Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Raymond J. MooneySuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works212 from public data
- A Machine Learning Approach to Coreference Resolution of Noun Phrases1,146
The learning approach to coreference resolution of noun phrases in unrestricted text is presented, indicating that on the general noun phrase coreference task, the learning approach holds promise and achieves accuracy comparable to that of nonlearning approaches.
- The CoNLL-2014 Shared Task on Grammatical Error Correction592
The task definition is given, the data sets are presented, and the evaluation metric and scorer used in the shared task are described, to give an overview of the various approaches adopted by the participating teams, and present the evaluation results.
- Feature selection, perception learning, and a usability case study for text categorization588
An automated learning approach to text categorization based on perception learning and a new feature selection metric, called correlation coefficient, is described and empirical results indicate that this approach outperforms the best published results on this % uters collection.
- A Neural Approach to Automated Essay Scoring570
This paper develops an approach based on recurrent neural networks to learn the relation between an essay and its assigned score, without any feature engineering.
- Integrating multiple knowledge sources to disambiguate word sense569
This approach integrates a diverse set of knowledge sources to disambiguate word sense, including part of speech of neighboring words, morphological form, the unordered set of surrounding words, local collocations, and verb-object syntactic relation.
- 554
- Towards Robust Linguistic Analysis using OntoNotes521
An analysis of the performance of publicly available, state-of-the-art tools on all layers and languages in the OntoNotes v5.0 corpus should set the benchmark for future development of various NLP components in syntax and semantics, and possibly encourage research towards an integrated system that makes use of the various layers jointly to improve overall performance.
- An Unsupervised Neural Attention Model for Aspect Extraction431
A novel neural approach that improves coherence by exploiting the distribution of word co-occurrences through the use of neural word embeddings, and uses an attention mechanism to de-emphasize irrelevant words during training, further improving the coherence of aspects.
- Better Evaluation for Grammatical Error Correction420
This work presents a novel method for evaluating grammatical error correction that is an algorithm for efficiently computing the sequence of phrase-level edits between a source sentence and a system hypothesis that achieves the highest overlap with the gold-standard annotation.
- Building a Large Annotated Corpus of Learner English: The NUS Corpus of Learner English407
The annotation schema and the data collection and annotation process of NUCLE are described and an unpublished study of annotator agreement for grammatical error correction is reported on.
- The CoNLL-2013 Shared Task on Grammatical Error Correction357
The CoNLL-2013 shared task was devoted to grammatical error correction and the task definition is given, the data sets are presented, and the evaluation metric and scorer used are described.
- It Makes Sense: A Wide-Coverage Word Sense Disambiguation System for Free Text355
The flexible framework of IMS allows users to integrate different preprocessing tools, additional features, and different classifiers, and it achieves state-of-the-art results on several SensEval and SemEval tasks.
- Named entity recognition337
It is shown that the maximum entropy framework is able to make use of global information directly, and achieves performance that is comparable to the best previous machine learning-based NERs on M UC-6 and MUC-7 test data.
- Word Sense Disambiguation Improves Statistical Machine Translation329
It is shown for the first time that integrating a WSD system improves the performance of a state-of-the-art statistical MT system on an actual translation task, and the improvement is statistically significant.
- Recognizing implicit discourse relations in the Penn Discourse Treebank313
An implicit discourse relation classifier is presented in the Penn Discourse Treebank that considers the context of the two arguments, word pair information, as well as the arguments' internal constituent and dependency parses.
- A PDTB-styled end-to-end discourse parser302
This work designed and developed an end-to-end discourse parser-to-parse free texts in the PDTB style in a fully data-driven approach and proposes and presents a comprehensive evaluation from both component-wise and error-cascading perspectives.
- An empirical evaluation of knowledge sources and learning algorithms for word sense disambiguation286
Evaluated knowledge sources include the part-of-speech of neighboring words, single words in the surrounding context, local collocations, and syntactic relations, and the SVM, Naive Bayes, AdaBoost, and decision tree algorithms.
- An Interactive Multi-Task Learning Network for End-to-End Aspect-Based Sentiment Analysis282
An interactive multi-task learning network (IMN) is proposed which is able to jointly learn multiple related tasks simultaneously at both the token level as well as the document level and introduces a message passing architecture where information is iteratively passed to different tasks through a shared set of latent variables.
- Effective Modeling of Encoder-Decoder Architecture for Joint Entity and Relation Extraction279
A representation scheme for relation tuples which enables the decoder to generate one word at a time like machine translation models and still finds all the tuples present in a sentence with full entity names of different length and with overlapping entities.
- Named entity recognition with a maximum entropy approach275
The named entity recognition (NER) task involves identifying noun phrases that are names, and assigning a class to each name.
- A Multilayer Convolutional Encoder-Decoder Neural Network for Grammatical Error Correction239
By ensembling multiple models, and incorporating an N-gram language model and edit features via rescoring, this novel method becomes the first neural approach to outperform the current state-of-the-art statistical machine translation-based approach, both in terms of grammaticality and fluency.
- Exploiting Document Knowledge for Aspect-level Sentiment Classification220
Two approaches that transfer knowledge from document-level data, which is much less expensive to obtain, are explored to improve the performance of aspect-level sentiment classification.
- Effective Attention Modeling for Aspect-Level Sentiment Classification212
This work proposes a method for target representation that better captures the semantic meaning of the opinion target and introduces an attention model that incorporates syntactic information into the attention mechanism.
- Chinese Part-of-Speech Tagging: One-at-a-Time or All-at-Once? Word-Based or Character-Based?199
An in-depth study on issues of processing architecture and feature representation for Chinese POS tagging, within a maximum entropy framework, and builds a state-of-the-art Chinese word segmenter, which outperforms the best SIGHAN 2003 word segmenters in the closed track on 3 out of 4 test corpora.
- Exploiting parallel texts for word sense disambiguation182
This paper evaluates an approach to automatically acquire sense-tagged training data from English-Chinese parallel corpora, which are then used for disambiguating the nouns in the SENSEVAL-2 English lexical sample task.
- Mining topic-specific concepts and definitions on the web180
The goal is to help people learn in-depth knowledge of a topic systematically on the Web, and the proposed techniques first identify those sub-topics or salient concepts of the topic, and then find and organize those informative pages, containing definitions and descriptions of thetopic and sub- topics, just like those in a book.
- A Maximum Entropy Approach to Chinese Word Segmentation176
This work evaluated the Chinese word segmenter in the open track, on all four corpora, namely Academia Sinica, City University of Hong Kong, Microsoft Research, and Peking University, and achieved the highest F measure for AS, CITYU, and PKU.
- Flexible Domain Adaptation for Automated Essay Scoring Using Correlated Linear Regression175
This work proposes domain adaptation as a solution to adapt an AES system from an initial prompt to a new prompt and proposes a novel domain adaptation technique that uses Bayesian linear ridge regression.
- Automatically Evaluating Text Coherence Using Discourse Relations166
A novel model is presented to represent and assess the discourse coherence of text that assumes that coherent text implicitly favors certain types of discourse relation transitions and is synergistic with the previous approach, demonstrating an error reduction of 73% when the features from both models are combined for the task.
- Bayesian online classifiers for text classification and filtering160
Empirical results indicate that these Bayesian online classifiers are comparable with the best text classification systems and the online approach offers the advantage of continuous learning in the batch-adaptive text filtering task.
- A generative model for parsing natural language to meaning representations158
The generative model is applied to the task of mapping sentences to hierarchical representations of their underlying meaning and achieves state-of-the-art performance when tested on two publicly available corpora.
- The CoNLL-2015 Shared Task on Shallow Discourse Parsing155
The CoNLL-2015 Shared Task is on Shallow Discourse Parsing, a task focusing on identifying individual discourse relations that are present in a natural language text, and the evaluation protocol and metric used during this shared task is presented.
- Grammatical Error Correction: A Survey of the State of the Art152
The field is condense into a single article and some of the linguistic challenges of the task are outlined, the most popular datasets that are available to researchers are introduced, and the various methods and techniques that have been developed with a particular focus on artificial error generation are summarized.
- 152
- Improving the Robustness of Question Answering Systems to Question Paraphrasing151
This work proposes a data augmentation approach that requires no human intervention to re-train the models for improved robustness to question paraphrasing and uses a neural paraphrase model trained to generate multiple paraphrased questions for a given source question and a set of paraphrase suggestions.
- Better Punctuation Prediction with Dynamic Conditional Random Fields144
Empirical results show that the proposed approach is designed to jointly perform both sentence boundary and sentence type prediction, and punctuation prediction on speech utterances, outperforms an approach based on linear-chain conditional random fields and other previous approaches.
- 143
- CoNLL 2016 Shared Task on Multilingual Shallow Discourse Parsing139
The task definition, the training and test sets, and the evaluation protocol and metric used during the CoNLL-2016 Shared Task are presented, which will serve as a benchmark for future research on shallow discourse parsing.
- Revisiting DocRED - Addressing the False Negative Problem in Relation Extraction136
To address the shortcoming of the incomplete annotation of DocRED, this work re-annotates 4,053 documents in the DocRED dataset by adding the missed relation triples back to the original DocRED.
- Supervised Word Sense Disambiguation with Support Vector Machines and multiple knowledge sources133
The knowledge sources used were part-of-speech of neighboring words, single words in the surrounding context, local collocations, and syntactic relations for the translation and sense subtask of the SENSEVAL-3 English lexical sample task.
- Domain Adaptation with Active Learning for Word Sense Disambiguation129
By using the predominant sense predicted by expectation-maximization (EM) and adopting a count-merging technique, this paper improves the effectiveness of the original adaptation process achieved by the basic active learning approach.
- 120
- Semi-Supervised Word Sense Disambiguation Using Word Embeddings in General and Specific Domains114
Two ways of incorporating word embeddings in a word sense disambiguation setting are investigated and the obtained results show that such representations consistently improve the accuracy of the selected supervised WSD system.
- Identification and Resolution of Chinese Zero Pronouns: A Machine Learning Approach110
This work is the first to perform both identification and resolution of Chinese anaphoric zero pronouns using a machine learning approach, with two sets of easily computable features.
- Improved statistical machine translation for resource-poor languages using related resource-rich languages104
A novel language-independent approach for improving statistical machine translation for resource-poor languages by exploiting their similarity to resource-rich ones while using much less additional data.
- Getting Serious about Word Sense Disambiguation104
It is argued that a large, human sensetagged corpus is also critical as well as necessary to achieve broad coverage, high accuracy word sense disambiguation, where the sense distinction is at the level of a good desk-top dictionary such as WORDNET.
- Word Sense Disambiguation Improves Information Retrieval102
This paper proposes a method to estimate sense distributions for short queries and proposes a novel approach to incorporate word senses into the language modeling approach to IR and also exploit the integration of synonym relations.
- Mining new word translations from comparable corpora100
This paper presents a new approach to mining new word translations from comparable corpora, by using context information to complement transliteration information in Chinese and English Gigaword corpora.
- One Million Sense-Tagged Instances for Word Sense Disambiguation and Induction95
It is shown that the open source IMS WSD system trained on the dataset achieves stateof-the-art results in standard disambiguation tasks and a recent word sense induction task, outperforming several task submissions and strong baselines.
- Improved Word Sense Disambiguation Using Pre-Trained Contextualized Word Representations93
Different strategies of integrating pre-trained contextualized word representations are explored and the best strategy achieves accuracies exceeding the best prior published accuracies by significant margins on multiple benchmark WSD datasets.
- MAXSIM: A Maximum Similarity Metric for Machine Translation Evaluation93
An automatic machine translation evaluation metric that calculates a similarity score (based on precision and recall) of a pair of sentences that achieves higher correlation with human judgements than all 11 automatic MT evaluation metrics that were evaluated during the workshop.
- How Far are We from Fully Automatic High Quality Grammatical Error Correction?91
It is concluded that inter-annotator agreement statistics in grammatical error correction are less informative in fields where there may be more than one correct answer and a new metric is proposed based on the ratio between human and system performance.
- A Case Study on Inter-Annotator Agreement for Word Sense Disambiguation91
This paper examines th~s msue by comparing the agreement rate on a large corpus of more than 30,000 sense-tagged instances of the WORDNET Semcor corpus and the DSO corpus, which has been independently tagged by two separate groups of human annotators.
- NUS-PT: exploiting parallel texts for word sense disambiguation in the English all-words tasks89
A supervised learning approach with SVM as the learning algorithm was used, using training examples from English-Chinese parallel corpora, SemCor, and DSO corpus to train the system employed for the fine-grained English all-words task.
- Learning to recognize tables in free text88
A new approach that learns to recognize tables in free text, including the boundary, rows and columns of tables, outperforms a deterministic table recognition algorithm that identifies tables based on a fixed set of conditions.
- A Beam-Search Decoder for Grammatical Error Correction87
A novel beam-search decoder for grammatical error correction that is able to perform correction of whole sentences with multiple and interacting errors while still taking advantage of powerful existing classifier approaches.
- Grammatical Error Correction with Alternating Structure Optimization85
The NUS Corpus of Learner English (NUCLE), a fully annotated one million words corpus of learner English available for research purposes, is introduced and a novel approach to grammatical error correction based on Alternating Structure Optimization is presented.
- Estimating class priors in domain adaptation for word sense disambiguation84
By using well calibrated probabilities, this paper is able to estimate the sense priors of words drawn from a new domain effectively to achieve significant improvements in WSD accuracy.
- Word sense disambiguation with distribution estimation83
Novel application of two distribution estimation algorithms are presented to provide estimates of the sense distribution of the new domain data set to achieve a relative improvement of 56% when incorporated into a word sense disambiguation system.
- 78
- 74
- Correcting Semantic Collocation Errors with L1-induced Paraphrases73
This work presents a novel approach for automatic collocation error correction in learner English which is based on paraphrases extracted from parallel corpora based on the key assumption that collocation errors are often caused by semantic similarity in the first language (L1-language) of the writer.
- Closing the gap71
This is the first research to have demonstrated that a learning approach to the full-scale information extraction task could achieve performance rivaling that of the knowledge engineering approach.
- 67
- 65
- On the role of coherence in abductive explanation64
Some problems encountered using abduction to understand text are described, and some solutions to overcome these problems are presented, around the use of a different criterion, called explanatory coherence, as the primary measure to evaluate the quality of an explanation.
- Adaptive Semi-supervised Learning for Cross-domain Sentiment Classification63
This work considers the cross-domain sentiment classification problem, where a sentiment classifier is to be learned from a source domain and to be generalized to a target domain, and explicitly minimizes the distance between source and target instances in an embedded feature space.
- Neural Quality Estimation of Grammatical Error Correction63
This work proposes the first neural approach to automatic quality estimation of GEC output sentences that does not employ any hand-crafted features, and shows that a state-of-the-art GEC system can be improved when quality scores are used as features for re-ranking the N-best candidates.
- Scaling up word sense disambiguation via parallel texts63
The approach of automatically gathering training examples from parallel texts is scalable to a large set of nouns and achieves accuracy comparable to the best system of SENSEVAL-2 English all-words task, and significantly outperforms the baseline of always choosing sense 1 of WordNet.
- A lattice-based approach to query-by-example spoken document retrieval62
Experimental results on a speech corpus of conversational English show that the use of statistics from lattices for both documents and query exemplars results in better retrieval accuracy than using only 1-best transcripts for either documents, or queries, or both.
- Model-based, multiple-fault diagnosis of dynamic, continuous physical devices62
A diagnosis algorithm that can generate diagnosis candidates incrementally and detect multiple faults is presented that models continuously changing device states using a discrete set of Qsim qualitative states over time.
- A Probabilistic Forest-to-String Model for Language Generation from Typed Lambda Calculus Expressions61
This paper describes a novel probabilistic approach for generating natural language sentences from their underlying semantics in the form of typed lambda calculus, built on top of a novel reduction-based weighted synchronous context free grammar formalism.
- A machine learning approach to answering questions for reading comprehension tests61
To the best of the knowledge, this is the first work that reports that the use of a machine learning approach achieves competitive results on answering questions for reading comprehension tests.
- Abductive Plan Recognition and Diagnosis: A Comprehensive Empirical Evaluation60
Detailed empirical results on applying a general abductive system, Ac-cel, to moderately complex problems in plan recognition and diagnosis indicate that general purpose abduction is an eective and ecient mechanism for solving problems in plan recognition and diagnosis.
- System Combination for Grammatical Error Correction58
This paper proposes to combine the output from a classification- based system and an SMT-based system to improve the correction quality, achieving an F0.5 score on the test set of the CoNLL-2014 shared task.
- Semantic Role Labeling of NomBank58
This paper describes the attempt at NomBank-based automatic Semantic Role Labeling (SRL) and reports the first reported automatic NomBank SRL system, to the knowledge of this paper.
- Decomposability of translation metrics for improved evaluation and efficient algorithms57
This work proposes a very conservative modification toBleu and a cross between Bleu and word error rate that address issues while improving correlation with human judgments.
- 56
- Corpus-Based Approaches to Semantic Interpretation in Natural Language Processing55
An introduction to some of the emerging research in the application of corpusbased learning techniques to problems in semantic interpretation, namely, word-sense disambiguation and semantic parsing.
- Unsupervised Domain Adaptation of a Pretrained Cross-Lingual Language Model51
This paper proposes a novel unsupervised feature decomposition method that can automatically extract domain-specific features and domain-invariant features from the entangled pretrained cross-lingual representations, given unlabeled raw texts in the source language.
- 51
- A Beam-Search Decoder for Normalization of Social Media Text with Application to Machine Translation51
A novel beam-search decoder is proposed to effectively integrate various normalization operations and shows statistically significant improvements over two strong baselines in both normalization and translation tasks, for both Chinese and English.
- Connecting the Dots: Towards Human-Level Grammatical Error Correction50
A grammatical error correction (GEC) system primarily based on the state-of-the-art statistical machine translation (SMT) approach is built, using task-specific features and tuning, and further enhance it with the modeling power of neural network joint models.
- Do Multi-Hop Question Answering Systems Know How to Answer the Single-Hop Sub-Questions?48
A neural decomposition model is adopted to generate sub-questions for a multi-hop question, followed by extracting the corresponding sub-answers in order to shed some light on explaining the reasoning process of QA systems in answering complex questions.
- Natural language generation with tree conditional random fields48
The method is built on top of a hybrid tree representation that jointly encodes both the meaning representation as well as the natural language in a tree structure that performs better than a previous state-of-the-art natural language generation model.
- Corpus-Based Approaches to Semantic Interpretation in NLP47
An introduction to some of the emerging research in the application of corpus-based learning techniques to problems in semantic interpretation, namely, word-sense disambiguation and semantic parsing.
- 44
- 44
- Inferring cancer disease response from radiology reports using large language models with data augmentation and prompting43
Large clinical language models demonstrate potential to infer cancer disease response from radiology reports at scale by applying transformer models, a bidirectional long short-term memory model, a convolutional neural network model, and conventional machine learning methods.
- Adapting Grammatical Error Correction Based on the Native Language of Writers with Neural Network Joint Models42
This paper adapts a neural network joint model (NNJM) using L1-specific learner text and integrates it into a statistical machine translation (SMT) based GEC system and shows that adaptation achieves F 0 .
- TESLA: Translation Evaluation of Sentences with Linear-Programming-Based Analysis41
TESLA-M and TESLA, two novel automatic machine translation evaluation metrics with state-of-the-art performances are presented and it is shown that they outperform all participating systems in most tasks.
- Better Evaluation Metrics Lead to Better Machine Translation40
It is demonstrated that tuning Joshua, a hierarchical phrase-based statistical machine translation system, with the TESLA metrics results in significantly better human-judged translation quality than the BLEU-tuned baseline.
- Statistical lattice-based spoken document retrieval40
Experimental results show that the method consistently achieves better retrieval performance than using only the 1-best transcripts in statistical retrieval, outperforms a recently proposed lattice-based vector space retrieval method, and also compares favorably with a lattICE-based retrieval method based on the Okapi BM25 model.
- A Question-Focused Multi-Factor Attention Network for Question Answering39
This paper proposes a novel end-to-end question-focused multi-factor attention network for answer extraction using tensor-based transformation that achieves significant improvements over the best prior state-of-the-art results on three large-scale challenging QA datasets, namely NewsQA, TriviaQA and SearchQA.
- Joint Syntactic and Semantic Parsing of Chinese39
Evaluation on Chinese TreeBank, Chinese PropBank, and Chinese NomBank shows that the integrated parsing approach outperforms the pipeline parsing approach on n-best parse trees, a natural extension of the widely used pipeline parse approach on the top- best parse tree.
- Word sense disambiguation with semi-supervised learning39
Empirical results show that unlabeled data can bring significant improvement in WSD accuracy, and four semisupervised leaming algorithms are evaluated on 29 nouns of Senseval-2 (SE2) English lexical sample task and SE2 English all-words task.
- 38
- PEM: A Paraphrase Evaluation Metric Exploiting Parallel Texts37
PEM is the first fully automatic metric to evaluate the quality of paraphrases, and consequently, that of paraphrase generation systems, based on three criteria: adequacy, fluency, and lexical dissimilarity.
- Learning Predictive Structures for Semantic Role Labeling of NomBank36
This paper presents a novel application of Alternating Structure Optimization to the task of Semantic Role Labeling (SRL) of noun predicates in NomBank, and achieves the highest accuracy published to date on the English NomBank SRL task.
- Joint learning of preposition senses and semantic roles of prepositional phrases35
A joint probabilistic model for word sense disambiguation of prepositions and semantic role labeling of prePOSitional phrases is proposed and experiments show that jointly learning the word sense and the semantic role leads to an improvement over state-of-the-art individual classifier models on the two tasks.
- Improved Word Sense Disambiguation with Enhanced Sense Representations32
This paper focuses on enhancing the sense representations via incorporating synonyms, example phrases or sentences showing usage of word senses, and sense gloss of hypernyms, and shows that incorporating such additional information boosts the performance on WSD.
- NUS at the HOO 2012 Shared Task32
The NUS submission to the HOO 2012 shared task uses a pipeline of confidence-weighted linear classifiers to correct determiner and preposition errors and achieves the highest correction F1 score on the official test set among all 14 participating teams.
- Dynamic conditional random fields for joint sentence boundary and punctuation prediction31
This paper combines lexical, prosodic, and modified n-gram score features into the DCRF framework for a joint sentence boundary and punctuation prediction task on TDT3 English broadcast news and shows that the joint prediction method outperforms the conventional two-stage method using LCRF or maximum entropy model (MaxEnt).
- Cross-Sentence Grammatical Error Correction30
This paper employs an auxiliary encoder that encodes previous sentences and incorporate the encoding in the decoder via attention and gating mechanisms and results in statistically significant improvements in overall GEC performance over strong baselines across multiple test sets.
- 29
- Effective Attention Modeling for Neural Relation Extraction29
A novel and effective attention model which incorporates syntactic information of the sentence and a multi-factor attention mechanism is proposed which outperforms prior state-of-the-art models on the New York Times corpus.
- Word sense disambiguation using OntoNotes29
This work conducts the first large-scale WSD evaluation involving hundreds of word types and tens of thousands of sense-tagged examples, while adopting a coarse-grained sense inventory and proposes a domain adaptation technique using feature augmentation with active learning.
- 28
- A Constituent-Based Approach to Argument Labeling with Joint Inference in Discourse Parsing28
A novel constituent-based approach to argument labeling is proposed, which integrates the advantages of both linear tagging and subtree extraction and unifies intra- and intersentence cases by treating the immediately preceding sentence as a special constituent.
- Combining Coherence Models and Machine Translation Evaluation Metrics for Summarization Evaluation28
A machine translation metric is adapted to measure content coverage, an enhanced discourse coherence model is applied to evaluate summary readability, and both are combined in a trained regression model to evaluate overall responsiveness.
- 25
- Frustratingly Easy System Combination for Grammatical Error Correction24
It is demonstrated that with the right problem formulation, a simple logistic regression algorithm can be highly effective for combining GEC models and generates better corrections with higher F0.5 scores than the conventional ensemble.
- A Nil-Aware Answer Extraction Framework for Question Answering24
This paper proposes a novel nil-aware answer span extraction framework that is capable of returning Nil or a text span from the associated passage as an answer in a single step and shows that the integration of the proposed framework significantly outperforms several strong baseline systems that use pipeline or threshold-based approaches.
- Source Language Adaptation Approaches for Resource-Poor Machine Translation24
Three novel, language-independent approaches to source language adaptation for resource-poor statistical machine translation are proposed by adapting and using a large bitext for a related resource-rich language RICH and the same target language TGT.
- TESLA at WMT 2011: Translation Evaluation and Tunable Metric24
This paper describes the submission from the National University of Singapore to the WMT 2011 Shared Evaluation Task and the Tunable Metric Task, where the entry is TESLA in three different configurations: TESla-M, TESLa-F, and the new TES LA-B.
- A Unified Tagging Approach to Text Normalization23
A unified tagging approach to perform the text normalization task using Conditional Random Fields (CRF) and Experimental results on email data cleaning show that the proposed method significantly outperforms the approach of using cascaded models and that of employing independent models.
- Source Language Adaptation for Resource-Poor Machine Translation22
A novel, language-independent approach for improving machine translation from a resource-poor language to X by adapting a large bi-text for a related resource-rich language and X (the same target language).
- Reasoning Models Hallucinate More: Factuality-Aware Reinforcement Learning for Large Reasoning Models20
Factuality-aware Step-wise Policy Optimization (FSPO) is proposed, an innovative RL fine-tuning algorithm incorporating explicit factuality verification at each reasoning step that effectively reduces hallucinations while enhancing reasoning accuracy, substantially improving both reliability and performance.
- Exploiting Zero Pronouns to Improve Chinese Coreference Resolution20
A simplified semantic role labeling framework is proposed to identify clauses and to detect zero pronouns effectively, and two effective methods are employed to exploit zero pronouns for Chinese coreference resolution.
- Word sense disambiguation for all words without hard labor20
The evaluation results show that the proposed completely automatic approach to scale up word sense disambiguation to all words of English is able to achieve high accuracy, outperforming the first-sense baseline and coming close to a prior reported approach that requires manual human efforts to provide Chinese translations of English senses.
- Information Retrieval Technology20
Evaluating Scalability in Information Retrieval with Multigraded Relevance and Improving Re-ranking of Search Results using Collaborative Filtering.
- Grammatical Error Correction Using Integer Linear Programming19
This work uses integer linear programming (ILP) to model the inference process, which can easily incorporate both the power of existing error classi ers and prior knowledge on grammatical error correction.
- An efficient first-order horn-clause abduction system based on the ATMS19
This paper presents an algorithm for first-order Horn-clause abduction that uses an ATMS to avoid redundant computation and presents a heuristic version of the algorithm that uses beam search to compute a subset of the simplest explanations.
- Does BERT Know that the IS-A Relation Is Transitive?18
This investigation reveals that BERT’s predictions do not fully obey the transitivity property of the IS-A relation, and aims to quantify how much BERT agrees with the transitive property ofIS-A relations, via a minimalist probing setting.
- A Reassessment of Reference-Based Grammatical Error Correction Metrics18
This paper re-evaluate reference-based GEC metrics by measuring the system-level correlations with humans on a large dataset of human judgments of GEC outputs, and by properly conducting statistical significance tests, and finds that apart from being less interpretable and non-deterministic, GLEU also produces counter-intuitive scores in commonly occurring test examples.
- 18
- Corpus-Based Learning for Noun Phrase Coreference Resolution18
Encouraging results are obtained, indicating that on the general noun phrase coreference task, the learning approach holds promise and achieves accuracy comparable to non-learning approaches.
- Unlocking Temporal Question Answering for Large Language Models with Tailor-Made Reasoning Logic17
A novel framework that combines the extraction capability of LLMs and the logical reasoning capability of a Python solver to tackle the challenge that LLMs face when it comes to temporal reasoning is proposed.
- Domain Adaptation with Active Learning for Coreference Resolution17
Experimental results show that domain adaptation with active learning and target domain instance weighting achieves performance on MEDLINE abstracts similar to a system trained on coreference annotation of only target domain training instances, but with a greatly reduced number of target domainTraining instances that the authors need to annotate.
- One class per named entity: exploiting unlabeled text for named entity recognition17
A simple yet novel method of exploiting unlabeled text to further improve the accuracy of a high-performance state-of-the art named entity recognition (NER) system using the empirical property that many named entities occur in one name class only.
- Teaching a weaker classifier17
This paper describes how a machine-learning named entity recognizer (NER) on upper case text can be improved by using a mixed case NER and some unlabeled text, which reduces the performance gap between the mixed caseNER and the upper case N ER substantially.
- Learning to Identify Follow-Up Questions in Conversational Question Answering16
A three-way attentive pooling network is proposed that determines the suitability of a follow-up question by capturing pair-wise interactions between the associated passage, the conversation history, and a candidate follow- up question.
- Automatic Evaluation of Chinese Translation Output: Word-Level or Character-Level?16
This paper compares word-level metrics with character- level metrics on the submitted output of English-to-Chinese translation systems in the IWSLT'08 CT-EC and NIST'08 EC tasks and reveals that character-level Metrics correlate with human assessment better than word- Level metrics.
- A machine learning approach to identification and resolution of one-anaphora16
This work presents the first learning-based system for the identification and resolution of one-anaphora, and evaluated its approach on written texts drawn from the informative domains of the British National Corpus.
- A General Abductive System with Application to Plan Recognition andDiagnosis16
This thesis views explanation as abduction, where an abduction explanation is a consistent set of assumptions which, together with background knowledge, logically entails a set of observations.
- 15
- Grammatical Error Correction with Contrastive Learning in Low Error Density Domains15
A contrastive learning approach to encourage the GEC model to assign a higher probability to a correct sentence while reducing the probability of incorrect sentences that the model tends to generate, so as to improve the accuracy of the model.
- Does word sense disambiguation improve information retrieval?15
Will semantic annotation of word senses improve information retrieval?
- SemEval-2007 task 11: English lexical sample task via English-Chinese parallel text15
The process of gathering examples from the parallel corpora, the differences with similar tasks in previous SENSEVAL evaluations, and the results of participating systems are described.
- Robust Question Answering against Distribution Shifts with Test-Time Adaptation: An Empirical Study14
This work finds that TTA is comparable to RT methods, and applying TTA after RT can significantly boost the performance on COLDQA, and proposes a novel TTA method called online imitation learning (OIL).
- 13
- A Statistical Language Modeling Approach to Lattice-Based Spoken Document Retrieval13
This paper presents a novel approach to lattice-based spoken document retrieval using statistical language models: a statistical model is estimated for each document, and probabilities derived from the document models are directly used to measure relevance.
- From Moments to Milestones: Incremental Timeline Summarization Leveraging Large Language Models12
This study introduces a novel approach that leverages large language models (LLMs) for generating both event and topic timelines, and outperforms the best prior published approaches, highlighting the potential of LLMs in timeline summarization for real-world applications.
- Semantic argument classification exploiting argument interdependence12
This paper proposes the use of the neighboring semantic arguments of a predicate as additional features in determining the class of the current semantic argument, and shows significant improvement in the accuracy of semantic argument classification after exploiting argument interdependence.
- Multi-Agent Sampling: Scaling Inference Compute for Data Synthesis with Tree Search-Based Agentic Collaboration11
This work introduces Tree Search-based Orchestrated Agents~(TOA), where the workflow evolves iteratively during the sequential sampling process, and leverages Monte Carlo Tree Search (MCTS), integrating a reward model to provide real-time feedback and accelerate exploration.
- Self-Judge: Selective Instruction Following with Alignment Self-Evaluation10
Self-J, a novel self-training framework for developing judge models without needing human-annotated quality scores, is introduced, which leverages the model's inherent self-evaluation capability to extract information about response quality from labeled instruction-tuning data.
- Unsupervised Grammatical Error Correction Rivaling Supervised Methods10
Empirical results show that the GEC system outperforms previous unsupervised GEC systems, and achieves performance comparable to supervised GEC systems without ensemble, and when combined with labeled training data, achieves new state-of-the-art results on the CoNLL-2014 and NLPCC-2018 test sets.
- System Combination for Grammatical Error Correction Based on Integer Programming10
Experiments of the IP approach on combining state-of-the-art standalone GEC systems show that the combined system outperforms all standalone systems, demonstrating IP’s competitive combination capability.
- 10
- ALLECS: A Lightweight Language Error Correction System9
ALLECS provides three state-of-the-art base GEC systems using two approaches (sequence-to-sequence generation and sequence tagging), as well as two state- of- the-art GEC system combination methods using two approach (edit-based and text-based).
- A Dependency-Based Neural Reordering Model for Statistical Machine Translation9
A novel reordering approach utilizing a neural network and dependency-based embeddings to predict whether the translations of two source words linked by a dependency relation should remain in the same order or should be swapped in the translated sentence is presented.
- Character-Level Machine Translation Evaluation for Languages with Ambiguous Word Boundaries9
This work introduces the TESLA-CELAB metric (Translation Evaluation of Sentences with Linear-programming-based Analysis -- Character-level Evaluation for Languages with Ambiguous word Boundaries) for automatic machine translation evaluation, and shows empirically that TESLAB significantly outperforms character-level BLEU in the English-Chinese translation evaluation tasks.
- FocusUI: Efficient UI Grounding via Position-Preserving Visual Token Selection8
This work proposes FocusUI, an efficient UI grounding framework that selects patches most relevant to the instruction while preserving positional continuity for precise grounding, and introduces a novel PosPad strategy, which compresses each contiguous sequence of dropped visual tokens into a single special marker placed at the sequence's last index to preserve positional continuity.
- 8
- Efficient and Interpretable Grammatical Error Correction with Mixture of Experts8
A mixture-of-experts model, MoECE, for grammatical error correction that achieves the performance of T5-XL with three times fewer effective parameters and produces interpretable corrections by also identifying the error type during inference.
- Mitigating Exposure Bias in Grammatical Error Correction with Data Augmentation and Reweighting8
A novel data manipulation approach to overcome the exposure bias problem in grammatical error correction, which includes a data augmentation method during training to mimic the decoder input at inference time, and a data reweighting method to automatically balance the importance of each kind of augmented samples.
- 8
- Combining Punctuation and Disfluency Prediction: An Empirical Study8
The results show that the various methods linking the two tasks are not significantly different from one another, although they perform better than the isolated prediction method by 0.5‐1.5% in the F1 score.
- NUS at the HOO 2011 Pilot Shared Task8
This paper describes the submission of the National University of Singapore (NUS) to the Helping The authors' Own (HOO) Pilot Shared Task, which targets spelling, article, and preposition errors in a sequential processing pipeline.
- 7
- Enriching document representation via translation for improved monolingual information retrieval7
This paper proposes the use of translated words to enrich document representation, going beyond the words in the original source language to represent a document, and uses monotonic translation by removing the time-consuming reordering component.
- A 2-poisson model for probabilistic coreference of named entities for improved text retrieval7
A novel 2-Poisson model is proposed to estimate the frequency of anaphic expressions of a named entity, without explicitly resolving the anaphoric expressions, and it is shown that CEEF achieves significant and consistent improvements over state-of-the-art retrieval methods using standard term frequency estimation.
- NUS at WMT097
The system developed by the team of the National University of Singapore for English to Spanish machine translation of News Commentary text for the WMT09 Shared Translation Task is described, based on domain adaptation, combining a small in-domain News Commentary bi-text and a large out-of-domain one from the Europarl corpus.
- 7
- Factorized Learning for Temporally Grounded Video-Language Models6
This work proposes D^{2}$ VLM, a framework that decouples the learning of these two tasks while also emphasizing their inherent dependency, and introduces evidence tokens for evidence grounding, which emphasize event-level visual semantic capture beyond the focus on timestamp representation in existing works.
- Maximum Metric Score Training for Coreference Resolution6
A novel approach comprising the use of instance weighting and beam search to maximize the evaluation metric score on the training corpus during training achieves significant improvement over the state-of-the-art.
- A Semi-supervised Learning Approach with Two Teachers to Improve Breakdown Identification in Dialogues5
A novel semi-supervised teacher-student learning framework to tackle identifying breakdowns in ongoing dialogues and leverages unlabeled data to improve classification in student training where it can achieve improvements over single-teacher performance.
- Learning from the Experience of Doctors: Automated Diagnosis of Appendicitis Based on Clinical Notes5
Visualization shows that the proposed novel neural network approach that learns to diagnose acute appendicitis based on doctors’ free-text ED notes without any feature engineering is able to learn important features, signs, and symptoms of patients from unstructured free- text ED notes, which will help doctors to make better diagnosis.
- To Swap or Not to Swap? Exploiting Dependency Word Pairs for Reordering in Statistical Machine Translation5
A novel reordering approach utilizing sparse features based on dependency word pairs that captures whether two words, which are related by a dependency link in the source sentence dependency parse tree, follow the same order or are swapped in the translation output.
- A Beam-Search Decoder for Disfluency Detection5
This paper proposes node-weighted max-margin Markov networks (M3N) to boost the performance on words belonging to specific part-of-speech (POS) classes and shows the importance of measuring the quality of cleaned-up sentences and performing multiple passes of disfluency detection.
- MaxSim: performance and effects of translation fluency5
The proposed automatic machine translation evaluation metric MaxSim calculates a similarity score between a pair of English system-reference sentences by comparing information items such as n-grams across the sentence pair and computes similarity scores between items.
- DSO at TREC-8: A Hybrid Algorithm for the Routing Task.5
A new hybrid algorithm is described that is able to give good performance on TREC-8 test data and achieved a slight improvement in average uninterpolated precision by using Dynamic Feedback Optimization as another weight tuning algorithm.
- On the Robustness of Question Rewriting Systems to Questions of Varying Hardness4
To enhance the robustness of QR systems to questions of varying hardness, a novel learning framework for QR is proposed that first trains a QR model independently on each subset of questions of a certain level of hardness, then combines these QR models as one joint model for inference.
- 4
- Diversity-Driven Combination for Grammatical Error Correction4
Diversity-Driven Combination for GEC is presented, a system combination strategy that encourages diversity among component systems that achieves significant performance gain with a small number of training examples and outperforms the component systems by a large margin.
- Upping the Ante: Towards a Better Benchmark for Chinese-to-English Machine Translation4
This paper proposes a benchmark in evaluation setup for Chinese-to-English machine translation, such that the effectiveness of a new proposed MT approach can be directly compared to previous approaches and builds a highly competitive state-of-the-art MT system trained on a large-scale training set.
- The NUS Statistical Machine Translation System for IWSLT 20094
The system developed by the team of the National University of Singapore for the Chinese-English BTEC task of the IWSLT 2009 evaluation campaign adopted a state-of-the-art phrase-based statistical machine translation approach and focused on experiments with different Chinese word segmentation standards.
- 4
- OpenSeal: Good, Fast, and Cheap Construction of an Open-Source Southeast Asian LLM via Parallel Data3
The findings show that using only parallel data is the most effective way to extend an LLM to new languages, and OpenSeal is built, the first truly open Southeast Asian LLM that rivals the performance of existing models of similar size.
- A Co-Attentive Cross-Lingual Neural Model for Dialogue Breakdown Detection3
A novel dialogue breakdown detection model that jointly incorporates a pretrained cross-lingual language model and a co-attention network is proposed that outperforms all previous approaches on all evaluation metrics in both the Japanese and English tracks in Dialogue Breakdown Detection Challenge 4.
- Automated Anonymization as Spelling Variant Detection3
The task of anonymizing clinical texts written in sentence fragments and which frequently contain symbols, abbreviations, and misspelled words is tackled, exploiting patients’ personal information in the structured fields to detect their spelling variants in clinical texts.
- de of Coherence lanation3
Some problems encountered using abduction to understand text are described, and some solutions to overcome these problems are presented, around the use of a different criterion, called explanatory coherence, as the primary measure to evaluate the quality of an explanation.
- 2
- 2
- Just What You Desire: Constrained Timeline Summarization with Self-Reflection for Enhanced Relevance2
- 2
- 2
- 2
- 2
- The SIGIR 2008 workshop program2
The SIGIR workshops program provides an informal but structured setting for extended dialog and brainstorming that supports cross fertilization between existing communities and the development of new ones, often bringing together practitioners from industry and academia.
- Refining the Wrapper Approach - Smoothed Error Estimates for Feature Selection2
It is shown empirically that smoothing the error estimate gives improved performance in feature selection and proposes using the jackknife to reduce the bias inherent in Bayesian estimators.
- Equity with Efficiency: An Empirical Study of Tokenizers for Multilingual Large Language Models1
The first systematic comparison of equitable tokenizers on a unified benchmark spanning 11 Southeast Asian languages is presented, demonstrating that cross-lingual fairness and tokenization efficiency are not fundamentally at odds, and offer practical guidance for designing equitable multilingual models.
- Game of Thought: Robust Information Seeking with Large Language Models Using Game Theory1
Game of Thought (GoT), a framework that applies game-theoretic techniques to approximate a Nash equilibrium strategy for the restricted variant of the game, is proposed and empirical results demonstrate that this approach consistently improves worst-case performance.
- 1
- WAMP: Writing, Annotation, and Marking Platform1
This paper proposes a web-based annotation tool – WAMP – that tackles the issue of generating annotated corpora by allowing annotators to annotate essays with ease and export the resulting annotated essays for use in GEC research.
- My Tenure as the Editor-in-Chief of Computational Linguistics1
This editorial will describe the changes that the editor-in-chief of Computational Linguistics introduced at the journal, and highlight the achievements and challenges of the journal.
- 1
- –
- –
- –
- Communicating Chess Strategies in Natural Language–
This work designs a pipeline for verbalizing strategies and an evaluation framework for objective evaluation of generated strategy descriptions, and shows that natural language is a promising and interpretable medium for communicating strategic information to both human and LLM players.
- –
- –
- –
- –
- A Hierarchical Entity Graph Convolutional Network for Relation Extraction across Documents–
This work creates a dataset for two-hop relation extraction, where each chain contains exactly two documents, and proposes a hierarchical entity graph convolutional network (HEGCN) model for this task that improves performance by 1.1% F1 score on this dataset.
- The State of the Journal–
TALIP has diligently stuck to its target of reviewing and publishing only articles related to Asian language information processing, with an acceptance rate of 15% in 2009, which is one of the most competitive acceptance rates among all ACM journals and transactions.
- –
- Information Retrieval Technology: Third Asia Information Retrieval Symposium, AIRS 2006, Singapore, October 16-18, 2006, Proceedings–
Evaluating Scalability in Information Retrieval with Multigraded Relevance and Improving Re-ranking of Search Results using Collaborative Filtering.
- –
- –
- Instructions for ACL-2005 Proceedings–
This document contains the instructions for preparing a camera-ready manuscript for the proceedings of ACL-2005, and is therefore an example of what your manuscript should look like.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-11. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.