presented.

2018

silfverberg-etal-2017-data

https://aclanthology.org/K17-2010

Data Augmentation for Morphological Reinflection

This paper presents the submission of the Linguistics Department of the University of Colorado at Boulder for the 2017 CoNLL-SIGMORPHON Shared Task on Universal Morphological Reinflection. The system is implemented as an RNN Encoder-Decoder. It is specifically geared toward a low-resource setting. To this end, it employs data augmentation for counteracting overfitting and a copy symbol for processing characters unseen in the training data. The system is an ensemble of ten models combined using a weighted voting scheme. It delivers substantial improvement in accuracy compared to a non-neural baseline system in presence of varying amounts of training data.

The third author has been partly sponsored by DARPA I20 in the program Low Resource Languages for Emergent Incidents (LORELEI) issued by DARPA/I20 under Contract No. HR0011-15-C-0113.

2017

lehmann-1981-pragmalinguistics

https://aclanthology.org/J81-3004

Pragmalinguistics Theory and Practice

"Pragmalinguistics" or the occupation with pragmatic aspects of language can be important where computational linguists or artificial intelligence researchers are concerned with natural language interfaces to computers, with modelling dialogue behavior, or the like. What speakers intend with their utterances, how hearers react to what they hear, and what they take the words to mean will all play a role of increasing importance when natural language systems have matured enough to cope readily with syntax and semantics. Asking a sensible question to a user or giving him a reasonable response often enough depends not only on the "pure" meaning of some previous utterances but also on attitudes, expectations, and intentions that the user may have. These are partly conveyed in the user's utterances and have to be taken into account, if a system is to do more than just give factual answers to factual requests. Blakar writes on language as a means of social power. His paper is anecdotal; he draws conclusions without stating from what premises; and he is on the whole not very explicit. Gregersen postulates in his article on the relationships between social class and language usage that an economic analysis of "objective class positions" has to precede sociolinguistic studies proper, but fails to show how the results of such an analysis will influence sociolinguistics. Haeberlin writes on class-specific vocabulary as a communication problem. His ideas have been published before and in more detail.

1981

aggarwal-mamidi-2017-automatic

https://aclanthology.org/P17-3012

Automatic Generation of Jokes in Hindi

When it comes to computational language generation systems, humour is a relatively unexplored domain, especially more so for Hindi (or rather, for most languages other than English). Most researchers agree that a joke consists of two main parts-the setup and the punchline, with humour being encoded in the incongruity between the two. In this paper, we look at Dur se Dekha jokes, a restricted domain of humorous three liner poetry in Hindi. We analyze their structure to understand how humour is encoded in them and formalize it. We then develop a system which is successfully able to generate a basic form of these jokes.

The authors would like to thank all the evaluators for their time and help in rating the jokes. We would also like to thank Kaveri Anuranjana for all the time she spent helping us put this work on paper.

2017

rohanian-etal-2020-verbal

https://aclanthology.org/2020.acl-main.259

Verbal Multiword Expressions for Identification of Metaphor

Metaphor is a linguistic device in which a concept is expressed by mentioning another. Identifying metaphorical expressions, therefore, requires a non-compositional understanding of semantics. Multiword Expressions (MWEs), on the other hand, are linguistic phenomena with varying degrees of semantic opacity and their identification poses a challenge to computational models. This work is the first attempt at analysing the interplay of metaphor and MWEs processing through the design of a neural architecture whereby classification of metaphors is enhanced by informing the model of the presence of MWEs. To the best of our knowledge, this is the first "MWE-aware" metaphor identification system paving the way for further experiments on the complex interactions of these phenomena. The results and analyses show that this proposed architecture reach state-of-the-art on two different established metaphor datasets.

2020

gupta-etal-2021-disfl

https://aclanthology.org/2021.findings-acl.293

Disfl-QA: A Benchmark Dataset for Understanding Disfluencies in Question Answering

Disfluencies is an under-studied topic in NLP, even though it is ubiquitous in human conversation. This is largely due to the lack of datasets containing disfluencies. In this paper, we present a new challenge question answering dataset, DISFL-QA, a derivative of SQUAD, where humans introduce contextual disfluencies in previously fluent questions. DISFL-QA contains a variety of challenging disfluencies that require a more comprehensive understanding of the text than what was necessary in prior datasets. Experiments show that the performance of existing state-of-the-art question answering models degrades significantly when tested on DISFL-QA in a zero-shot setting. We show data augmentation methods partially recover the loss in performance and also demonstrate the efficacy of using gold data for fine-tuning. We argue that we need large-scale disfluency datasets in order for NLP models to be robust to them. The dataset is publicly available at: https://github.com/ google-research-datasets/disfl-qa.

Constructing datasets for spoken problems. We would also like to bring attention to the fact that being a speech phenomenon, a spoken setup would have been an ideal choice for disfluencies dataset. This would have accounted for higher degree of confusion, hesitations, corrections, etc. while recalling parts of context on the fly, which otherwise one may find hard to create synthetically when given enough time to think. However, such a spoken setup is extremely tedious for data collection mainly due to: (i) privacy concerns with acquiring speech data from real world speech transcriptions, (ii) creating scenarios for simulated environment is a challenging task, and (iii) relatively low yield for cases containing disfluencies. In such cases, we believe that a targeted and purely textual mode of data collection can be more effective both in terms of cost and specificity.

2021

mckeown-1983-paraphrasing

https://aclanthology.org/J83-1001

Paraphrasing Questions Using Given and new information

The design and implementation of a paraphrase component for a natural language question-answering system (CO-OP) is presented. The component is used to produce a paraphrase of a user's question to the system, which is presented to the user before the question is evaluated and answered. A major point made is the role of given and new information in formulating a paraphrase that differs in a meaningful way from the user's question. A description is also given of the transformational grammar that is used by the paraphraser. 2 For example, in the question "Which users work on projects sponsored by NASA?", the speaker makes the existential presupposition that there are projects sponsored by NASA.

This work was partially supported by an IBM fellowship and NSF grant MC78-08401. I would like to thank Dr. Aravind K. Joshi and Dr. Bonnie Webber for their invaluable comments on the style and content of this paper.

1983

chiu-etal-2022-joint

https://aclanthology.org/2022.spnlp-1.5

A Joint Learning Approach for Semi-supervised Neural Topic Modeling

Topic models are some of the most popular ways to represent textual data in an interpretable manner. Recently, advances in deep generative models, specifically auto-encoding variational Bayes (AEVB), have led to the introduction of unsupervised neural topic models, which leverage deep generative models as opposed to traditional statistics-based topic models. We extend upon these neural topic models by introducing the Label-Indexed Neural Topic Model (LI-NTM), which is, to the extent of our knowledge, the first effective upstream semisupervised neural topic model. We find that LI-NTM outperforms existing neural topic models in document reconstruction benchmarks, with the most notable results in low labeled data regimes and for data-sets with informative labels; furthermore, our jointly learned classifier outperforms baseline classifiers in ablation studies.

AS is supported by R01MH123804, and FDV is supported by NSF IIS-1750358. All authors acknowledge insightful feedback from members of CS282 Fall 2021.

2022

ws-1993-acquisition

https://aclanthology.org/W93-0100

Acquisition of Lexical Knowledge from Text

1993

vanni-miller-2002-scaling

http://www.lrec-conf.org/proceedings/lrec2002/pdf/306.pdf

Scaling the ISLE Framework: Use of Existing Corpus Resources for Validation of MT Evaluation Metrics across Languages

This paper describes a machine translation (MT) evaluation (MTE) research program which has benefited from the availability of two collections of source language texts and the results of processing these texts with several commercial MT engines (DARPA 1994, Doyon, Taylor, & White 1999). The methodology entails the systematic development of a predic tive relationship between discrete, well-defined MTE metrics and specific information processing tasks that can be reliably performed with output of a given MT system. Unlike tests used in initial experiments on automated scoring (Jones and Rusk 2000), we employ traditional measures of MT output quality, selected from the International Standards for Language Engineering (ISLE) framework: Coherence, Clarity, Syntax, Morphology, General and Domain-specific Lexical robustness, to include Named-entity translation. Each test was originally validated on MT output produced by three Spanish-to-English systems (1994 DARPA MTE). We validate tests in the present work, however, with material taken from the MT Scale Evaluation research program produced by Japanese-to-English MT systems. Since Spanish and Japanese differ structurally on the morphological, syntactic, and discourse levels, a comparison of scores on tests measuring these output qualities should reveal how structural similarity, such as that enjoyed by Spanish and English, and structural contrast, such as that found between Japanese and English, affect the linguistic distinctions which must be accommodated by MT systems. Moreover, we show that metrics developed using Spanish-English MT output are equally effective when applied to Japanese-English MT output.

2002

shimizu-etal-2013-constructing

https://aclanthology.org/2013.iwslt-papers.3

Constructing a speech translation system using simultaneous interpretation data

There has been a fair amount of work on automatic speech translation systems that translate in real-time, serving as a computerized version of a simultaneous interpreter. It has been noticed in the field of translation studies that simultaneous interpreters perform a number of tricks to make the content easier to understand in real-time, including dividing their translations into small chunks, or summarizing less important content. However, the majority of previous work has not specifically considered this fact, simply using translation data (made by translators) for learning of the machine translation system. In this paper, we examine the possibilities of additionally incorporating simultaneous interpretation data (made by simultaneous interpreters) in the learning process. First we collect simultaneous interpretation data from professional simultaneous interpreters of three levels, and perform an analysis of the data. Next, we incorporate the simultaneous interpretation data in the learning of the machine translation system. As a result, the translation style of the system becomes more similar to that of a highly experienced simultaneous interpreter. We also find that according to automatic evaluation metrics, our system achieves performance similar to that of a simultaneous interpreter that has 1 year of experience.

2013

uehara-thepkanjana-2014-called

https://aclanthology.org/Y14-1016

The So-called Person Restriction of Internal State Predicates in Japanese in Contrast with Thai

Internal state predicates or ISPs refer to internal states of sentient beings, such as emotions, sensations and thought processes. Japanese ISPs with zero pronouns exhibit the "person restriction" in that the zero form of their subjects must be first person at the utterance time. This paper examines the person restriction of ISPs in Japanese in contrast with those in Thai, which is a zero pronominal language like Japanese. It is found that the person restriction is applicable to Japanese ISPs but not to Thai ones. This paper argues that the person restriction is not adequate to account for Japanese and Thai ISPs. We propose a new constraint to account for this phenomenon, i.e., the Experiencer-Conceptualizer Identity (ECI) Constraint, which states that "The experiencer of the situation/event must be identical with the conceptualizer of that situation/event." It is argued that both languages conventionalize the ECI constraint in ISP expressions but differ in how the ECI constraint is conventionalized.

This research work is partially supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (No. 24520416) awarded to the first author and the Ratchadaphiseksomphot Endowment Fund of Chulalongkorn University (RES560530179-HS) awarded to the second author.

2014

chen-etal-2021-system

https://aclanthology.org/2021.autosimtrans-1.4

System Description on Automatic Simultaneous Translation Workshop

This paper shows our submission on the second automatic simultaneous translation workshop at NAACL2021. We participate in all the two directions of Chinese-to-English translation, Chinese audio→English text and Chinese text→English text. We do data filtering and model training techniques to get the best BLEU score and reduce the average lagging. We propose a two-stage simultaneous translation pipeline system which is composed of Quartznet and BPE-based transformer. We propose a competitive simultaneous translation system and achieves a BLEU score of 24.39 in the audio input track.

2021

oflazer-1995-error

https://aclanthology.org/1995.iwpt-1.24

Error-tolerant Finite State Recognition

Error-tolerant recognition enables the recognition of strings that deviate slightly fro� any string in the regular set recognized by the underlying finite state recognizer. In the context of natural language processing, it has applications in error-tolerant morphological analysis, and spe�g correction. After a description of the concepts and algorithms involved, we give examples from these two applications: In morphological analysis, error-tolerant recognition allows misspelled input word forms to� corrected, and morphologically analyzed concurrently. The algorithm can be applied to the moiphological analysis of any language whose morphology is fully captured by a single (and possibly very large) finite state transducer, regardless of the word formation processes (such as agglutination or productive compounding) and morphographemic phenomena involved. We present an .application to error tolerant analysis of agglutinative morphology of Turkish words. In spelling correction, error-tolerant recognition can be used to enumerate correct candidate forms from a given misspelled string within a certain edit distance. It can be applied to any language whose morphology is fully described by • a finite state transducer, or with a word list comprising all inflected forms with very large word lists of root• and inflected forms (some containing well over 200,000 forms), generating all candida� solutions within 10 to 45 milliseconds (with edit distance 1) on a SparcStation 10/41. For spelling correction in Turkish, error-tolerant recognition operating with a (circular) recognizer ofTurkish words (with about 29,000 states and 119,000 transitions) can generate all candidate words in less than 20 milliseconds (with edit distance 1). Spelling correction using a recognizer constructed from a large word German list that simulates compounding, also indicates that the approach is applicable in such cases.

This research was supported in part by a NATO Science for Stability Project Grant TU-LANGUAGE. I would like to thank Xerox Advanced Document Systems, and Lauri Karttunen of Xerox Pare and of Rank Xerox Research Centre (Grenoble) for providing us with the two-level transducer development software. Kemal Olku and Kurtulu� Yorulmaz of Bilkent University implemented some of the algorithms.

1995

mangeot-2014-motamot

http://www.lrec-conf.org/proceedings/lrec2014/pdf/128_Paper.pdf

Mot\`aMot project: conversion of a French-Khmer published dictionary for building a multilingual lexical system

Economic issues related to the information processing techniques are very important. The development of such technologies is a major asset for developing countries like Cambodia and Laos, and emerging ones like Vietnam, Malaysia and Thailand. The MotAMot project aims to computerize an under-resourced language: Khmer, spoken mainly in Cambodia. The main goal of the project is the development of a multilingual lexical system targeted for Khmer. The macrostructure is a pivot one with each word sense of each language linked to a pivot axi. The microstructure comes from a simplification of the explanatory and combinatory dictionary. The lexical system has been initialized with data coming mainly from the conversion of the French-Khmer bilingual dictionary of Denis Richer from Word to XML format. The French part was completed with pronunciation and parts-of-speech coming from the FeM French-english-Malay dictionary. The Khmer headwords noted in IPA in the Richer dictionary were converted to Khmer writing with OpenFST, a finite state transducer tool. The resulting resource is available online for lookup, editing, download and remote programming via a REST API on a Jibiki platform.

2014

francopoulo-etal-2016-providing

https://aclanthology.org/W16-4711

Providing and Analyzing NLP Terms for our Community

By its own nature, the Natural Language Processing (NLP) community is a priori the best equipped to study the evolution of its own publications, but works in this direction are rare and only recently have we seen a few attempts at charting the field. In this paper, we use the algorithms, resources, standards, tools and common practices of the NLP field to build a list of terms characteristic of ongoing research, by mining a large corpus of scientific publications, aiming at the largest possible exhaustivity and covering the largest possible time span. Study of the evolution of this term list through time reveals interesting insights on the dynamics of field and the availability of the term database and of the corpus (for a large part) make possible many further comparative studies in addition to providing a test field for a new graphic interface designed to perform visual time analytics of large sized thesauri.

2016

goldberg-elhadad-2007-svm

https://aclanthology.org/P07-1029

SVM Model Tampering and Anchored Learning: A Case Study in Hebrew NP Chunking

We study the issue of porting a known NLP method to a language with little existing NLP resources, specifically Hebrew SVM-based chunking. We introduce two SVM-based methods-Model Tampering and Anchored Learning. These allow fine grained analysis of the learned SVM models, which provides guidance to identify errors in the training corpus, distinguish the role and interaction of lexical features and eventually construct a model with ∼10% error reduction. The resulting chunker is shown to be robust in the presence of noise in the training corpus, relies on less lexical features than was previously understood and achieves an F-measure performance of 92.2 on automatically PoS-tagged text. The SVM analysis methods also provide general insight on SVM-based chunking.

2007

hendrix-1982-natural

https://aclanthology.org/J82-2002

Natural-Language Interface

A major problem faced by would-be users of computer systems is that computers generally make use of special-purpose languages familiar only to those trained in computer science. For a large number of applications requiring interaction between humans and computer systems, it would be highly desirable for machines to converse in English or other natural languages familiar to their human users.

1982

irvine-etal-2014-american

http://www.lrec-conf.org/proceedings/lrec2014/pdf/914_Paper.pdf

The American Local News Corpus

We present the American Local News Corpus (ALNC), containing over 4 billion words of text from 2, 652 online newspapers in the United States. Each article in the corpus is associated with a timestamp, state, and city. All 50 U.S. states and 1, 924 cities are represented. We detail our method for taking daily snapshots of thousands of local and national newspapers and present two example corpus analyses. The first explores how different sports are talked about over time and geography. The second compares per capita murder rates with news coverage of murders across the 50 states. The ALNC is about the same size as the Gigaword corpus and is growing continuously. Version 1.0 is available for research use.

We would like to thank the creators of the Newspaper Map website for providing us with their database of U.S. newspapers. This material is based on research sponsored by DARPA under contract HR0011-09-1-0044 and by the Johns Hopkins University Human Language Technology Center of Excellence. The views and conclusions contained in this publication are those of the authors and should not be interpreted as representing official policies or endorsements of DARPA or the U.S. Government.

2014

wissing-etal-2004-spoken

http://www.lrec-conf.org/proceedings/lrec2004/pdf/71.pdf

A Spoken Afrikaans Language Resource Designed for Research on Pronunciation Variations

In this contribution, the design, collection, annotation and planned distribution of a new spoken language resource of Afrikaans (SALAR) is discussed. The corpus contains speech of mother tongue speakers of Afrikaans, and is intended to become a primary national language resource for phonetic research and research on pronunciation variations. As such, the corpus is designed to expose pronunciation variations due to regional accents, speech rate (normal and fast speech) and speech mode (read and spontaneous speech). The corpus is collected by the Potchefstroom Campus of the NorthWest University, but in all phases of the corpus creation process there was a close collaboration with ELIS-UG (Belgium), one of the institutions that has been engaged in the creation of the Spoken Dutch Corpus (CGN).

2004

lu-paladini-adell-2012-beyond

https://aclanthology.org/2012.amta-commercial.10

Beyond MT: Source Content Quality and Process Automation

This document introduces the strategy implemented at CA Technologies to exploit Machine Translation (MT) at the corporate-wide level. We will introduce the different approaches followed to further improve the quality of the output of the machine translation engine once the engines have reached a maximum level of customization. Senior team support, clear communication between the parties involved and improvement measurement are the key components for the success of the initiative.

2012

liu-etal-2018-towards-less

https://aclanthology.org/D18-1297

Towards Less Generic Responses in Neural Conversation Models: A Statistical Re-weighting Method

Sequence-to-sequence neural generation models have achieved promising performance on short text conversation tasks. However, they tend to generate generic/dull responses, leading to unsatisfying dialogue experience. We observe that in conversation tasks, each query could have multiple responses, which forms a 1-ton or m-ton relationship in the view of the total corpus. The objective function used in standard sequence-to-sequence models will be dominated by loss terms with generic patterns. Inspired by this observation, we introduce a statistical re-weighting method that assigns different weights for the multiple responses of the same query, and trains the standard neural generation model with the weights. Experimental results on a large Chinese dialogue corpus show that our method improves the acceptance rate of generated responses compared with several baseline models and significantly reduces the number of generated generic responses.

2018

ide-suderman-2009-bridging

https://aclanthology.org/W09-3004

Bridging the Gaps: Interoperability for GrAF, GATE, and UIMA

This paper explores interoperability for data represented using the Graph Annotation Framework (GrAF) (Ide and Suderman, 2007) and the data formats utilized by two general-purpose annotation systems: the General Architecture for Text Engineering (GATE) (Cunningham, 2002) and the Unstructured Information Management Architecture (UIMA). GrAF is intended to serve as a "pivot" to enable interoperability among different formats, and both GATE and UIMA are at least implicitly designed with an eye toward interoperability with other formats and tools. We describe the steps required to perform a round-trip rendering from GrAF to GATE and GrAF to UIMA CAS and back again, and outline the commonalities as well as the differences and gaps that came to light in the process.

This work was supported by an IBM UIMA Innovation Award and National Science Foundation grant INT-0753069.

2009

nangia-bowman-2019-human

https://aclanthology.org/P19-1449

Human vs. Muppet: A Conservative Estimate of Human Performance on the GLUE Benchmark

The GLUE benchmark (Wang et al., 2019b) is a suite of language understanding tasks which has seen dramatic progress in the past year, with average performance moving from 70.0 at launch to 83.9, state of the art at the time of writing (May 24, 2019). Here, we measure human performance on the benchmark, in order to learn whether significant headroom remains for further progress. We provide a conservative estimate of human performance on the benchmark through crowdsourcing: Our annotators are non-experts who must learn each task from a brief set of instructions and 20 examples. In spite of limited training, these annotators robustly outperform the state of the art on six of the nine GLUE tasks and achieve an average score of 87.1. Given the fast pace of progress however, the headroom we observe is quite limited. To reproduce the datapoor setting that our annotators must learn in, we also train the BERT model (Devlin et al., 2019) in limited-data regimes, and conclude that low-resource sentence classification remains a challenge for modern neural network approaches to text understanding.

This work was made possible in part by a donation to NYU from Eric and Wendy Schmidt made by recommendation of the Schmidt Futures program and by funding from Samsung Research. We gratefully acknowledge the support of NVIDIA Corporation with the donation of a Titan V GPU used at NYU for this research. We thank Alex Wang and Amanpreet Singh for their help with conducting GLUE evaluations, and we thank Jason Phang for his help with training the BERT model.

2019

heyer-etal-1990-knowledge

https://aclanthology.org/C90-3073

Knowledge Representation and Semantics in a Complex Domain: The UNIX Natural Language Help System GOETHE

Natural language help systems for complex domains requirc, in our view, an integration of semantic representation and knowledge base in order to adequately and efficiently deal with cognitively misconceived user input ut. We present such an integration by way of the notiml of a frame-semae~tics that has been implemented for the purposes of a natural language help system for UNIX.

1990

zhou-etal-2022-hierarchical

https://aclanthology.org/2022.findings-acl.170

Hierarchical Recurrent Aggregative Generation for Few-Shot NLG

Large pretrained models enable transfer learning to low-resource domains for language generation tasks. However, previous end-to-end approaches do not account for the fact that some generation sub-tasks, specifically aggregation and lexicalisation, can benefit from transfer learning to different extents. To exploit these varying potentials for transfer learning, we propose a new hierarchical approach for few-shot and zero-shot generation. Our approach consists of a three-moduled jointly trained architecture: the first module independently lexicalises the distinct units of information in the input as sentence sub-units (e.g. phrases), the second module recurrently aggregates these sub-units to generate a unified intermediate output, while the third module subsequently post-edits it to generate a coherent and fluent final text. We perform extensive empirical analysis and ablation studies on fewshot and zero-shot settings across 4 datasets. Automatic and human evaluation shows that the proposed hierarchical approach is consistently capable of achieving state-of-the-art results when compared to previous work. 1

2022

chen-etal-2021-employing

https://aclanthology.org/2021.rocling-1.30

Employing low-pass filtered temporal speech features for the training of ideal ratio mask in speech enhancement

x [n] x m [n] m m m D × 1 M X = [ 0 1 ... M −1 ], D × M

2021

hempelmann-etal-2005-evaluating

https://aclanthology.org/W05-0211

Evaluating State-of-the-Art Treebank-style Parsers for Coh-Metrix and Other Learning Technology Environments

This paper evaluates a series of freely available, state-of-the-art parsers on a standard benchmark as well as with respect to a set of data relevant for measuring text cohesion. We outline advantages and disadvantages of existing technologies and make recommendations. Our performance report uses traditional measures based on a gold standard as well as novel dimensions for parsing evaluation. To our knowledge this is the first attempt to evaluate parsers accross genres and grade levels for the implementation in learning technology.

Quality Education

This research was funded by Institute for Educations Science Grant IES R3056020018-02. Any opinions, findings, and conclusions or recommendations expressed in this article are those of the authors and do not necessarily reflect the views of the IES. We are grateful to Philip M. McCarthy for his assistance in preparing some of our data.

2005

roark-etal-2009-deriving

https://aclanthology.org/D09-1034

Deriving lexical and syntactic expectation-based measures for psycholinguistic modeling via incremental top-down parsing

A number of recent publications have made use of the incremental output of stochastic parsers to derive measures of high utility for psycholinguistic modeling, following the work of Hale (2001; 2003; 2006). In this paper, we present novel methods for calculating separate lexical and syntactic surprisal measures from a single incremental parser using a lexicalized PCFG. We also present an approximation to entropy measures that would otherwise be intractable to calculate for a grammar of that size. Empirical results demonstrate the utility of our methods in predicting human reading times.

Thanks to Michael Collins, John Hale and Shravan Vasishth for valuable discussions about this work. This research was supported in part by NSF Grant #BCS-0826654. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the NSF.

2009

su-etal-1989-smoothing

https://aclanthology.org/O89-1010

Smoothing Statistic Databases in a Machine Translation System

1989

raake-2002-content

http://www.lrec-conf.org/proceedings/lrec2002/pdf/107.pdf

Does the Content of Speech Influence its Perceived Sound Quality?

From a user's perspective, the speech quality of modern telecommunication systems often differs from that of traditional wireline telephone systems. One aspect is a changed sound of the interlocutor's voice-introduced by an expansion of the transmissionbandwidth to wide-band, by low-bitrate coding and/or by the acoustic properties of specific user-interfaces. In order to quantify the effect of transmission on speech quality, subjective data to be correlated to transmission characteristics have to be collected in auditory tests. In this paper, a study is presented investigating in how far the content of specific speech material used in a listening-only test impacts its perceived sound quality. A set of French speech data was presented to two different groups of listeners: French native speakers and listeners without knowledge of French. The speech material consists of different text types, such as everyday speech or semantically unpredictable sentences (SUS). The listeners were asked to rate the sound quality of the transmitted voice on a onedimensional category rating scale. The French listeners' ratings were found to be lower for SUS, while those of the non-French listeners did not show any major dependency on text material. Hence, it can be stated that if a given speech sign is understood by the listeners, they are unable to separate form from function and reflect content in their ratings of sound.

This work has been carried out at the Institute of Communication Acoustics, Ruhr-University Bochum (Prof. J. Blauert, PD. U. Jekosch). It was performed in the framework of a PROCOPE co-operation with the LMA, CNRS, Marseille, France (Dr. G. Canévet). The author would like to thank U. Jekosch, S. Möller and S. Schaden for fruitful discussions and S. Meunier (CNRS Marseille) for her help in auditory test organization at the LMA.

2002

duan-etal-2012-twitter

https://aclanthology.org/C12-1047

Twitter Topic Summarization by Ranking Tweets using Social Influence and Content Quality

In this paper, we propose a time-line based framework for topic summarization in Twitter. We summarize topics by sub-topics along time line to fully capture rapid topic evolution in Twitter. Specifically, we rank and select salient and diversified tweets as a summary of each sub-topic. We have observed that ranking tweets is significantly different from ranking sentences in traditional extractive document summarization. We model and formulate the tweet ranking in a unified mutual reinforcement graph, where the social influence of users and the content quality of tweets are taken into consideration simultaneously in a mutually reinforcing manner. Extensive experiments are conducted on 3.9 million tweets. The results show that the proposed approach outperforms previous approaches by 14% improvement on average ROUGE-1. Moreover, we show how the content quality of tweets and the social influence of users effectively improve the performance of measuring the salience of tweets.

Peace, Justice and Strong Institutions

2012

goodman-1985-repairing

https://aclanthology.org/P85-1026

Repairing Reference Identification Failures by Relaxation

The goal of thls work is the enrichment of human-machlne mteractIons in a natural language

1985

bruce-wiebe-1998-word

https://aclanthology.org/W98-1507

Word-Sense Distinguishability and Inter-Coder Agreement

It. is common in NLP that the categories into which text is classified do not have fully objective definitions. Examples of such categories are lexical distinctions such as part-of-speech tags and wordsense distinctions, sentence level distinctions such as phrase attachment, and discourse level distinct.icms such as topic or speech-act categorization. This p>1per presents an approach to analy?-ing the agrcen1ent arnong lnnnan judges for the purpose of formulating a refined and more reliable set of category designations. We use these techniques to analyze the sense tags assigned by five judgps to the noun intcr•est. The initial tag set is takmi from Longman's Dictionary of Contemporary i:nglish. Through this process of analysis, we automatically identify and assign a revised set of sense tags for the data. The revised tags exhibit high reliability as measured by Cohen's r;.. Such techniques are important for formulating and evaluating both human and automated classification systems.

1998

lauscher-etal-2018-investigating

https://aclanthology.org/D18-1370

Investigating the Role of Argumentation in the Rhetorical Analysis of Scientific Publications with Neural Multi-Task Learning Models

Exponential growth in the number of scientific publications yields the need for effective automatic analysis of rhetorical aspects of scientific writing. Acknowledging the argumentative nature of scientific text, in this work we investigate the link between the argumentative structure of scientific publications and rhetorical aspects such as discourse categories or citation contexts. To this end, we (1) augment a corpus of scientific publications annotated with four layers of rhetoric annotations with argumentation annotations and (2) investigate neural multi-task learning architectures combining argument extraction with a set of rhetorical classification tasks. By coupling rhetorical classifiers with the extraction of argumentative components in a joint multi-task learning setting, we obtain significant performance gains for different rhetorical analysis tasks.

This research was partly funded by the German Research Foundation (DFG), grant number EC 477/5-1 (LOC-DB). We thank our four annotators for their dedicated annotation effort and the anonymous reviewers for constructive and insightful comments.

2018

martschat-strube-2015-latent

https://aclanthology.org/Q15-1029

Latent Structures for Coreference Resolution

Machine learning approaches to coreference resolution vary greatly in the modeling of the problem: while early approaches operated on the mention pair level, current research focuses on ranking architectures and antecedent trees. We propose a unified representation of different approaches to coreference resolution in terms of the structure they operate on. We represent several coreference resolution approaches proposed in the literature in our framework and evaluate their performance. Finally, we conduct a systematic analysis of the output of these approaches, highlighting differences and similarities.

This work has been funded by the Klaus Tschira Foundation, Heidelberg, Germany. The first author has been supported by a HITS PhD scholarship. We thank the anonymous reviewers and our colleagues Benjamin Heinzerling, Yufang Hou and Nafise Moosavi for feedback on earlier drafts of this paper. Furthermore, we are grateful to Anders Björkelund for helpful comments on cost functions.

2015

muller-2020-pymmax2

https://aclanthology.org/2020.law-1.16

pyMMAX2: Deep Access to MMAX2 Projects from Python

pyMMAX2 is an API for processing MMAX2 stand-off annotation data in Python. It provides a lightweight basis for the development of code which opens up the Java-and XML-based ecosystem of MMAX2 for more recent, Python-based NLP and data science methods. While pyMMAX2 is pure Python, and most functionality is implemented from scratch, the API re-uses the complex implementation of the essential business logic for MMAX2 annotation schemes by interfacing with the original MMAX2 Java libraries. pyMMAX2 is available for download at http://github.com/nlpAThits/pyMMAX2.

The work described in this paper was done as part of the project DeepCurate, which is funded by the German Federal Ministry of Education and Research (BMBF) (No. 031L0204) and the Klaus Tschira Foundation, Heidelberg, Germany. We thank the anonymous reviewers for their helpful suggestions.

2020

bhattasali-etal-2018-processing

https://aclanthology.org/W18-4904

Processing MWEs: Neurocognitive Bases of Verbal MWEs and Lexical Cohesiveness within MWEs

Multiword expressions have posed a challenge in the past for computational linguistics since they comprise a heterogeneous family of word clusters and are difficult to detect in natural language data. In this paper, we present a fMRI study based on language comprehension to provide neuroimaging evidence for processing MWEs. We investigate whether different MWEs have distinct neural bases, e.g. if verbal MWEs involve separate brain areas from non-verbal MWEs and if MWEs with varying levels of cohesiveness activate dissociable brain regions. Our study contributes neuroimaging evidence illustrating that different MWEs elicit spatially distinct patterns of activation. We also adapt an association measure, usually used to detect MWEs, as a cognitively plausible metric for language processing.

This material is based upon work supported by the National Science Foundation under Grant No. 1607441.

2018

farrow-dzikovska-2009-context

https://aclanthology.org/W09-1502

Context-Dependent Regression Testing for Natural Language Processing

Regression testing of natural language systems is problematic for two main reasons: component input and output is complex, and system behaviour is context-dependent. We have developed a generic approach which solves both of these issues. We describe our regression tool, CONTEST, which supports context-dependent testing of dialogue system components, and discuss the regression test sets we developed, designed to effectively isolate components from changes and problems earlier in the pipeline. We believe that the same approach can be used in regression testing for other dialogue systems, as well as in testing any complex NLP system containing multiple components.

This work has been supported in part by Office of Naval Research grant N000140810043. We thank Charles Callaway for help with generation and tutoring tests.

2009

manuvinakurike-etal-2018-conversational

https://aclanthology.org/W18-5033

Conversational Image Editing: Incremental Intent Identification in a New Dialogue Task

We present "conversational image editing", a novel real-world application domain combining dialogue, visual information, and the use of computer vision. We discuss the importance of dialogue incrementality in this task, and build various models for incremental intent identification based on deep learning and traditional classification algorithms. We show how our model based on convolutional neural networks outperforms models based on random forests, long short term memory networks, and conditional random fields. By training embeddings based on image-related dialogue corpora, we outperform pre-trained out-of-the-box embeddings, for intention identification tasks. Our experiments also provide evidence that incremental intent processing may be more efficient for the user and could save time in accomplishing tasks.

This work was supported by a generous gift of Adobe Systems Incorporated to USC/ICT, and the first author's internship at Adobe Research. The first and last authors were also funded by the U.S. Army Research Laboratory. Statements and opinions expressed do not necessarily reflect the position or policy of the U.S. Government, and no official endorsement should be inferred.

2018

roth-2017-role

https://aclanthology.org/W17-6934

Role Semantics for Better Models of Implicit Discourse Relations

Predicting the structure of a discourse is challenging because relations between discourse segments are often implicit and thus hard to distinguish computationally. I extend previous work to classify implicit discourse relations by introducing a novel set of features on the level of semantic roles. My results demonstrate that such features are helpful, yielding results competitive with other feature-rich approaches on the PDTB. My main contribution is an analysis of improvements that can be traced back to role-based features, providing insights into why and when role semantics is helpful.

This research was supported in part by the Cluster of Excellence "Multimodal Computing and Interaction" of the German Excellence Initiative, and a DFG Research Fellowship (RO 4848/1-1).

2017

yamaguchi-etal-2021-frustratingly

https://aclanthology.org/2021.emnlp-main.249

Frustratingly Simple Pretraining Alternatives to Masked Language Modeling

Masked language modeling (MLM), a selfsupervised pretraining objective, is widely used in natural language processing for learning text representations. MLM trains a model to predict a random sample of input tokens that have been replaced by a [MASK] placeholder in a multi-class setting over the entire vocabulary. When pretraining, it is common to use alongside MLM other auxiliary objectives on the token or sequence level to improve downstream performance (e.g. next sentence prediction). However, no previous work so far has attempted in examining whether other simpler linguistically intuitive or not objectives can be used standalone as main pretraining objectives. In this paper, we explore five simple pretraining objectives based on token-level classification tasks as replacements of MLM. Empirical results on GLUE and SQUAD show that our proposed methods achieve comparable or better performance to MLM using a BERT-BASE architecture. We further validate our methods using smaller models, showing that pretraining a model with 41% of the BERT-BASE's parameters, BERT-MEDIUM results in only a 1% drop in GLUE scores with our best objective. 1 * Work was done while at the University of Sheffield.

NA is supported by EPSRC grant EP/V055712/1, part of the European Commission CHIST-ERA programme, call 2019 XAI: Explainable Machine Learning-based Artificial Intelligence. KM is supported by Amazon through the Alexa Fellowship scheme.

2021

herzig-etal-2016-classifying

https://aclanthology.org/W16-3609

Classifying Emotions in Customer Support Dialogues in Social Media

Providing customer support through social media channels is gaining increasing popularity. In such a context, automatic detection and analysis of the emotions expressed by customers is important, as is identification of the emotional techniques (e.g., apology, empathy, etc.) in the responses of customer service agents. Result of such an analysis can help assess the quality of such a service, help and inform agents about desirable responses, and help develop automated service agents for social media interactions. In this paper, we show that, in addition to text based turn features, dialogue features can significantly improve detection of emotions in social media customer service dialogues and help predict emotional techniques used by customer service agents. Got excited to pick up the latest bundle since it was on sale today, but now I can't download it at all. Bummer. =/ Yeah, no problems there. The error is coming when I actually try to download the games. Error code: 412344 Uh oh! To check, were you able to purchase that title? Let's confirm by signing in at http://t.co/53fsdfd real quick. Appreciate that! Let's power cycle and unplug modem/router for 2 mins then try again. Seems to be working now. Weird. I tried that 3 different times earlier. Thanks. Odd, but glad to hear that's sorted! Happy gaming, and we'll be here to help if any other questions or concerns arise.

2016

pighin-etal-2012-analysis

http://www.lrec-conf.org/proceedings/lrec2012/pdf/337_Paper.pdf

An Analysis (and an Annotated Corpus) of User Responses to Machine Translation Output

We present an annotated resource consisting of open-domain translation requests, automatic translations and user-provided corrections collected from casual users of the translation portal http://reverso.net. The layers of annotation provide: 1) quality assessments for 830 correction suggestions for translations into English, at the segment level, and 2) 814 usefulness assessments for English-Spanish and English-French translation suggestions, a suggestion being useful if it contains at least local clues that can be used to improve translation quality. We also discuss the results of our preliminary experiments concerning 1) the development of an automatic filter to separate useful from non-useful feedback, and 2) the incorporation in the machine translation pipeline of bilingual phrases extracted from the suggestions. The annotated data, available for download from ftp://mi.eng.cam.ac.uk/data/faust/LW-UPC-Oct11-FAUST-feedback-annotation.tgz, is released under a Creative Commons license. To our best knowledge, this is the first resource of this kind that has ever been made publicly available.

This research has been partially funded by the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement numbers 247762 (FAUST project, FP7-ICT-2009-4-247762) and 247914 (MOLTO project, FP7-ICT-2009-4-247914) and by the Spanish Ministry of Education and Science (OpenMT-2, TIN2009-14675-C03).

2012

yonezaki-enomoto-1980-database

https://aclanthology.org/C80-1032

Database System Based on Intensional Logic

Model theoretic semantics of database systems is studied. As Rechard Montague has done in his work, 5 we translate statements of DDL and DML into intensional logic and the latter is interpreted with reference to a suitable model. Major advantages of its approach include (i) it leads itself to the design of database systems which can handle historical data, (ii) it provides with a formal description of database semantics.

Our thanks are due to Mr. Kenichi Murata for fruitful discussions and encouragement and to Prof. Takuya Katayama and many other people whose ideas we have unwittingly absorbed over the years.

1980

wing-baldridge-2011-simple

https://aclanthology.org/P11-1096

Simple supervised document geolocation with geodesic grids

We investigate automatic geolocation (i.e. identification of the location, expressed as latitude/longitude coordinates) of documents. Geolocation can be an effective means of summarizing large document collections and it is an important component of geographic information retrieval. We describe several simple supervised methods for document geolocation using only the document's raw text as evidence. All of our methods predict locations in the context of geodesic grids of varying degrees of resolution. We evaluate the methods on geotagged Wikipedia articles and Twitter feeds. For Wikipedia, our best method obtains a median prediction error of just 11.8 kilometers. Twitter geolocation is more challenging: we obtain a median error of 479 km, an improvement on previous results for the dataset.

This research was supported by a grant from the Morris Memorial Trust Fund of the New York Community Trust and from the Longhorn Innovation Fund for Technology. This paper benefited from reviewer comments and from discussion in the Natural Language Learning reading group at UT Austin, with particular thanks to Matt Lease.

2011

schulz-etal-2020-named

https://aclanthology.org/2020.lrec-1.553

Named Entities in Medical Case Reports: Corpus and Experiments

We present a new corpus comprising annotations of medical entities in case reports, originating from PubMed Central's open access library. In the case reports, we annotate cases, conditions, findings, factors and negation modifiers. Moreover, where applicable, we annotate relations between these entities. As such, this is the first corpus of this kind made available to the scientific community in English. It enables the initial investigation of automatic information extraction from case reports through tasks like Named Entity Recognition, Relation Extraction and (sentence/paragraph) relevance detection. Additionally, we present four strong baseline systems for the detection of medical entities made available through the annotated dataset.

Good Health and Well-Being

The research presented in this article is funded by the German Federal Ministry of Education and Research (BMBF) through the project QURATOR (Unternehmen Region, Wachstumskern, grant no. 03WKDA1A), see http://qurator. ai. We want to thank our medical experts for their help annotating the data set, especially Ashlee Finckh and Sophie Klopfenstein.

2020

nakashole-mitchell-2014-language

https://aclanthology.org/P14-1095

Language-Aware Truth Assessment of Fact Candidates

This paper introduces FactChecker, language-aware approach to truth-finding. FactChecker differs from prior approaches in that it does not rely on iterative peer voting, instead it leverages language to infer believability of fact candidates. In particular, FactChecker makes use of linguistic features to detect if a given source objectively states facts or is speculative and opinionated. To ensure that fact candidates mentioned in similar sources have similar believability, FactChecker augments objectivity with a co-mention score to compute the overall believability score of a fact candidate. Our experiments on various datasets show that FactChecker yields higher accuracy than existing approaches.

Peace, Justice and Strong Institutions

We thank members of the NELL team at CMU for their helpful comments. This research was supported by DARPA under contract number FA8750-13-2-0005.

2014

gu-etal-2022-phrase

https://aclanthology.org/2022.acl-long.444

Phrase-aware Unsupervised Constituency Parsing

Recent studies have achieved inspiring success in unsupervised grammar induction using masked language modeling (MLM) as the proxy task. Despite their high accuracy in identifying low-level structures, prior arts tend to struggle in capturing high-level structures like clauses, since the MLM task usually only requires information from local context. In this work, we revisit LM-based constituency parsing from a phrase-centered perspective. Inspired by the natural reading process of human readers, we propose to regularize the parser with phrases extracted by an unsupervised phrase tagger to help the LM model quickly manage low-level structures. For a better understanding of high-level structures, we propose a phrase-guided masking strategy for LM to emphasize more on reconstructing nonphrase words. We show that the initial phrase regularization serves as an effective bootstrap, and phrase-guided masking improves the identification of high-level structures. Experiments on the public benchmark with two different backbone models demonstrate the effectiveness and generality of our method.

Research was supported in part by US DARPA KAIROS Program No. FA8750-19-2-1004, So-cialSim Program No. W911NF-17-C-0099, and INCAS Program No. HR001121C0165, National Science Foundation IIS-19-56151, IIS-17-41317, and IIS 17-04532, IBM-Illinois Discovery Accelerator Institute, and the Molecule Maker Lab Institute: An AI Research Institutes program supported by NSF under Award No. 2019897. Any opinions, findings, and conclusions or recommendations expressed herein are those of the authors and do not necessarily represent the views, either expressed or implied, of DARPA or the U.S. Government. The views and conclusions contained in this paper are those of the authors and should not be interpreted as representing any funding agencies.

2022

lagarda-casacuberta-2008-applying

https://aclanthology.org/2008.eamt-1.14

Applying boosting to statistical machine translation

Boosting is a general method for improving the accuracy of a given learning algorithm under certain restrictions. In this work, AdaBoost, one of the most popular boosting algorithms, is adapted and applied to statistical machine translation. The appropriateness of this technique in this scenario is evaluated on a real translation task. Results from preliminary experiments confirm that statistical machine translation can take advantage from this technique, improving the translation quality.

2008

darwish-etal-2014-verifiably

https://aclanthology.org/D14-1154

Verifiably Effective Arabic Dialect Identification

Several recent papers on Arabic dialect identification have hinted that using a word unigram model is sufficient and effective for the task. However, most previous work was done on a standard fairly homogeneous dataset of dialectal user comments. In this paper, we show that training on the standard dataset does not generalize, because a unigram model may be tuned to topics in the comments and does not capture the distinguishing features of dialects. We show that effective dialect identification requires that we account for the distinguishing lexical, morphological, and phonological phenomena of dialects. We show that accounting for such can improve dialect detection accuracy by nearly 10% absolute.

2014

miranda-etal-2018-multilingual

https://aclanthology.org/D18-1483

Multilingual Clustering of Streaming News

Clustering news across languages enables efficient media monitoring by aggregating articles from multilingual sources into coherent stories. Doing so in an online setting allows scalable processing of massive news streams. To this end, we describe a novel method for clustering an incoming stream of multilingual documents into monolingual and crosslingual story clusters. Unlike typical clustering approaches that consider a small and known number of labels, we tackle the problem of discovering an ever growing number of cluster labels in an online fashion, using real news datasets in multiple languages. Our method is simple to implement, computationally efficient and produces state-of-the-art results on datasets in German, English and Spanish.

We would like to thank Esma Balkır, Nikos Papasarantopoulos, Afonso Mendes, Shashi Narayan and the anonymous reviewers for their feedback. This project was supported by the European H2020 project SUMMA, grant agreement 688139 (see http://www.summa-project.eu) and by a grant from Bloomberg.

2018

popowich-etal-1997-lexicalist

https://aclanthology.org/1997.tmi-1.9

A lexicalist approach to the translation of colloquial text

Colloquial English (CE) as found in television programs or typical conversations is different than text found in technical manuals, newspapers and books. Phrases tend to be shorter and less sophisticated. In this paper, we look at some of the theoretical and implementational issues involved in translating CE. We present a fully automatic large-scale multilingual natural language processing system for translation of CE input text, as found in the commercially transmitted closed-caption television signal, into simple target sentences. Our approach is based on the Whitelock's Shake and Bake machine translation paradigm, which relies heavily on lexical resources. The system currently translates from English to Spanish with the translation modules for Brazilian Portuguese under development.

1997

stajner-popovic-2018-improving

https://aclanthology.org/W18-7006

Improving Machine Translation of English Relative Clauses with Automatic Text Simplification

This article explores the use of automatic sentence simplification as a preprocessing step in neural machine translation of English relative clauses into grammatically complex languages. Our experiments on English-to-Serbian and Englishto-German translation show that this approach can reduce technical post-editing effort (number of post-edit operations) to obtain correct translation. We find that larger improvements can be achieved for more complex target languages, as well as for MT systems with lower overall performance. The improvements mainly originate from correctly simplified sentences with relatively complex structure, while simpler structures are already translated sufficiently well using the original source sentences.

Acknowledgements: This research was supported by the ADAPT Centre for Digital Content Technology at Dublin City University, funded under the Science Foundation Ireland Research Centres Programme (Grant 13/RC/2106) and co-

2018

jansen-etal-2018-worldtree

https://aclanthology.org/L18-1433

WorldTree: A Corpus of Explanation Graphs for Elementary Science Questions supporting Multi-hop Inference

Developing methods of automated inference that are able to provide users with compelling human-readable justifications for why the answer to a question is correct is critical for domains such as science and medicine, where user trust and detecting costly errors are limiting factors to adoption. One of the central barriers to training question answering models on explainable inference tasks is the lack of gold explanations to serve as training data. In this paper we present a corpus of explanations for standardized science exams, a recent challenge task for question answering. We manually construct a corpus of detailed explanations for nearly all publicly available standardized elementary science question (approximately 1,680 3 rd through 5 th grade questions) and represent these as "explanation graphs"-sets of lexically overlapping sentences that describe how to arrive at the correct answer to a question through a combination of domain and world knowledge. We also provide an explanation-centered tablestore, a collection of semi-structured tables that contain the knowledge to construct these elementary science explanations. Together, these two knowledge resources map out a substantial portion of the knowledge required for answering and explaining elementary science exams, and provide both structured and free-text training data for the explainable inference task.

Quality Education

2018

skantze-etal-2013-exploring

https://aclanthology.org/W13-4029

Exploring the effects of gaze and pauses in situated human-robot interaction

In this paper, we present a user study where a robot instructs a human on how to draw a route on a map, similar to a Map Task. This setup has allowed us to study user reactions to the robot's conversational behaviour in order to get a better understanding of how to generate utterances in incremental dialogue systems. We have analysed the participants' subjective rating, task completion, verbal responses, gaze behaviour, drawing activity, and cognitive load. The results show that users utilise the robot's gaze in order to disambiguate referring expressions and manage the flow of the interaction. Furthermore, we show that the user's behaviour is affected by how pauses are realised in the robot's speech.

Gabriel Skantze is supported by the Swedish research council (VR) project Incremental processing in multimodal conversational systems (2011-6237). Anna Hjalmarsson is supported by the Swedish Research Council (VR) project Classifying and deploying pauses for flow control in conversational systems . Catharine Oertel is supported by GetHomeSafe (EU 7th Framework STREP 288667).

2013

gao-etal-2021-learning

https://aclanthology.org/2021.textgraphs-1.6

Learning Clause Representation from Dependency-Anchor Graph for Connective Prediction

Semantic representation that supports the choice of an appropriate connective between pairs of clauses inherently addresses discourse coherence, which is important for tasks such as narrative understanding, argumentation, and discourse parsing. We propose a novel clause embedding method that applies graph learning to a data structure we refer to as a dependencyanchor graph. The dependency anchor graph incorporates two kinds of syntactic information, constituency structure and dependency relations, to highlight the subject and verb phrase relation. This enhances coherencerelated aspects of representation. We design a neural model to learn a semantic representation for clauses from graph convolution over latent representations of the subject and verb phrase. We evaluate our method on two new datasets: a subset of a large corpus where the source texts are published novels, and a new dataset collected from students' essays. The results demonstrate a significant improvement over tree-based models, confirming the importance of emphasizing the subject and verb phrase. The performance gap between the two datasets illustrates the challenges of analyzing student's written text, plus a potential evaluation task for coherence modeling and an application for suggesting revisions to students.

2021

nan-etal-2020-reasoning

https://aclanthology.org/2020.acl-main.141

Reasoning with Latent Structure Refinement for Document-Level Relation Extraction

Document-level relation extraction requires integrating information within and across multiple sentences of a document and capturing complex interactions between inter-sentence entities. However, effective aggregation of relevant information in the document remains a challenging research question. Existing approaches construct static document-level graphs based on syntactic trees, co-references or heuristics from the unstructured text to model the dependencies. Unlike previous methods that may not be able to capture rich non-local interactions for inference, we propose a novel model that empowers the relational reasoning across sentences by automatically inducing the latent document-level graph. We further develop a refinement strategy, which enables the model to incrementally aggregate relevant information for multi-hop reasoning. Specifically, our model achieves an F 1 score of 59.05 on a large-scale documentlevel dataset (DocRED), significantly improving over the previous results, and also yields new state-of-the-art results on the CDR and GDA dataset. Furthermore, extensive analyses show that the model is able to discover more accurate inter-sentence relations. * * Equally Contributed. † † Work done during internship at SUTD.

We would like to thank the anonymous reviewers for their thoughtful and constructive comments. This research is supported by Ministry of Education, Singapore, under its Academic Research Fund (AcRF) Tier 2 Programme (MOE AcRF Tier 2 Award No: MOE2017-T2-1-156). Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of the Ministry of Education, Singapore.

2020

martin-etal-2020-leveraging

https://aclanthology.org/2020.cllrd-1.5

Leveraging Non-Specialists for Accurate and Time Efficient AMR Annotation

Meaning Representations (AMRs), a syntax-free representation of phrase semantics (Banarescu et al., 2013), are useful for capturing the meaning of a phrase and reflecting the relationship between concepts that are referred to. However, annotating AMRs is time consuming and expensive. The existing annotation process requires expertly trained workers who have knowledge of an extensive set of guidelines for parsing phrases. In this paper, we propose a cost-saving two-step process for the creation of a corpus of AMR-phrase pairs for spatial referring expressions. The first step uses non-specialists to perform simple annotations that can be leveraged in the second step to accelerate the annotation performed by the experts. We hypothesize that our process will decrease the cost per annotation and improve consistency across annotators. Few corpora of spatial referring expressions exist and the resulting language resource will be valuable for referring expression comprehension and generation modeling.

This work is partially supported by the National Science Foundation award number 1849357.

2020

elder-etal-2020-make

https://aclanthology.org/2020.emnlp-main.230

How to Make Neural Natural Language Generation as Reliable as Templates in Task-Oriented Dialogue

Neural Natural Language Generation (NLG) systems are well known for their unreliability. To overcome this issue, we propose a data augmentation approach which allows us to restrict the output of a network and guarantee reliability. While this restriction means generation will be less diverse than if randomly sampled, we include experiments that demonstrate the tendency of existing neural generation approaches to produce dull and repetitive text, and we argue that reliability is more important than diversity for this task. The system trained using this approach scored 100% in semantic accuracy on the E2E NLG Challenge dataset, the same as a template system.

We thank the anonymous reviewers for their helpful comments. This research is supported by Science Foundation Ireland in the ADAPT Centre for Digital Content Technology. The ADAPT Centre for Digital Content Technology is funded under the SFI Research Centres Programme (Grant 13/RC/2106) and is co-funded under the European Regional Development Fund.

2020

nagesh-2015-exploring

https://aclanthology.org/N15-2006

Exploring Relational Features and Learning under Distant Supervision for Information Extraction Tasks

Information Extraction (IE) has become an indispensable tool in our quest to handle the data deluge of the information age. IE can broadly be classified into Named-entity Recognition (NER) and Relation Extraction (RE). In this thesis, we view the task of IE as finding patterns in unstructured data, which can either take the form of features and/or be specified by constraints. In NER, we study the categorization of complex relational 1 features and outline methods to learn feature combinations through induction. We demonstrate the efficacy of induction techniques in learning : i) rules for the identification of named entities in text-the novelty is the application of induction techniques to learn in a very expressive declarative rule language ii) a richer sequence labeling model-enabling optimal learning of discriminative features. In RE, our investigations are in the paradigm of distant supervision, which facilitates the creation of large albeit noisy training data. We devise an inference framework in which constraints can be easily specified in learning relation extractors. In addition, we reformulate the learning objective in a max-margin framework. To the best of our knowledge, our formulation is the first to optimize multi-variate non-linear performance measures such as F β for a latent variable structure prediction task.

2015

iosif-etal-2012-associative

http://www.lrec-conf.org/proceedings/lrec2012/pdf/536_Paper.pdf

Associative and Semantic Features Extracted From Web-Harvested Corpora

We address the problem of automatic classification of associative and semantic relations between words, and particularly those that hold between nouns. Lexical relations such as synonymy, hypernymy/hyponymy, constitute the fundamental types of semantic relations. Associative relations are harder to define, since they include a long list of diverse relations, e.g., "Cause-Effect", "Instrument-Agency". Motivated by findings from the literature of psycholinguistics and corpus linguistics, we propose features that take advantage of general linguistic properties. For evaluation we merged three datasets assembled and validated by cognitive scientists. A proposed priming coefficient that measures the degree of asymmetry in the order of appearance of the words in text achieves the best classification results, followed by context-based similarity metrics. The web-based features achieve classification accuracy that exceeds 85%.

2012

sayeed-etal-2012-grammatical

https://aclanthology.org/N12-1085

Grammatical structures for word-level sentiment detection

Existing work in fine-grained sentiment analysis focuses on sentences and phrases but ignores the contribution of individual words and their grammatical connections. This is because of a lack of both (1) annotated data at the word level and (2) algorithms that can leverage syntactic information in a principled way. We address the first need by annotating articles from the information technology business press via crowdsourcing to provide training and testing data. To address the second need, we propose a suffix-tree data structure to represent syntactic relationships between opinion targets and words in a sentence that are opinion-bearing. We show that a factor graph derived from this data structure acquires these relationships with a small number of word-level features. We demonstrate that our supervised model performs better than baselines that ignore syntactic features and constraints.

This paper is based upon work supported by the US National Science Foundation under Grant IIS-0729459. Additional support came from the Cluster of Excellence "Multimodal Computing and Innovation", Germany. Jordan Boyd-Graber is also supported by US National Science Foundation Grant NSF grant #1018625 and the Army Research Laboratory through ARL Cooperative Agreement W911NF-09-2-0072. Any opinions, findings, conclusions, or recommendations expressed are the authors' and do not necessarily reflect those of the sponsors.

2012

feng-2003-cooperative

https://aclanthology.org/N03-3010

Cooperative model-based language understanding

In this paper, we propose a novel Cooperative Model for natural language understanding in a dialogue system. We build this based on both Finite State Model (FSM) and Statistical Learning Model (SLM). FSM provides two strategies for language understanding and have a high accuracy but little robustness and flexibility. Statistical approach is much more robust but less accurate. Cooperative Model incorporates all the three strategies together and thus can suppress all the shortcomings of different strategies and has all the advantages of the three strategies.

The author would like to thank Deepak Ravichandran for his invaluable help of the whole work.

2003

romanov-etal-2019-adversarial

https://aclanthology.org/N19-1088

Adversarial Decomposition of Text Representation

In this paper, we present a method for adversarial decomposition of text representation. This method can be used to decompose a representation of an input sentence into several independent vectors, each of them responsible for a specific aspect of the input sentence. We evaluate the proposed method on two case studies: the conversion between different social registers and diachronic language change. We show that the proposed method is capable of fine-grained controlled change of these aspects of the input sentence. It is also learning a continuous (rather than categorical) representation of the style of the sentence, which is more linguistically realistic. The model uses adversarial-motivational training and includes a special motivational loss, which acts opposite to the discriminator and encourages a better decomposition. Furthermore, we evaluate the obtained meaning embeddings on a downstream task of paraphrase detection and show that they significantly outperform the embeddings of a regular autoencoder.

2019

reeves-1982-terminology

https://aclanthology.org/1982.tc-1.12

Terminology for translators

Should a technical translator be a subject specialist with additional linguistic skills? Or a trained linguist with some specialist knowledge? It is an old debate and plainly in practice successful translators can derive from both categories. Indeed in the past entry into the profession was often largely determined by personal circumstances -an engineer who had acquired linguistic knowledge through overseas postings might turn later in his career, or as a side-line, to translating engineering texts. A language graduate, finding him or usually herself, confined to earning a living from the home, acquired knowledge of a technical area in a self-teaching process. Today, however, the enormous growth of scientific discovery and technological innovation together with the internationalisation of trade make, as we all know, the systematic training of translators a necessity. Decisions therefore have to be taken about the most efficacious methods to be adopted in the training process and the old question of linguist versus specialist recurs with fresh urgency. Or at least it would appear to. But there is a further complicating factor: the technical translator is principally concerned with the language in which the message is expressed, whereas the sender was principally concerned with the topic of the message. The sender used the special language of his area to describe and analyse the extra-linguistic reality that was his primary interest. But the translator's primary interest is the special language itself: in short a subject's terminology assumes first importance for the translator. And the moment we speak of terminology in this context as 'an aggregate of terms representing the system of concepts in an individual subject field'(1), we are reminded that the translator also needs to understand the principles according to which a particular terminology is established, the relationship between various monolingual terminologies and between the specialist terminologies of one language and those of another language(2). Thus the technical translator has to be an expert in three discrete disciplines: translation itself, a technical specialism and the theory and practice of terminology.

1982

yusupov-kuratov-2018-nips

https://aclanthology.org/C18-1312

NIPS Conversational Intelligence Challenge 2017 Winner System: Skill-based Conversational Agent with Supervised Dialog Manager

We present bot#1337: a dialog system developed for the 1 st NIPS Conversational Intelligence Challenge 2017 (ConvAI). The aim of the competition was to implement a bot capable of conversing with humans based on a given passage of text. To enable conversation, we implemented a set of skills for our bot, including chitchat , topic detection, text summarization, question answering and question generation. The system has been trained in a supervised setting using a dialogue manager to select an appropriate skill for generating a response. The latter allows a developer to focus on the skill implementation rather than the finite state machine based dialog manager. The proposed system bot#1337 won the competition with an average dialogue quality score of 2.78 out of 5 given by human evaluators. Source code and trained models for the bot#1337 are available on GitHub.

We thank Mikhail Burtsev, Luiza Sayfullina and Mikhail Pavlov for comments that greatly improved the manuscript. We would also like to thank the Reason8.ai company for providing computational resources and grant for NIPS 2017 ticket. We thank Neural Systems and Deep Learning Lab of MIPT for ideas and support.

2018

trawinski-2003-licensing

https://aclanthology.org/W03-1813

Licensing Complex Prepositions via Lexical Constraints

In this paper, we will investigate a cross-linguistic phenomenon referred to as complex prepositions (CPs), which is a frequent type of multiword expressions (MWEs) in many languages. Based on empirical data, we will point out the problems of the traditional treatment of CPs as complex lexical categories, and, thus, propose an analysis using the formal paradigm of the HPSG in the tradition of (Pollard and Sag, 1994). Our objective is to provide an approach to CPs which (1) convincingly explains empirical data, (2) is consistent with the underlying formal framework and does not require any extensions or modifications of the existing description apparatus, (3) is computationally tractable.

I would like to thank Manfred Sailer, Frank Richter, and the anonymous reviewers of the ACL-2003 Workshop on Multiword Expressions: Analysis, Acquisition and Treatment in Sapporo for their interesting comments on the issue presented in this paper and Carmella Payne for help with English.

2003

yeniterzi-oflazer-2010-syntax

https://aclanthology.org/P10-1047

Syntax-to-Morphology Mapping in Factored Phrase-Based Statistical Machine Translation from English to Turkish

We present a novel scheme to apply factored phrase-based SMT to a language pair with very disparate morphological structures. Our approach relies on syntactic analysis on the source side (English) and then encodes a wide variety of local and non-local syntactic structures as complex structural tags which appear as additional factors in the training data. On the target side (Turkish), we only perform morphological analysis and disambiguation but treat the complete complex morphological tag as a factor, instead of separating morphemes. We incrementally explore capturing various syntactic substructures as complex tags on the English side, and evaluate how our translations improve in BLEU scores. Our maximal set of source and target side transformations, coupled with some additional techniques, provide an 39% relative improvement from a baseline 17.08 to 23.78 BLEU, all averaged over 10 training and test sets. Now that the syntactic analysis on the English side is available, we also experiment with more long distance constituent reordering to bring the English constituent order close to Turkish, but find that these transformations do not provide any additional consistent tangible gains when averaged over the 10 sets.

We thank Joakim Nivre for providing us with the parser. This publication was made possible by the generous support of the Qatar Foundation through Carnegie Mellon University's Seed Research program. The statements made herein are solely the responsibility of the authors.

2010

zhang-etal-2012-whitepaper

https://aclanthology.org/W12-4401

Whitepaper of NEWS 2012 Shared Task on Machine Transliteration

Transliteration is defined as phonetic translation of names across languages. Transliteration of Named Entities (NEs) is necessary in many applications, such as machine translation, corpus alignment, cross-language IR, information extraction and automatic lexicon acquisition. All such systems call for high-performance transliteration, which is the focus of shared task in the NEWS 2012 workshop. The objective of the shared task is to promote machine transliteration research by providing a common benchmarking platform for the community to evaluate the state-of-the-art technologies.

2012

lee-lee-2014-postech

https://aclanthology.org/W14-1709

POSTECH Grammatical Error Correction System in the CoNLL-2014 Shared Task

This paper describes the POSTECH grammatical error correction system. Various methods are proposed to correct errors such as rule-based, probability n-gram vector approaches and router-based approach. Google N-gram count corpus is used mainly as the correction resource. Correction candidates are extracted from NUCLE training data and each candidate is evaluated with development data to extract high precision rules and n-gram frames. Out of 13 participating teams, our system is ranked 4 th on both the original and revised annotation.

This research was supported by the MSIP(The Ministry of Science, ICT and Future Planning), Korea and Microsoft Research, under IT/SW Creative research program supervised by the NIPA(National IT Industry Promotion Agency) (NIPA-2013-H0503-13-1006) and this research was supported by the Basic Science Research Program through the National Research Foundation of Korea(NRF) funded by the Ministry of Education, Science and Technology(2010-0019523).

2014

aoki-yamamoto-2007-opinion

https://aclanthology.org/Y07-1007

Opinion Extraction based on Syntactic Pieces

This paper addresses a task of opinion extraction from given documents and its positive/negative classification. We propose a sentence classification method using a notion of syntactic piece. Syntactic piece is a minimum unit of structure, and is used as an alternative processing unit of n-gram and whole tree structure. We compute its semantic orientation, and classify opinion sentences into positive or negative. We have conducted an experiment on more than 5000 opinion sentences of multiple domains, and have proven that our approach attains high performance at 91% precision.

2007

junczys-dowmunt-grundkiewicz-2016-phrase

https://aclanthology.org/D16-1161

Phrase-based Machine Translation is State-of-the-Art for Automatic Grammatical Error Correction

In this work, we study parameter tuning towards the M 2 metric, the standard metric for automatic grammar error correction (GEC) tasks. After implementing M 2 as a scorer in the Moses tuning framework, we investigate interactions of dense and sparse features, different optimizers, and tuning strategies for the CoNLL-2014 shared task. We notice erratic behavior when optimizing sparse feature weights with M 2 and offer partial solutions. We find that a bare-bones phrase-based SMT setup with task-specific parameter-tuning outperforms all previously published results for the CoNLL-2014 test set by a large margin (46.37% M 2 over previously 41.75%, by an SMT system with neural features) while being trained on the same, publicly available data. Our newly introduced dense and sparse features widen that gap, and we improve the state-of-the-art to 49.49% M 2 .

The authors would like to thank Colin Cherry for his help with Batch Mira hyper-parameters and Kenneth Heafield for many helpful comments and discussions. This work was partially funded by the Polish National Science Centre (Grant No. 2014/15/N/ST6/02330) and by Facebook. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of Facebook.

2016

childs-etal-1998-coreference

https://aclanthology.org/X98-1010

Coreference Resolution Strategies From an Application Perspective

As part of our TIPSTER III research program, we have continued our research into strategies to resolve coreferences within a free text document; this research was begun during our TIPSTER II research program. In the TIPSTER II Proceedings paper, "An Evaluation of Coreference Resolution Strategies for Acquiring Associated Information," the goal was to evaluate the contributions of various techniques for associating an entity with three types of information: 1) name variations, 2) descriptive phrases, and 3) location information.

1998

costa-jussa-etal-2020-multilingual

https://aclanthology.org/2020.cl-2.1

Multilingual and Interlingual Semantic Representations for Natural Language Processing: A Brief Introduction

We introduce the Computational Linguistics special issue on Multilingual and Interlingual Semantic Representations for Natural Language Processing. We situate the special issue's five articles in the context of our fast-changing field, explaining our motivation for this project. We offer a brief summary of the work in the issue, which includes developments on lexical and sentential semantic representations, from symbolic and neural perspectives. 1. Motivation This special issue arose from our observation of two trends in the fields of computational linguistics and natural language processing. The first trend is a matter of increasing demand for language technologies that serve diverse populations, particularly those whose languages have received little attention in the research community.

We thank Kyle Lo for assistance with the S2ORC data. MRC is supported in part by a Google Faculty Research Award 2018, Spanish Ministerio de Economía y Competitividad, the European Regional Development Fund and the Agencia Estatal de Investigación, through the postdoctoral senior grant Ramón y Cajal, the contract TEC2015-69266-P (MINECO/FEDER,EU) and the contract PCIN-2017-079 (AEI/MINECO). N. A. S. is supported by National Science Foundation grant IIS-1562364. C. E. B. is funded by the German Federal Ministry of Education and Research under the funding code 01IW17001 (Deeplee). Responsibility for the content of this publication is with the authors.

2020

kotlerman-etal-2012-sentence

https://aclanthology.org/S12-1005

Sentence Clustering via Projection over Term Clusters

This paper presents a novel sentence clustering scheme based on projecting sentences over term clusters. The scheme incorporates external knowledge to overcome lexical variability and small corpus size, and outperforms common sentence clustering methods on two reallife industrial datasets.

2012