Datasets:

feradauto

/

NLP4SGPapers

like 2

yuan-etal-2021-cambridge

https://aclanthology.org/2021.semeval-1.74

Cambridge at SemEval-2021 Task 1: An Ensemble of Feature-Based and Neural Models for Lexical Complexity Prediction

This paper describes our submission to the SemEval-2021 shared task on Lexical Complexity Prediction. We approached it as a regression problem and present an ensemble combining four systems, one feature-based and three neural with fine-tuning, frequency pre-training and multi-task learning, achieving Pearson scores of 0.8264 and 0.7556 on the trial and test sets respectively (sub-task 1). We further present our analysis of the results and discuss our findings.

We thank Sian Gooding and Ekaterina Kochmar for support and advice. This paper reports on research supported by Cambridge Assessment, University of Cambridge. This work was performed using resources provided by the Cambridge Service for Data Driven Discovery operated by the University of Cambridge Research Computing Service, provided by Dell EMC and Intel using Tier-2 funding from the Engineering and Physical Sciences Research Council (capital grant EP/P020259/1), and DiRAC funding from the Science and Technology Facilities Council. We acknowledge NVIDIA for an Academic Hardware Grant.

2021

filice-etal-2017-kelp

https://aclanthology.org/S17-2053

KeLP at SemEval-2017 Task 3: Learning Pairwise Patterns in Community Question Answering

This paper describes the KeLP system participating in the SemEval-2017 community Question Answering (cQA) task. The system is a refinement of the kernel-based sentence pair modeling we proposed for the previous year challenge. It is implemented within the Kernel-based Learning Platform called KeLP, from which we inherit the team's name. Our primary submission ranked first in subtask A, and third in subtasks B and C, being the only systems appearing in the top-3 ranking for all the English subtasks. This shows that the proposed framework, which has minor variations among the three subtasks, is extremely flexible and effective in tackling learning tasks defined on sentence pairs.

This work has been partially supported by the EC project CogNet, 671625 (H2020-ICT-2014-2, Research and Innovation action).

2017

zhou-etal-2021-low

https://aclanthology.org/2021.sustainlp-1.1

Low Resource Quadratic Forms for Knowledge Graph Embeddings

We address the problem of link prediction between entities and relations of knowledge graphs. State of the art techniques that address this problem, while increasingly accurate, are computationally intensive. In this paper we cast link prediction as a sparse convex program whose solution defines a quadratic form that is used as a ranking function. The structure of our convex program is such that standard support vector machine software packages, which are numerically robust and efficient, can solve it. We show that on benchmark data sets, our model's performance is competitive with state of the art models, but training times can be reduced by a factor of 40 using only CPUbased (and not GPU-accelerated) computing resources. This approach may be suitable for applications where balancing the demands of graph completion performance against computational efficiency is a desirable trade-off.

2021

li-etal-2015-dependency-parsing

https://aclanthology.org/Y15-2039

Dependency parsing for Chinese long sentence: A second-stage main structure parsing method

This paper explores the problem of parsing Chinese long sentences. Inspired by human sentence processing, a second-stage parsing method, referred as main structure parsing in this paper, are proposed to improve the parsing performance as well as maintaining its high accuracy and efficiency on Chinese long sentences. Three different methods have attempted in this paper and the result shows that the best performance comes from the method using Chinese comma as the boundary of the sub-sentence. According to our experiment about testing on the Chinese dependency Treebank 1.0 data, it improves long dependency accuracy by around 6.0% than the baseline parser and 3.2% than the previous best model.

2015

tkachenko-etal-2018-searching

https://aclanthology.org/P18-1112

Searching for the X-Factor: Exploring Corpus Subjectivity for Word Embeddings

We explore the notion of subjectivity, and hypothesize that word embeddings learnt from input corpora of varying levels of subjectivity behave differently on natural language processing tasks such as classifying a sentence by sentiment, subjectivity, or topic. Through systematic comparative analyses, we establish this to be the case indeed. Moreover, based on the discovery of the outsized role that sentiment words play on subjectivity-sensitive tasks such as sentiment classification, we develop a novel word embedding SentiVec which is infused with sentiment information from a lexical resource, and is shown to outperform baselines on such tasks.

This research is supported by the National Research Foundation, Prime Minister's Office, Singapore under its NRF Fellowship Programme (Award No. NRF-NRFF2016-07).

2018

shuster-etal-2020-image

https://aclanthology.org/2020.acl-main.219

Image-Chat: Engaging Grounded Conversations

To achieve the long-term goal of machines being able to engage humans in conversation, our models should captivate the interest of their speaking partners. Communication grounded in images, whereby a dialogue is conducted based on a given photo, is a setup naturally appealing to humans (Hu et al., 2014). In this work we study large-scale architectures and datasets for this goal. We test a set of neural architectures using state-of-the-art image and text representations, considering various ways to fuse the components. To test such models, we collect a dataset of grounded human-human conversations, where speakers are asked to play roles given a provided emotional mood or style, as the use of such traits is also a key factor in engagingness (Guo et al., 2019). Our dataset, Image-Chat, consists of 202k dialogues over 202k images using 215 possible style traits. Automatic metrics and human evaluations of engagingness show the efficacy of our approach; in particular, we obtain state-of-the-art performance on the existing IGC task, and our best performing model is almost on par with humans on the Image-Chat test set (preferred 47.7% of the time).

2020

xia-etal-2000-comparing

https://aclanthology.org/W00-1208

Comparing Lexicalized Treebank Grammars Extracted from Chinese, Korean, and English Corpora

In this paper, we present a method for comparing Lexicalized Tree Adjoining Grammars extracted from annotated corpora for three languages: English, Chinese and Korean. This method makes it possible to do a quantitative comparison between the syntactic structures of each language, thereby providing a way of testing the Universal Grammar Hypothesis, the foundation of modern linguistic theories.

2000

lubis-etal-2018-unsupervised

https://aclanthology.org/W18-5017

Unsupervised Counselor Dialogue Clustering for Positive Emotion Elicitation in Neural Dialogue System

Positive emotion elicitation seeks to improve user's emotional state through dialogue system interaction, where a chatbased scenario is layered with an implicit goal to address user's emotional needs. Standard neural dialogue system approaches still fall short in this situation as they tend to generate only short, generic responses. Learning from expert actions is critical, as these potentially differ from standard dialogue acts. In this paper, we propose using a hierarchical neural network for response generation that is conditioned on 1) expert's action, 2) dialogue context, and 3) user emotion, encoded from user input. We construct a corpus of interactions between a counselor and 30 participants following a negative emotional exposure to learn expert actions and responses in a positive emotion elicitation scenario. Instead of relying on the expensive, labor intensive, and often ambiguous human annotations, we unsupervisedly cluster the expert's responses and use the resulting labels to train the network. Our experiments and evaluation show that the proposed approach yields lower perplexity and generates a larger variety of responses.

Good Health and Well-Being

Part of this work was supported by JSPS KAKENHI Grant Numbers JP17H06101 and JP17K00237.

2018

van-halteren-2008-source

https://aclanthology.org/C08-1118

Source Language Markers in EUROPARL Translations

This paper shows that it is very often possible to identify the source language of medium-length speeches in the EU-ROPARL corpus on the basis of frequency counts of word n-grams (87.2%-96.7% accuracy depending on classification method). The paper also examines in detail which positive markers are most powerful and identifies a number of linguistic aspects as well as culture-and domain-related ones.1

2008

karimi-tang-2019-learning

https://aclanthology.org/N19-1347

Learning Hierarchical Discourse-level Structure for Fake News Detection

On the one hand, nowadays, fake news articles are easily propagated through various online media platforms and have become a grand threat to the trustworthiness of information. On the other hand, our understanding of the language of fake news is still minimal. Incorporating hierarchical discourse-level structure of fake and real news articles is one crucial step toward a better understanding of how these articles are structured. Nevertheless, this has rarely been investigated in the fake news detection domain and faces tremendous challenges. First, existing methods for capturing discourse-level structure rely on annotated corpora which are not available for fake news datasets. Second, how to extract out useful information from such discovered structures is another challenge. To address these challenges, we propose Hierarchical Discourselevel Structure for Fake news detection. HDSF learns and constructs a discourse-level structure for fake/real news articles in an automated and data-driven manner. Moreover, we identify insightful structure-related properties, which can explain the discovered structures and boost our understating of fake news. Conducted experiments show the effectiveness of the proposed approach. Further structural analysis suggests that real and fake news present substantial differences in the hierarchical discourse-level structures.

Peace, Justice and Strong Institutions

2019

bari-etal-2021-uxla

https://aclanthology.org/2021.acl-long.154

UXLA: A Robust Unsupervised Data Augmentation Framework for Zero-Resource Cross-Lingual NLP

Transfer learning has yielded state-of-the-art (SoTA) results in many supervised NLP tasks. However, annotated data for every target task in every target language is rare, especially for low-resource languages. We propose UXLA a novel unsupervised data augmentation framework for zero-resource transfer learning scenarios. In particular, UXLA aims to solve crosslingual adaptation problems from a source language task distribution to an unknown target language task distribution, assuming no training label in the target language. At its core, UXLA performs simultaneous selftraining with data augmentation and unsupervised sample selection. To show its effectiveness, we conduct extensive experiments on three diverse zero-resource cross-lingual transfer tasks. UXLA achieves SoTA results in all the tasks, outperforming the baselines by a good margin. With an in-depth framework dissection, we demonstrate the cumulative contributions of different components to its success.

2021

navigli-etal-2010-annotated

http://www.lrec-conf.org/proceedings/lrec2010/pdf/20_Paper.pdf

An Annotated Dataset for Extracting Definitions and Hypernyms from the Web

This paper presents and analyzes an annotated corpus of definitions, created to train an algorithm for the automatic extraction of definitions and hypernyms from Web documents. As an additional resource, we also include a corpus of non-definitions with syntactic patterns similar to those of definition sentences, e.g.: "An android is a robot" vs. "Snowcap is unmistakable". Domain and style independence is obtained thanks to the annotation of a sample of the Wikipedia corpus and to a novel pattern generalization algorithm based on wordclass lattices (WCL). A lattice is a directed acyclic graph (DAG), a subclass of nondeterministic finite state automata (NFA). The lattice structure has the purpose of preserving the salient differences among distinct sequences, while eliminating redundant information. The WCL algorithm will be integrated into an improved version of the GlossExtractor Web application (Velardi et al., 2008). This paper is mostly concerned with a description of the corpus, the annotation strategy, and a linguistic analysis of the data. A summary of the WCL algorithm is also provided for the sake of completeness.

2010

deville-etal-1996-anthem

https://aclanthology.org/1996.amta-1.27

ANTHEM: advanced natural language interface for multilingual text generation in healthcare (LRE 62-007)

The ANTHEM project: "Advanced Natural Language Interface for Multilingual Text Generation in Healthcare" (LRE 62-007) is co-financed by the European Union within the "Linguistic Research and Engineering" program. The ANTHEM consortium is coordinated by W. Ceusters of RAMIT vzw (Ghent University Hospital) and further consists of the Institute of Modern Languages of the University of Namur (G. Deville), the IAI of the University of Saarbrücken (O. Streiter), the CRP-CU of Luxembourg (P. Mousel), the University of Liege (C. Gérardy), Datasoft Management nv -Oostende (J. Devlies) and the Military Hospital in Brussels (D. Penson).

Good Health and Well-Being

1996

sapena-etal-2010-global

https://aclanthology.org/C10-2125

A Global Relaxation Labeling Approach to Coreference Resolution

This paper presents a constraint-based graph partitioning approach to coreference resolution solved by relaxation labeling. The approach combines the strengths of groupwise classifiers and chain formation methods in one global method. Experiments show that our approach significantly outperforms systems based on separate classification and chain formation steps, and that it achieves the best results in the state of the art for the same dataset and metrics.

2010

artetxe-etal-2015-building

https://aclanthology.org/2015.eamt-1.3

Building hybrid machine translation systems by using an EBMT preprocessor to create partialtranslations

This paper presents a hybrid machine translation framework based on a preprocessor that translates fragments of the input text by using example-based machine translation techniques. The preprocessor resembles a translation memory with named-entity and chunk generalization, and generates a high quality partial translation that is then completed by the main translation engine, which can be either rule-based (RBMT) or statistical (SMT). Results are reported for both RBMT and SMT hybridization as well as the preprocessor on its own, showing the effectiveness of our approach.

The research leading to these results was carried out as part of the TACARDI project (Spanish Ministry of Education and Science, TIN2012-38523-C02-011, with FEDER funding) and the QTLeap project funded by the European Commission (FP7-ICT-2013.4.1-610516).

2015

pawar-etal-2015-noun

https://aclanthology.org/W15-5905

Noun Phrase Chunking for Marathi using Distant Supervision

Information Extraction from Indian languages requires effective shallow parsing, especially identification of "meaningful" noun phrases. Particularly, for an agglutinative and free word order language like Marathi, this problem is quite challenging. We model this task of extracting noun phrases as a sequence labelling problem. A Distant Supervision framework is used to automatically create a large labelled data for training the sequence labelling model. The framework exploits a set of heuristic rules based on corpus statistics for the automatic labelling. Our approach puts together the benefits of heuristic rules, a large unlabelled corpus as well as supervised learning to model complex underlying characteristics of noun phrase occurrences. In comparison to a simple English-like chunking baseline and a publicly available Marathi Shallow Parser, our method demonstrates a better performance.

2015

webb-etal-2008-cross

http://www.lrec-conf.org/proceedings/lrec2008/pdf/502_paper.pdf

Cross-Domain Dialogue Act Tagging

We present recent work in the area of Cross-Domain Dialogue Act (DA) tagging. We have previously reported on the use of a simple dialogue act classifier based on purely intra-utterance features-principally involving word n-gram cue phrases automatically generated from a training corpus. Such a classifier performs surprisingly well, rivalling scores obtained using far more sophisticated language modelling techniques. In this paper, we apply these automatically extracted cues to a new annotated corpus, to determine the portability and generality of the cues we learn.

2008

volokh-neumann-2012-parsing

https://aclanthology.org/W12-5615

Parsing Hindi with MDParser

We describe our participation in the MTPIL Hindi Parsing Shared Task-2012. Our system achieved the following results: 82.44% LAS/90.91% UAS (auto) and 85.31% LAS/92.88% UAS (gold). Our parser is based on the linear classification, which is suboptimal as far as the accuracy is concerned. The strong point of our approach is its speed. For parsing development the system requires 0.935 seconds, which corresponds to a parsing speed of 1318 sentences per second. The Hindi Treebank contains much less different part of speech tags than many other treebanks and therefore it was absolutely necessary to use the additional morphosyntactic features available in the treebank. We were able to build classifiers predicting those, using only the standard word form and part of speech features, with a high accuracy.

The work presented here was partially supported by a research grant from the German Federal Ministry of Education and Research (BMBF) to the DFKI project Deependance (FKZ. 01IW11003).

2012

moeljadi-etal-2015-building

https://aclanthology.org/W15-3302

Building an HPSG-based Indonesian Resource Grammar (INDRA)

This paper presents the creation and the initial stage development of a broad

Thanks to Michael Wayne Goodman and Dan Flickinger for teaching us how to use GitHub and FFTB. Thanks to Fam Rashel for helping us with POS Tagger and to Lian Tze Lim for helping us improve Wordnet Bahasa. This research was supported in part by the MOE Tier 2 grant That's what you meant: a Rich Representation for Manipulation of Meaning (MOE ARC41/13).

2015

borovikov-etal-2009-edeal

https://aclanthology.org/2009.mtsummit-government.7

The EDEAL Project for Automated Processing of African Languages

The EDEAL project seeks to identify, collect, evaluate, and enhance resources relevant to processing collected material in African languages. Its priority languages are Swahili, Hausa, Oromo, and Yoruba. Resources of interest include software for OCR, Machine Translation (MT), and Named Entity Extraction (NEE), as well as data resources for developing and evaluating tools for these languages, and approaches-whether automated or manual-for developing capabilities for languages that lack significant data resources and reference material. We have surveyed the available resources, and the project is now in its first execution phase, focused on providing end-to-end capabilities and solid data coverage for a single language; we have chosen Swahili since it has the best existing coverage to build on. The results of the work will be freely available to the U.S. Government community.

The work described here is performed by a team at CACI that includes, in addition to the authors, Marta Cruz, Mark Turner, and a large team of native speakers of different African languages.This work is sponsored by funding from the Defense Intelligence Agency (DIA) under contract GS-35F-0342N. We are very grateful for the wise guidance of Nick Bemish and Theresa Williams.

2009

offersgaard-hansen-2016-facilitating

https://aclanthology.org/L16-1398

Facilitating Metadata Interoperability in CLARIN-DK

The issue for CLARIN archives at the metadata level is to facilitate the user's possibility to describe their data, even with their own standard, and at the same time make these metadata meaningful for a variety of users with a variety of resource types, and ensure that the metadata are useful for search across all resources both at the national and at the European level. We see that different people from different research communities fill in the metadata in different ways even though the metadata was defined and documented. This has impacted when the metadata are harvested and displayed in different environments. A loss of information is at stake. In this paper we view the challenges of ensuring metadata interoperability through examples of propagation of metadata values from the CLARIN-DK archive to the VLO. We see that the CLARIN Community in many ways support interoperability, but argue that agreeing upon standards, making clear definitions of the semantics of the metadata and their content is inevitable for the interoperability to work successfully. The key points are clear and freely available definitions, accessible documentation and easily usable facilities and guidelines for the metadata creators.

2016

ito-etal-2020-langsmith

https://aclanthology.org/2020.emnlp-demos.28

Langsmith: An Interactive Academic Text Revision System

Despite the current diversity and inclusion initiatives in the academic community, researchers with a non-native command of English still face significant obstacles when writing papers in English. This paper presents the Langsmith editor, which assists inexperienced, non-native researchers to write English papers, especially in the natural language processing (NLP) field. Our system can suggest fluent, academic-style sentences to writers based on their rough, incomplete phrases or sentences. The system also encourages interaction between human writers and the computerized revision system. The experimental results demonstrated that Langsmith helps non-native English-speaker students write papers in English. The system is available at https://emnlp-demo.editor. langsmith.co.jp/. * The authors contributed equally 1 The 58th Annual Meeting of the Association for Computational Linguistics 2 See https://www.youtube.com/channel/ UCjHeZPe0tT6bWxVVvum1bFQ for the screencast.

Industry, Innovation and Infrastructure

We are grateful to Ana Brassard for her feedback on English. We also appreciate the participants of our user studies. This work was supported by Grant-in-Aid for JSPS Fellows Grant Number JP20J22697. 21 We conducted the one-side sign test. The difference is significant with p ≤ 0.05.

2020

chang-etal-2019-bias

https://aclanthology.org/D19-2004

Bias and Fairness in Natural Language Processing

Reduced Inequalities

Gender Equality

2019

mccoy-etal-2020-berts

https://aclanthology.org/2020.blackboxnlp-1.21

BERTs of a feather do not generalize together: Large variability in generalization across models with similar test set performance

If the same neural network architecture is trained multiple times on the same dataset, will it make similar linguistic generalizations across runs? To study this question, we finetuned 100 instances of BERT on the Multigenre Natural Language Inference (MNLI) dataset and evaluated them on the HANS dataset, which evaluates syntactic generalization in natural language inference. On the MNLI development set, the behavior of all instances was remarkably consistent, with accuracy ranging between 83.6% and 84.8%. In stark contrast, the same models varied widely in their generalization performance. For example, on the simple case of subject-object swap (e.g., determining that the doctor visited the lawyer does not entail the lawyer visited the doctor), accuracy ranged from 0.0% to 66.2%. Such variation is likely due to the presence of many local minima in the loss surface that are equally attractive to a low-bias learner such as a neural network; decreasing the variability may therefore require models with stronger inductive biases.

We are grateful to Emily Pitler, Dipanjan Das, and the members of the Johns Hopkins Computation and Psycholinguistics lab group for helpful comments. Any errors are our own.This project is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. 1746891 and by a gift to TL from Google, and it was conducted using computational resources from the Maryland Advanced Research Computing Center (MARCC). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation, Google, or MARCC.

2020

jimenez-etal-2013-unal

https://aclanthology.org/S13-2020

UNAL: Discriminating between Literal and Figurative Phrasal Usage Using Distributional Statistics and POS tags

In this paper we describe the system used to participate in the sub task 5b in the Phrasal Semantics challenge (task 5) in SemEval 2013. This sub task consists in discriminating literal and figurative usage of phrases with compositional and non-compositional meanings in context. The proposed approach is based on part-of-speech tags, stylistic features and distributional statistics gathered from the same development-training-test text collection. The system obtained a relative improvement in accuracy against the most-frequentclass baseline of 49.8% in the "unseen contexts" (LexSample) setting and 8.5% in "unseen phrases" (AllWords).

This research was funded in part by the Systems and Industrial Engineering Department, the Office of Student Welfare of the National University of Colombia, Bogotá, and through a grant from the Colombian Department for Science, Technology and Innovation, Colciencias, proj. 1101-521-28465 with funding from "El Patrimonio Autónomo Fondo Nacional de Financiamiento para la Ciencia, la Tecnología y la Innovación, Francisco José de Caldas." The third author recognizes the support from Mexican Government (SNI, COFAA-IPN, SIP 20131702, CONACYT 50206-H) and CONACYT-DST India (proj. 122030 "Answer Validation through Textual Entailment").

2013

zhao-etal-2015-auditory

https://aclanthology.org/Y15-1036

Auditory Synaesthesia and Near Synonyms: A Corpus-Based Analysis of sheng1 and yin1 in Mandarin Chinese

This paper explores the nature of linguistic synaesthesia in the auditory domain through a corpus-based lexical semantic study of near synonyms. It has been established that the near synonyms 聲 sheng "sound" and 音 yin "sound" in Mandarin Chinese have different semantic functions in representing auditory production and auditory perception respectively. Thus, our study is devoted to testing whether linguistic synaesthesia is sensitive to this semantic dichotomy of cognition in particular, and to examining the relationship between linguistic synaesthesia and cognitive modelling in general. Based on the corpus, we find that the near synonyms exhibit both similarities and differences on synaesthesia. The similarities lie in that both 聲 and 音 are productive recipients of synaesthetic transfers, and vision acts as the source domain most frequently. Besides, the differences exist in selective constraints for 聲 and 音 with synaesthetic modifiers as well as syntactic functions of the whole combinations. We propose that the similarities can be explained by the cognitive characteristics of the sound, while the differences are determined by the influence of the semantic dichotomy of production/perception on synaesthesia. Therefore, linguistic synaesthesia is not a random association, but can be motivated and predicted by cognition. 1 The terms, "lower domains" and "higher domains", are copied from Ullmann (1957), where the former refers to touch, taste and smell, and the later includes hearing and vision.

We would like to give thanks to Dennis Tay from the Hong Kong Polytechnic University for his insightful comments on this work.

2015

zheng-etal-2019-boundary

https://aclanthology.org/D19-1034

A Boundary-aware Neural Model for Nested Named Entity Recognition

In natural language processing, it is common that many entities contain other entities inside them. Most existing works on named entity recognition (NER) only deal with flat entities but ignore nested ones. We propose a boundary-aware neural model for nested NER which leverages entity boundaries to predict entity categorical labels. Our model can locate entities precisely by detecting boundaries using sequence labeling models. Based on the detected boundaries, our model utilizes the boundary-relevant regions to predict entity categorical labels, which can decrease computation cost and relieve error propagation problem in layered sequence labeling model. We introduce multitask learning to capture the dependencies of entity boundaries and their categorical labels, which helps to improve the performance of identifying entities. We conduct our experiments on nested NER datasets and the experimental results demonstrate that our model outperforms other state-of-the-art methods.

This work was supported by the Fundamental Research Funds for the Central Universities, SCUT (No. 2017ZD048, D2182480), the Science and Technology Planning Project of Guangdong Province (No.2017B050506004), the Science and Technology Programs of Guangzhou (No. 201704030076,201802010027,201902010046) and a CUHK Research Committee Funding (Direct Grants) (Project Code: EE16963).

2019

fares-etal-2019-arabic

https://aclanthology.org/W19-4626

Arabic Dialect Identification with Deep Learning and Hybrid Frequency Based Features

Studies on Dialectical Arabic are growing more important by the day as it becomes the primary written and spoken form of Arabic online in informal settings. Among the important problems that should be explored is that of dialect identification. This paper reports different techniques that can be applied towards such goal and reports their performance on the Multi Arabic Dialect Applications and Resources (MADAR) Arabic Dialect Corpora. Our results show that improving on traditional systems using frequency based features and non deep learning classifiers is a challenging task. We propose different models based on different word and document representations. Our top model is able to achieve an F1 macro averaged score of 65.66 on MADAR's smallscale parallel corpus of 25 dialects and Modern Standard Arabic (MSA).

2019

hillard-etal-2003-detection

https://aclanthology.org/N03-2012

Detection Of Agreement vs. Disagreement In Meetings: Training With Unlabeled Data

To support summarization of automatically transcribed meetings, we introduce a classifier to recognize agreement or disagreement utterances, utilizing both word-based and prosodic cues. We show that hand-labeling efforts can be minimized by using unsupervised training on a large unlabeled data set combined with supervised training on a small amount of data. For ASR transcripts with over 45% WER, the system recovers nearly 80% of agree/disagree utterances with a confusion rate of only 3%.

Partnership for the goals

This work is supported in part by the NSF under grants 0121396 and 0619921, DARPA grant N660019928924, and NASA grant NCC 2-1256. Any opinions, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of these agencies.

2003

zajac-1999-aspects

https://aclanthology.org/W99-0506

On Some Aspects of Lexical Standardization

In developing and using many large mult~-hngual multt-purpose lexicons at CRL, we ~denttfied three dlstmct problem areas (1) an appropriate metalanguage (formahsm) tot representing and processing lex~cal knowledge (2) a standard generic lex~cal framework defimng a common lex~cal entry structure (names ot features and types ot content), and (3) shared umversal hngu~st~c types In th~s paper, we present the solutions developed at CRL addressing d~mens~ons 1 and 2, and we mention the ongoing research addressing dlmens~on 3

1999

dagan-2009-time

https://aclanthology.org/W09-3701

It's time for a semantic inference engine

A common computational goal is to encapsulate the modeling of a target phenomenon within a unified and comprehensive "engine", which addresses a broad range of the required processing tasks. This goal is followed in common modeling of the morphological and syntactic levels of natural language, where most processing tasks are encapsulated within morphological analyzers and syntactic parsers. In this talk I suggest that computational modeling of the semantic level should also focus on encapsulating the various processing tasks within a unified module (engine). The input/output specification of such engine (API) can be based on the textual entailment paradigm, which will be described in brief and suggested as an attractive framework for applied semantic inference. The talk will illustrate an initial proposal for the engine's API, designed to be embedded within the prominent language processing applications. Finally, I will sketch the entailment formalism and efficient inference algorithm developed at Bar-Ilan University, which illustrates a principled transformational (rather than interpretational) approach towards developing a comprehensive semantic engine.

2009

el-baff-etal-2018-challenge

https://aclanthology.org/K18-1044

Challenge or Empower: Revisiting Argumentation Quality in a News Editorial Corpus

News editorials are said to shape public opinion, which makes them a powerful tool and an important source of political argumentation. However, rarely do editorials change anyone's stance on an issue completely, nor do they tend to argue explicitly (but rather follow a subtle rhetorical strategy). So, what does argumentation quality mean for editorials then? We develop the notion that an effective editorial challenges readers with opposing stance, and at the same time empowers the arguing skills of readers that share the editorial's stance-or even challenges both sides. To study argumentation quality based on this notion, we introduce a new corpus with 1000 editorials from the New York Times, annotated for their perceived effect along with the annotators' political orientations. Analyzing the corpus, we find that annotators with different orientation disagree on the effect significantly. While only 1% of all editorials changed anyone's stance, more than 5% meet our notion. We conclude that our corpus serves as a suitable resource for studying the argumentation quality of news editorials.

Peace, Justice and Strong Institutions

2018

ji-etal-2021-discrete

https://aclanthology.org/2021.naacl-main.431

Discrete Argument Representation Learning for Interactive Argument Pair Identification

In this paper, we focus on identifying interactive argument pairs from two posts with opposite stances to a certain topic. Considering opinions are exchanged from different perspectives of the discussing topic, we study the discrete representations for arguments to capture varying aspects in argumentation languages (e.g., the debate focus and the participant behavior). Moreover, we utilize hierarchical structure to model post-wise information incorporating contextual knowledge. Experimental results on the large-scale dataset collected from CMV show that our proposed framework can significantly outperform the competitive baselines. Further analyses reveal why our model yields superior performance and prove the usefulness of our learned representations.

This work is partially supported by National Natural Science Foundation of China (No.71991471), Science and Technology Commission of Shanghai Municipality Grant (No.20dz1200600). Jing Li is supported by CCF-Tencent Rhino-Bird Young Faculty Open Research Fund (R-ZDCJ), the Hong Kong Polytechnic University internal funds (1-BE2W and 1-ZVRH), and NSFC Young Scientists Fund 62006203.

2021

shen-etal-2006-jhu

https://aclanthology.org/2006.iwslt-evaluation.8

The JHU workshop 2006 IWSLT system

This paper describes the SMT we built during the 2006 JHU Summer Workshop for the IWSLT 2006 evaluation. Our effort focuses on two parts of the speech translation problem: 1) efficient decoding of word lattices and 2) novel applications of factored translation models to IWSLT-specific problems. In this paper, we present results from the open-track Chinese-to-English condition. Improvements of 5-10% relative BLEU are obtained over a high performing baseline. We introduce a new open-source decoder that implements the state-of-the-art in statistical machine translation.

We would like to thank our JHU summer workshop team members (Philipp Koehn, Hieu Hoang, Chris Dyer, Ondrej Bojar, Chris Callison-Burch, Brooke Cowan, Christine Moran, Alexandra Constantin and Evan Herbst) who made this construction of this system possible. We wish to acknowledge their diligent efforts to make the moses decoder stable in a six-week period.We would also like to thank the staff and faculty of CLSP at John's Hopkins University for graciously hosting us during the summer workshop.

2006

ali-etal-2013-hear

https://aclanthology.org/I13-1077

Can I Hear You? Sentiment Analysis on Medical Forums

Text mining studies have started to investigae relations between positive and negative opinions and patients' physical health. Several studies linked the personal lexicon with health and the health-related behavior of the individual. However, few text mining studies were performed to analyze opinions expressed in a large volume of user-written Web content. Our current study focused on performing sentiment analysis on several medical forums dedicated to Hearing Loss (HL). We categorized messages posted on the forums as positive, negative and neutral. Our study had two stages: first, we applied manual annotation of the posts with two annotators and have 82.01% overall agreement with kappa 0.65 and then we applied Machine Learning techniques to classify the posts.

Good Health and Well-Being

This work in part has been funded by a Natural Sciences and Engineering Research Council of Canada Discovery Research Grant and by a Children's Hospital of Eastern Ontario Department of Surgery Research Grant.

2013

strzalkowski-vauthey-1991-fast

https://aclanthology.org/H91-1068

Fast Text Processing for Information Retrieval

We describe an advanced text processing system for information retrieval from natural language document collections. We use both syntactic processing as well as statistical term clustering to obtain a representation of documents which would be more accurate than those obtained with more traditional keyword methods. A reliable top-down parser has been developed that allows for fast processing of large amounts of text, and for a precise identification of desired types of phrases for statistical analysis. Two statistical measures are computed: the measure of informational contribution of words in phrases, and the similarity measure between words.

1991

kokkinakis-gerdin-2009-issues

https://aclanthology.org/W09-4505

Issues on Quality Assessment of SNOMED CT® Subsets -- Term Validation and Term Extraction

The aim of this paper is to apply and develop methods based on Natural Language Processing for automatically testing the validity, reliability and coverage of various Swedish SNOMED-CT subsets, the Systematized NOmenclature of MEDicine-Clinical Terms a multiaxial, hierarchical classification system which is currently being translated from English to Swedish. Our work has been developed across two dimensions. Initially a Swedish electronic text collection of scientific medical documents has been collected and processed to a uniform format. Secondly, a term processing activity has been taken place. In the first phase of this activity, various SNOMED CT subsets have been mapped to the text collection for evaluating the validity and reliability of the translated terms. In parallel, a large number of term candidates have been extracted from the corpus in order to examine the coverage of SNOMED CT. Term candidates that are currently not included in the Swedish SNOMED CT can be either parts of compounds, parts of potential multiword terms, terms that are not yet been translated or potentially new candidates. In order to achieve these goals a number of automatic term recognition algorithms have been applied to the corpus. The results of the later process is to be reviewed by domain experts (relevant to the subsets extracted) through a relevant interface who can decide whether a new set of terms can be incorporated in the Swedish translation of SNOMED CT or not.

We would like to thank the editors of the Journal of the Swedish Medical Association and DiabetologNytt for making the electronic versions available to this study.

2009

xiao-etal-2013-learning

https://aclanthology.org/D13-1016

Learning Latent Word Representations for Domain Adaptation using Supervised Word Clustering

Domain adaptation has been popularly studied on exploiting labeled information from a source domain to learn a prediction model in a target domain. In this paper, we develop a novel representation learning approach to address domain adaptation for text classification with automatically induced discriminative latent features, which are generalizable across domains while informative to the prediction task. Specifically, we propose a hierarchical multinomial Naive Bayes model with latent variables to conduct supervised word clustering on labeled documents from both source and target domains, and then use the produced cluster distribution of each word as its latent feature representation for domain adaptation. We train this latent graphical model using a simple expectation-maximization (EM) algorithm. We empirically evaluate the proposed method with both cross-domain document categorization tasks on Reuters-21578 dataset and cross-domain sentiment classification tasks on Amazon product review dataset. The experimental results demonstrate that our proposed approach achieves superior performance compared with alternative methods.

2013

iida-etal-2010-incorporating

https://aclanthology.org/P10-1128

Incorporating Extra-Linguistic Information into Reference Resolution in Collaborative Task Dialogue

This paper proposes an approach to reference resolution in situated dialogues by exploiting extra-linguistic information. Recently, investigations of referential behaviours involved in situations in the real world have received increasing attention by researchers (Di Eugenio et al., 2000; Byron, 2005; van Deemter, 2007; Spanger et al., 2009). In order to create an accurate reference resolution model, we need to handle extra-linguistic information as well as textual information examined by existing approaches (Soon et al., 2001; Ng and Cardie, 2002, etc.). In this paper, we incorporate extra-linguistic information into an existing corpus-based reference resolution model, and investigate its effects on reference resolution problems within a corpus of Japanese dialogues. The results demonstrate that our proposed model achieves an accuracy of 79.0% for this task.

2010

imamura-2002-application

https://aclanthology.org/2002.tmi-papers.9

Application of translation knowledge acquired by hierarchical phrase alignment for pattern-based MT

Hierarchical phrase alignment is a method for extracting equivalent phrases from bilingual sentences, even though they belong to different language families. The method automatically extracts transfer knowledge from about 125K English and Japanese bilingual sentences and then applies it to a pattern-based MT system. The translation quality is then evaluated. The knowledge needs to be cleaned, since the corpus contains various translations and the phrase alignment contains errors. Various cleaning methods are applied in this paper. The results indicate that when the best cleaning method is used, the knowledge acquired by hierarchical phrase alignment is comparable to manually acquired knowledge.

2002

hua-wang-2017-pilot

https://aclanthology.org/W17-4513

A Pilot Study of Domain Adaptation Effect for Neural Abstractive Summarization

We study the problem of domain adaptation for neural abstractive summarization. We make initial efforts in investigating what information can be transferred to a new domain. Experimental results on news stories and opinion articles indicate that neural summarization model benefits from pre-training based on extractive summaries. We also find that the combination of in-domain and out-of-domain setup yields better summaries when in-domain data is insufficient. Further analysis shows that, the model is capable to select salient content even trained on out-of-domain data, but requires in-domain data to capture the style for a target domain.

This work was supported in part by National Science Foundation Grant IIS-1566382 and a GPU gift from Nvidia. We thank three anonymous reviewers for their valuable suggestions on various aspects of this work.

2017

cotterell-etal-2016-sigmorphon

https://aclanthology.org/W16-2002

The SIGMORPHON 2016 Shared Task---Morphological Reinflection

The 2016 SIGMORPHON Shared Task was devoted to the problem of morphological reinflection. It introduced morphological datasets for 10 languages with diverse typological characteristics. The shared task drew submissions from 9 teams representing 11 institutions reflecting a variety of approaches to addressing supervised learning of reinflection. For the simplest task, inflection generation from lemmas, the best system averaged 95.56% exact-match accuracy across all languages, ranging from Maltese (88.99%) to Hungarian (99.30%). With the relatively large training datasets provided, recurrent neural network architectures consistently performed best-in fact, there was a significant margin between neural and non-neural approaches. The best neural approach, averaged over all tasks and languages, outperformed the best nonneural one by 13.76% absolute; on individual tasks and languages the gap in accuracy sometimes exceeded 60%. Overall, the results show a strong state of the art, and serve as encouragement for future shared tasks that explore morphological analysis and generation with varying degrees of supervision.

2016

wang-etal-2021-easy

https://aclanthology.org/2021.findings-acl.415

As Easy as 1, 2, 3: Behavioural Testing of NMT Systems for Numerical Translation

Mistranslated numbers have the potential to cause serious effects, such as financial loss or medical misinformation. In this work we develop comprehensive assessments of the robustness of neural machine translation systems to numerical text via behavioural testing. We explore a variety of numerical translation capabilities a system is expected to exhibit and design effective test examples to expose system underperformance. We find that numerical mistranslation is a general issue: major commercial systems and state-of-the-art research models fail on many of our test examples, for high-and low-resource languages. Our tests reveal novel errors that have not previously been reported in NMT systems, to the best of our knowledge. Lastly, we discuss strategies to mitigate numerical mistranslation.

We thank all anonymous reviewers for their constructive comments. The authors acknowledge funding support by Facebook.

2021

pulman-1980-parsing

https://aclanthology.org/C80-1009

Parsing and Syntactic Theory

It is argued that many constraints on syntactic rules are a consequence of simple assumptions about parsing mechanisms. If generally true, this suggests an interesting new line of research for syntactic theory.

1980

hajishirzi-etal-2013-joint

https://aclanthology.org/D13-1029

Joint Coreference Resolution and Named-Entity Linking with Multi-Pass Sieves

Many errors in coreference resolution come from semantic mismatches due to inadequate world knowledge. Errors in named-entity linking (NEL), on the other hand, are often caused by superficial modeling of entity context. This paper demonstrates that these two tasks are complementary. We introduce NECO, a new model for named entity linking and coreference resolution, which solves both problems jointly, reducing the errors made on each. NECO extends the Stanford deterministic coreference system by automatically linking mentions to Wikipedia and introducing new NEL-informed mention-merging sieves. Linking improves mention-detection and enables new semantic attributes to be incorporated from Freebase, while coreference provides better context modeling by propagating named-entity links within mention clusters. Experiments show consistent improvements across a number of datasets and experimental conditions, including over 11% reduction in MUC coreference error and nearly 21% reduction in F1 NEL error on ACE 2004 newswire data.

The research was supported in part by grants from DARPA under the DEFT program through the AFRL (FA8750-13-2-0019) and the CSSG (N11AP20020), the ONR (N00014-12-1-0211), and the NSF (IIS-1115966). Support was also provided by a gift from Google, an NSF Graduate Research Fellowship, and the WRF / TJ Cable Professorship. The authors thank Greg Durrett, Heeyoung Lee, Mitchell Koch, Xiao Ling, Mark Yatskar, Kenton Lee, Eunsol Choi, Gabriel Schubiner, Nicholas FitzGerald, Tom Kwiatkowski, and the anonymous reviewers for helpful comments and feedback on the work.

2013

moraes-etal-2014-adapting

https://aclanthology.org/W14-4409

Adapting Graph Summaries to the Users' Reading Levels

Deciding on the complexity of a generated text in NLG systems is a contentious task. Some systems propose the generation of simple text for low-skilled readers; some choose what they anticipate to be a "good measure" of complexity by balancing sentence length and number of sentences (using scales such as the D-level sentence complexity) for the text; while others target high-skilled readers. In this work, we discuss an approach that aims to leverage the experience of the reader when reading generated text by matching the syntactic complexity of the generated text to the reading level of the surrounding text. We propose an approach for sentence aggregation and lexical choice that allows generated summaries of line graphs in multimodal articles available online to match the reading level of the text of the article in which the graphs appear. The technique is developed in the context of the SIGHT (Summarizing Information Graphics Textually) system. This paper tackles the micro planning phase of sentence generation discussing additionally the steps of lexical choice, and pronominalization.

Quality Education

2014

navigli-velardi-2002-automatic

http://www.lrec-conf.org/proceedings/lrec2002/pdf/47.pdf

Automatic Adaptation of WordNet to Domains

2002

calixto-etal-2019-latent

https://aclanthology.org/P19-1642

Latent Variable Model for Multi-modal Translation

In this work, we propose to model the interaction between visual and textual features for multi-modal neural machine translation (MMT) through a latent variable model. This latent variable can be seen as a multi-modal stochastic embedding of an image and its description in a foreign language. It is used in a target-language decoder and also to predict image features. Importantly, our model formulation utilises visual and textual inputs during training but does not require that images be available at test time. We show that our latent variable MMT formulation improves considerably over strong baselines, including a multi-task learning approach (Elliott and Kádár, 2017) and a conditional variational auto-encoder approach (Toyama et al., 2016). Finally, we show improvements due to (i) predicting image features in addition to only conditioning on them, (ii) imposing a constraint on the KL term to promote models with nonnegligible mutual information between inputs and latent variable, and (iii) by training on additional target-language image descriptions (i.e. synthetic data).

This work is supported by the Dutch Organisation for Scientific Research (NWO) VICI Grant nr. 277-89-002.

2019

reynaert-etal-2010-balancing

http://www.lrec-conf.org/proceedings/lrec2010/pdf/549_Paper.pdf

Balancing SoNaR: IPR versus Processing Issues in a 500-Million-Word Written Dutch Reference Corpus

In The Low Countries, a major reference corpus for written Dutch is currently being built. In this paper, we discuss the interplay between data acquisition and data processing during the creation of the SoNaR Corpus. Based on recent developments in traditional corpus compiling and new web harvesting approaches, SoNaR is designed to contain 500 million words, balanced over 36 text types including both traditional and new media texts. Beside its balanced design, every text sample included in SoNaR will have its IPR issues settled to the largest extent possible. This data collection task presents many challenges because every decision taken on the level of text acquisition has ramifications for the level of processing and the general usability of the corpus later on. As far as the traditional text types are concerned, each text brings its own processing requirements and issues. For new media texts-SMS, chat-the problem is even more complex, issues such as anonimity, recognizability and citation right, all present problems that have to be tackled one way or another. The solutions may actually lead to the creation of two corpora: a gigaword SoNaR, IPR-cleared for research purposes, and the smallerof commissioned size-more privacy compliant SoNaR, IPR-cleared for commercial purposes as well.

The SoNaR project is funded by the Nederlandse Taalunie (NTU: Dutch Language Union) within the framework of the STEVIN programme under grant number STE07014. See also http://taalunieversum.org/taal/technologie/stevin/

2010

li-etal-2009-chinese

https://aclanthology.org/W09-0433

Chinese Syntactic Reordering for Adequate Generation of Korean Verbal Phrases in Chinese-to-Korean SMT

Chinese and Korean belong to different language families in terms of word-order and morphological typology. Chinese is an SVO and morphologically poor language while Korean is an SOV and morphologically rich one. In Chinese-to-Korean SMT systems, systematic differences between the verbal systems of the two languages make the generation of Korean verbal phrases difficult. To resolve the difficulties, we address two issues in this paper. The first issue is that the verb position is different from the viewpoint of word-order typology. The second is the difficulty of complex morphology generation of Korean verbs from the viewpoint of morphological typology. We propose a Chinese syntactic reordering that is better at generating Korean verbal phrases in Chinese-to-Korean SMT. Specifically, we consider reordering rules targeting Chinese verb phrases (VPs), preposition phrases (PPs), and modality-bearing words that are closely related to Korean verbal phrases. We verify our system with two corpora of different domains. Our proposed approach significantly improves the performance of our system over a baseline phrased-based SMT system. The relative improvements in the two corpora are +9.32% and +5.43%, respectively.

This work was supported in part by MKE & II-TA through the IT Leading R&D Support Project and also in part by the BK 21 Project in 2009.

2009

yuen-etal-2004-morpheme

https://aclanthology.org/C04-1145

Morpheme-based Derivation of Bipolar Semantic Orientation of Chinese Words

The evaluative character of a word is called its semantic orientation (SO). A positive SO indicates desirability (e.g. Good, Honest) and a negative SO indicates undesirability (e.g., Bad, Ugly). This paper presents a method, based on Turney (2003), for inferring the SO of a word from its statistical association with strongly-polarized words and morphemes in Chinese. It is noted that morphemes are much less numerous than words, and that also a small number of fundamental morphemes may be used in the modified system to great advantage. The algorithm was tested on 1,249 words (604 positive and 645 negative) in a corpus of 34 million words, and was run with 20 and 40 polarized words respectively, giving a high precision (79.96% to 81.05%), but a low recall (45.56% to 59.57%). The algorithm was then run with 20 polarized morphemes, or single characters, in the same corpus, giving a high precision of 80.23% and a high recall of 85.03%. We concluded that morphemes in Chinese, as in any language, constitute a distinct sub-lexical unit which, though small in number, has greater linguistic significance than words, as seen by the significant enhancement of results with a much smaller corpus than that required by Turney.

2004

demirsahin-etal-2020-open

https://aclanthology.org/2020.lrec-1.804

Open-source Multi-speaker Corpora of the English Accents in the British Isles

This paper presents a dataset of transcribed highquality audio of English sentences recorded by volunteers speaking with different accents of the British Isles. The dataset is intended for linguistic analysis as well as use for speech technologies. The recording scripts were curated specifically for accent elicitation, covering a variety of phonological phenomena and providing a high phoneme coverage. The scripts include pronunciations of global locations, major airlines and common personal names in different accents; and native speaker pronunciations of local words. Overlapping lines for all speakers were included for idiolect elicitation, which include the same or similar lines with other existing resources such as the CSTR VCTK corpus and the Speech Accent Archive to allow for easy comparison of personal and regional accents. The resulting corpora include over 31 hours of recordings from 120 volunteers who selfidentify as native speakers of Southern England, Midlands, Northern England, Welsh, Scottish and Irish varieties of English.

The authors would like to thank Dawn Knight, Anna Jones and Alex Thomas from Cardiff University for their assis tance in collecting the Welsh English data presented in this paper. The authors also thank Richard Sproat for his com ments on the earlier drafts of this paper. Finally, the authors thank the anonymous reviewers for many helpful sugges tions.

2020

kruengkrai-etal-2021-multi

https://aclanthology.org/2021.findings-acl.217

A Multi-Level Attention Model for Evidence-Based Fact Checking

Evidence-based fact checking aims to verify the truthfulness of a claim against evidence extracted from textual sources. Learning a representation that effectively captures relations between a claim and evidence can be challenging. Recent state-of-the-art approaches have developed increasingly sophisticated models based on graph structures. We present a simple model that can be trained on sequence structures. Our model enables inter-sentence attentions at different levels and can benefit from joint training. Results on a large-scale dataset for Fact Extraction and VERification (FEVER) show that our model outperforms the graphbased approaches and yields 1.09% and 1.42% improvements in label accuracy and FEVER score, respectively, over the best published model. 1

Peace, Justice and Strong Institutions

We thank Erica Cooper (NII) for providing valuable feedback on an earlier draft of this paper. This work is supported by JST CREST Grants (JPMJCR18A6 and JPMJCR20D3) and MEXT KAKENHI Grants (21H04906), Japan.

2021

liu-etal-2021-self

https://aclanthology.org/2021.naacl-main.334

Self-Alignment Pretraining for Biomedical Entity Representations

Despite the widespread success of selfsupervised learning via masked language models (MLM), accurately capturing fine-grained semantic relationships in the biomedical domain remains a challenge. This is of paramount importance for entity-level tasks such as entity linking where the ability to model entity relations (especially synonymy) is pivotal. To address this challenge, we propose SAPBERT, a pretraining scheme that selfaligns the representation space of biomedical entities. We design a scalable metric learning framework that can leverage UMLS, a massive collection of biomedical ontologies with 4M+ concepts. In contrast with previous pipelinebased hybrid systems, SAPBERT offers an elegant one-model-for-all solution to the problem of medical entity linking (MEL), achieving a new state-of-the-art (SOTA) on six MEL benchmarking datasets. In the scientific domain, we achieve SOTA even without taskspecific supervision. With substantial improvement over various domain-specific pretrained MLMs such as BIOBERT, SCIBERT and PUB-MEDBERT, our pretraining scheme proves to be both effective and robust. 1

Good Health and Well-Being

We thank the three reviewers and the Area Chair for their insightful comments and suggestions. FL is supported by Grace & Thomas C.H. Chan Cambridge Scholarship. NC and MB would like to

2021

liu-emerson-2022-learning

https://aclanthology.org/2022.acl-long.275

Learning Functional Distributional Semantics with Visual Data

Functional Distributional Semantics is a recently proposed framework for learning distributional semantics that provides linguistic interpretability. It models the meaning of a word as a binary classifier rather than a numerical vector. In this work, we propose a method to train a Functional Distributional Semantics model with grounded visual data. We train it on the Visual Genome dataset, which is closer to the kind of data encountered in human language acquisition than a large text corpus. On four external evaluation datasets, our model outperforms previous work on learning semantics from Visual Genome. 1

2022

pultrova-2019-correlation

https://aclanthology.org/W19-8504

Correlation between the gradability of Latin adjectives and the ability to form qualitative abstract nouns

Comparison is distinctly limited in scope among grammatical categories in that it is unable, for semantic reasons, to produce comparative and superlative forms for many representatives of the word class to which it applies as a category (adjectives and their derived adverbs). In Latin and other dead languages, it is nontrivial to decide with certainty whether an adjective is gradable or not: being non-native speakers, we cannot rely on linguistic intuition; nor can a definitive answer be reached by consulting the corpus of Latin texts (the fact that an item is not attested in the surviving corpus obviously does not mean that it did not exist in Latin). What needs to be found are properties of adjectives correlated with gradability/ nongradability that are directly discernible at the level of written language. The present contribution gives one such property, showing that there is a strong correlation between gradability and the ability of an adjective to form abstract nouns. 1 Comparison: conceptual vs grammatical category Comparison is a grammatical category that has for a long time practically escaped the attention of linguists studying Latin. Only relatively recently were detailed studies published on the phenomenon of comparison on a cognitive and functional basis, 1 investigating how two or more entities could be compared in a language, what patterns are used in these various ways of comparison in Latin, and what different meanings comparatives and superlatives may have. These studies clearly demonstratewhich is true in other languages as wellthat it does not hold that comparison in Latin is always carried out using the forms of comparative and superlative, nor does it hold that comparatives and superlatives always perform the basic function of simple comparison of two or more entities. It follows that it is useful, even necessary, as with other grammatical categories, to differentiate between comparison on the one hand as a conceptual category that is expressed at the level of the whole proposition ("Paul is higher than John" = "John is not as high as Paul"), and on the other hand comparison as a grammatical/morphological category ("the formal modification of some predicative wordmost often an adjective-representing a parameter of gradation or comparison" 2). The present author is currently working on a monograph that examines the morphological category of Latin comparison. Put simply, she does not ask which means may be employed in Latin to express comparison, but how the forms of comparative and superlative are used. The present contribution deals with one question falling within the scope of this work. 2 Specific nature of category of comparison The grammatical category of comparison is distinctly limited, not being able to produce the forms of comparative and superlative from all the representatives of the word class to which it applies as a category (i.e. adjectives and their derived adverbs). A certain degree of limitation is not exceptional in itself (e.g. in the category of number there are singularia tantum and pluralia tantum; in the category of verb voice, intransitive verbs, for instance, cannot form personal passive forms; etc.); however, comparison is restricted to an exceptional degree. For example, according to the Czech National Corpus,

2019

webber-di-eugenio-1990-free

https://aclanthology.org/C90-2068

Free Adjuncts in Natural Language Instructions

In thi,~ paper, we give a brief account of our project Animation from Instructions, the view of instructions it reflects, and the semantics of one construction-the free adjunct-that is common in Natural Language instructions. *We thank Mark Steedman, Hans Karlgren and Breck Baldwin for comments and advice. They are not to blame for any er-~-ors in the translation of their advice into the present form. The ,:esem'ch was supported by DARPA grant no. N0014-85-K0018, and ARO grant no. DAAL03-89-C0031. 1Tiffs is not to suggest that animation can be driven solely from that common representation: other types of knowledge axe clearly needed as well-including knowledge of motor skills and other performance characteristics.

1990

watanabe-sumita-2002-bidirectional

https://aclanthology.org/C02-1050

Bidirectional Decoding for Statistical Machine Translation

This paper describes the right-to-left decoding method, which translates an input string by generating in right-to-left direction. In addition, presented is the bidirectional decoding method, that can take both of the advantages of left-to-right and right-to-left decoding method by generating output in both ways and by merging hypothesized partial outputs of two directions. The experimental results on Japanese and English translation showed that the right-to-left was better for Englith-to-Japanese translation, while the left-to-right was suitable for Japanese-to-English translation. It was also observed that the bidirectional method was better for English-to-Japanese translation.

The research reported here was supported in part by a contract with the Telecommunications Advancement Organization of Japan entitled, "A study of speech dialogue translation technology based on a large corpus".

2002

lee-1999-spoken

https://aclanthology.org/Y99-1019

Spoken Language Systems - Technical Challenges for Speech and Natural Language Processing

Speech is the most natural means of communication among humans. It is also believed that spoken language processing will play a major role in establishing a universal interface between humans and machines. Most of the existing spoken language systems are rather primitive. For example, speech synthesizers for reading unrestrict text of any language is only producing machine-sounding speech. Automatic speech recognizers are capable of recognizing spoken language from a selective population doing a highly restricted task. In this talk, we present some examples of spoken language translation and dialogue systems and examine the capabilities and limitations of current spoken language technologies. We also discuss technical challenges for language researchers to help realize the vision of natural human-machine communication to allow humans to converse with machines in any language to access information and solve problems.

1999

peng-hu-2016-web

https://aclanthology.org/2016.amta-users.4

Web App UI Layout Sniffer

layout doesn't work. ciency is painfully low. Consequently, it dramatically slows down product delivery in today's

2016

wang-cardie-2012-focused

https://aclanthology.org/W12-1642

Focused Meeting Summarization via Unsupervised Relation Extraction

We present a novel unsupervised framework for focused meeting summarization that views the problem as an instance of relation extraction. We adapt an existing in-domain relation learner (Chen et al., 2011) by exploiting a set of task-specific constraints and features. We evaluate the approach on a decision summarization task and show that it outperforms unsupervised utterance-level extractive summarization baselines as well as an existing generic relation-extraction-based summarization method. Moreover, our approach produces summaries competitive with those generated by supervised methods in terms of the standard ROUGE score.

Acknowledgments This work was supported in part by National Science Foundation Grants IIS-0968450 and IIS-1111176, and by a gift from Google.

2012

ritchie-etal-2006-find

https://aclanthology.org/W06-0804

How to Find Better Index Terms Through Citations

We consider the question of how information from the textual context of citations in scientific papers could improve indexing of the cited papers. We first present examples which show that the context should in principle provide better and new index terms. We then discuss linguistic phenomena around citations and which type of processing would improve the automatic determination of the right context. We present a case study, studying the effect of combining the existing index terms of a paper with additional terms from papers citing that paper in our corpus. Finally, we discuss the need for experimentation for the practical validation of our claim.

2006

lee-etal-2017-ntnu

https://aclanthology.org/S17-2165

The NTNU System at SemEval-2017 Task 10: Extracting Keyphrases and Relations from Scientific Publications Using Multiple Conditional Random Fields

This study describes the design of the NTNU system for the ScienceIE task at the SemEval 2017 workshop. We use self-defined feature templates and multiple conditional random fields with extracted features to identify keyphrases along with categorized labels and their relations from scientific publications. A total of 16 teams participated in evaluation scenario 1 (subtasks A, B, and C), with only 7 teams competing in all subtasks. Our best micro-averaging F1 across the three subtasks is 0.23, ranking in the middle among all 16 submissions.

Industry, Innovation and Infrastructure

This study was partially supported by the Ministry of Science and Technology, under the grant MOST 105-2221-E-003-020-MY2 and the "Aim for the Top University Project" and "Center of Learning Technology for Chinese" of National Taiwan Normal University, sponsored by the Ministry of Education, Taiwan.

2017

polisciuc-etal-2015-understanding

https://aclanthology.org/W15-2810

Understanding Urban Land Use through the Visualization of Points of Interest

Semantic data regarding points of interest in urban areas are hard to visualize. Due to the high number of points and categories they belong, as well as the associated textual information, maps become heavily cluttered and hard to read. Using traditional visualization techniques (e.g. dot distribution maps, typographic maps) partially solve this problem. Although, these techniques address different issues of the problem, their combination is hard and typically results in an efficient visualization. In our approach, we present a method to represent clusters of points of interest as shapes, which is based on vacuum package metaphor. The calculated shapes characterize sets of points and allow their use as containers for textual information. Additionally, we present a strategy for placing text onto polygons. The suggested method can be used in interactive visual exploration of semantic data distributed in space, and for creating maps with similar characteristics of dot distribution maps, but using shapes instead of points.

Sustainable Cities and Communities

This work was supported by the InfoCrowds project -FCT-PTDC/ECM-TRA/1898/2012FCT.

2015

rothlisberger-2002-cls

https://aclanthology.org/2002.tc-1.11

CLS Workflow - a translation workflow system

As the translation industry is faced with ever more challenging deadlines to meet and production costs to keep under tight control, translation companies need to find efficient ways of managing their work processes. CLS Corporate Language Services AG, a translation provider for the financial services and telecoms industries has tackled this issue by developing their own workflow application based on Lotus Notes. The system's various modules have been developed and enhanced over the last four years. Current enhancement projects include interfaces to the accounting tool used in the company, web-based information systems for clients and translation providers and the close integration of some of the CAT tools the company uses.

2002

dobrovoljc-etal-2019-improving

https://aclanthology.org/W19-8004

Improving UD processing via satellite resources for morphology

This paper presents the conversion of the reference language resources for Croatian and Slovenian morphology processing to UD morphological specifications. We show that the newly available training corpora and inflectional dictionaries improve the baseline stanfordnlp performance obtained on officially released UD datasets for lemmatization, morphology prediction and dependency parsing, illustrating the potential value of such satellite UD resources for languages with rich morphology.

The authors acknowledge the financial support from the Slovenian Research Agency through the research core funding no. P6-0411 (Language resources and technologies for Slovene language), the research project no. J6-8256 (New grammar of contemporary standard Slovene: sources and methods) and the Slovenian research infrastructure CLARIN.SI.

2019

wolf-sonkin-etal-2018-structured

https://aclanthology.org/P18-1245

A Structured Variational Autoencoder for Contextual Morphological Inflection

Statistical morphological inflectors are typically trained on fully supervised, type-level data. One remaining open research question is the following: How can we effectively exploit raw, token-level data to improve their performance? To this end, we introduce a novel generative latent-variable model for the semi-supervised learning of inflection generation. To enable posterior inference over the latent variables, we derive an efficient variational inference procedure based on the wake-sleep algorithm. We experiment on 23 languages, using the Universal Dependencies corpora in a simulated low-resource setting, and find improvements of over 10% absolute accuracy in some cases.

2018

ahmed-etal-2020-multilingual

https://aclanthology.org/2020.lrec-1.516

Multilingual Corpus Creation for Multilingual Semantic Similarity Task

In natural language processing, the performance of a semantic similarity task relies heavily on the availability of a large corpus. Various monolingual corpora are available (mainly English); but multilingual resources are very limited. In this work, we describe a semiautomated framework to create a multilingual corpus which can be used for the multilingual semantic similarity task. The similar sentence pairs are obtained by crawling bilingual websites, whereas the dissimilar sentence pairs are selected by applying topic modeling and an Open-AI GPT model on the similar sentence pairs. We focus on websites in the government, insurance, and banking domains to collect English-French and English-Spanish sentence pairs; however, this corpus creation approach can be applied to any other industry vertical provided that a bilingual website exists. We also show experimental results for multilingual semantic similarity to verify the quality of the corpus and demonstrate its usage.

This research was supported by Mitacs through the Mitacs Accelerate program. We also acknowledge the helpful comments provided by the reviewers.

2020

hermann-etal-2012-unsupervised

https://aclanthology.org/S12-1021

An Unsupervised Ranking Model for Noun-Noun Compositionality

We propose an unsupervised system that learns continuous degrees of lexicality for noun-noun compounds, beating a strong baseline on several tasks. We demonstrate that the distributional representations of compounds and their parts can be used to learn a finegrained representation of semantic contribution. Finally, we argue such a representation captures compositionality better than the current status-quo which treats compositionality as a binary classification problem.

The authors would like to acknowledge the use of the Oxford Supercomputing Centre (OSC) in carrying out this work.

2012

babych-etal-2009-evaluation

https://aclanthology.org/2009.eamt-1.6

Evaluation-Guided Pre-Editing of Source Text: Improving MT-Tractability of Light Verb Constructions

This paper reports an experiment on evaluating and improving MT quality of light-verb construction (LVCs)-combinations of a 'semantically depleted' verb and its complement. Our method uses construction-level human evaluation for systematic discovery of mistranslated contexts and creating automatic pre-editing rules, which make the constructions more tractable for Rule-Based Machine Translation (RBMT) systems. For rewritten phrases we achieve about 40% reduction in the number of incomprehensible translations into English from both French and Russian. The proposed method can be used for enhancing automatic pre-editing functionality of state-of-theart MT systems. It will allow MT users to create their own rewriting rules for frequently mistranslated constructions and contexts, going beyond existing systems' capabilities offered by user dictionaries and do-not translate lists.

2009

pan-etal-2019-twitter

https://aclanthology.org/P19-1252

Twitter Homophily: Network Based Prediction of User's Occupation

In this paper, we investigate the importance of social network information compared to content information in the prediction of a Twitter user's occupational class. We show that the content information of a user's tweets, the profile descriptions of a user's follower/following community, and the user's social network provide useful information for classifying a user's occupational group. In our study, we extend an existing dataset for this problem, and we achieve significantly better performance by using social network homophily that has not been fully exploited in previous work. In our analysis, we found that by using the graph convolutional network to exploit social homophily, we can achieve competitive performance on this dataset with just a small fraction of the training data.

We would like to thank the reviewers for their helpful comments on our work. This work is supported by DSO grant DSOCL17061.

2019

nguyen-chiang-2017-transfer

https://aclanthology.org/I17-2050

Transfer Learning across Low-Resource, Related Languages for Neural Machine Translation

We present a simple method to improve neural translation of a low-resource language pair using parallel data from a related, also low-resource, language pair. The method is based on the transfer method of Zoph et al., but whereas their method ignores any source vocabulary overlap, ours exploits it. First, we split words using Byte Pair Encoding (BPE) to increase vocabulary overlap. Then, we train a model on the first language pair and transfer its parameters, including its source word embeddings, to another model and continue training on the second language pair. Our experiments show that transfer learning helps word-based translation only slightly, but when used on top of a much stronger BPE baseline, it yields larger improvements of up to 4.3 BLEU.

This research was supported in part by University of Southern California subcontract 67108176 under DARPA contract HR0011-15-C-0115. Nguyen was supported by a fellowship from the Vietnam Education Foundation. We would like to express our great appreciation to Dr. Sharon Hu for letting us use her group's GPU cluster (supported by NSF award 1629914), and to NVIDIA corporation for the donation of a Titan X GPU.

2017

kochmar-shutova-2017-modelling

https://aclanthology.org/W17-5033

Modelling semantic acquisition in second language learning

Using methods of statistical analysis, we investigate how semantic knowledge is acquired in English as a second language and evaluate the pace of development across a number of predicate types and content word combinations, as well as across the levels of language proficiency and native languages. Our exploratory study helps identify the most problematic areas for language learners with different backgrounds and at different stages of learning.

Quality Education

We are grateful to the BEA reviewers for their helpful and instructive feedback. Ekaterina Kochmar's research is supported by Cambridge English Language Assessment via the ALTA Institute. Ekaterina Shutova's research is supported by the Leverhulme Trust Early Career Fellowship.

2017

alfonseca-etal-2013-heady

https://aclanthology.org/P13-1122

HEADY: News headline abstraction through event pattern clustering

This paper presents HEADY: a novel, abstractive approach for headline generation from news collections. From a web-scale corpus of English news, we mine syntactic patterns that a Noisy-OR model generalizes into event descriptions. At inference time, we query the model with the patterns observed in an unseen news collection, identify the event that better captures the gist of the collection and retrieve the most appropriate pattern to generate a headline. HEADY improves over a state-of-theart open-domain title abstraction method, bridging half of the gap that separates it from extractive methods using humangenerated titles in manual evaluations, and performs comparably to human-generated headlines as evaluated with ROUGE.

The research leading to these results has received funding from: the EU's 7 th Framework Programme (FP7/2007-2013) under grant agreement number 257790; the Spanish Ministry of Science and Innovation's project Holopedia (TIN2010-21128-C02); and the Regional Government of Madrid's MA2VICMR (S2009/TIC1542). We would like to thank Katja Filippova and the anonymous reviewers for their insightful comments.

2013

gupta-etal-2014-text

https://aclanthology.org/S14-1010

Text Summarization through Entailment-based Minimum Vertex Cover

Sentence Connectivity is a textual characteristic that may be incorporated intelligently for the selection of sentences of a well meaning summary. However, the existing summarization methods do not utilize its potential fully. The present paper introduces a novel method for singledocument text summarization. It poses the text summarization task as an optimization problem, and attempts to solve it using Weighted Minimum Vertex Cover (WMVC), a graph-based algorithm. Textual entailment, an established indicator of semantic relationships between text units, is used to measure sentence connectivity and construct the graph on which WMVC operates. Experiments on a standard summarization dataset show that the suggested algorithm outperforms related methods.

2014

wilson-wiebe-2003-annotating

https://aclanthology.org/W03-2102

Annotating Opinions in the World Press

In this paper we present a detailed scheme for annotating expressions of opinions, beliefs, emotions, sentiment and speculation (private states) in the news and other discourse. We explore inter-annotator agreement for individual private state expressions, and show that these low-level annotations are useful for producing higher-level subjective sentence annotations.

Peace, Justice and Strong Institutions

2003