_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d555420 | This paper focuses on two aspects of Machine Translation: parallel corpora and translation model. First, we present a method to automatically build parallel corpora from subtitle files. We use subtitle files gathered from the Internet. This leads to useful data for Subtitling Machine Translation. Our method is based on Dynamic Time Warping. We evaluated this alignment method by comparing it with a sample aligned by hand and we obtained a precision of alignment equal to 0.92. Second, we use the notion of inter-lingual triggers in order to build from the subtitle parallel corpora multilingual dictionaries and translation tables for machine translation. Inter-lingual triggers allow to detect couple of source and target words from parallel corpora. The Mutual Information measure used to determine inter-lingual triggers allows to hypothesize that a word in the source language is a translation of another word in the target language. We evaluate the obtained dictionary by comparing it to two existing dictionaries. Then, we integrated the obtained translation tables into an entire translation decoding process supplied by Pharaoh(Koehn, 2004). We compared the translation performance using our translation tables with the performance obtained by the Giza++ tool(Al-Onaizan et al., 1999). The results showed that the system tuned for our tables improves the Bleu (Papineni et al., 2001) value by 2.2% compared to the ones obtained by Giza++. | Building a bilingual dictionary from movie subtitles based on inter-lingual triggers |
d5604513 | Distributed word representations are very useful for capturing semantic information and have been successfully applied in a variety of NLP tasks, especially on English. In this work, we innovatively develop two component-enhanced Chinese character embedding models and their bigram extensions. Distinguished from English word embeddings, our models explore the compositions of Chinese characters, which often serve as semantic indictors inherently. The evaluations on both word similarity and text classification demonstrate the effectiveness of our models. | Component-Enhanced Chinese Character Embeddings |
d252780283 | In this paper, we focus on video-to-text summarization and investigate how to best utilize multimodal information for summarizing long inputs (e.g., an hour-long TV show) into long outputs (e.g., a multi-sentence summary). We extend SummScreen (Chen et al., 2022), a dialogue summarization dataset consisting of transcripts of TV episodes with reference summaries, and create a multimodal variant by collecting corresponding full-length videos. We incorporate multimodal information into a pretrained textual summarizer efficiently using adapter modules augmented with a hierarchical structure while tuning only 3.8% of model parameters. Our experiments demonstrate that multimodal adapters outperform more memoryheavy and fully fine-tuned textual summarization methods. | Hierarchical3D Adapters for Long Video-to-text Summarization |
d407178 | 1In this paper we present recent advances in acoustic and language modeling that improve recognition performance when children read out loud within digital books. First we extend previous work by incorporating crossutterance word history information and dynamic n-gram language modeling. By additionally incorporating Vocal Tract Length Normalization (VTLN), Speaker-Adaptive Training (SAT) and iterative unsupervised structural maximum a posteriori linear regression (SMAPLR) adaptation we demonstrate a 54% reduction in word error rate. Next, we show how data from children's read-aloud sessions can be utilized to improve accuracy in a spontaneous story summarization task. An error reduction of 15% over previous published results is shown. Finally we describe a novel real-time implementation of our research system that incorporates time-adaptive acoustic and language modeling. | Advances in Children's Speech Recognition within an Interactive Literacy Tutor |
d6431039 | In this paper, we attempt to explain the emergence of the linguistic diversity that exists across the consonant inventories of some of the major language families of the world through a complex network based growth model. There is only a single parameter for this model that is meant to introduce a small amount of randomness in the otherwise preferential attachment based growth process. The experiments with this model parameter indicates that the choice of consonants among the languages within a family are far more preferential than it is across the families. Furthermore, our observations indicate that this parameter might bear a correlation with the period of existence of the language families under investigation. These findings lead us to argue that preferential attachement seems to be an appropriate high level abstraction for language acquisition and change. | Language Diversity across the Consonant Inventories: A Study in the Framework of Complex Networks |
d1470 | We address the problem of structural disambiguation in syntactic parsing.In psycholinguistics, a number of principles of disambiguation have been proposed, notably the Lexical Preference Rule (LPR), the Right Association Principle (RAP), and the Attach Low and Parallel Principle (ALPP).We argue that in order to improve disambiguation results it is necessary to implement these principles on the basis of a probabilistic methodology.We define a 'three-word probability' for implementing LPR, and a 'length probability' for implementing RAP and ALPP.Furthermore, we adopt the 'back-off' method to combine these two types of probabilities.Our experimental results indicate our method to be effective, attaining an accuracy of 89.2%.1 A representation of a probability distribution is called a 'probability model,' or simplely a 'model.' | A Probabilistic Disambiguation Method Based on Psycholinguistic Principles |
d279533 | We describe tree edit models for representing sequences of tree transformations involving complex reordering phenomena and demonstrate that they offer a simple, intuitive, and effective method for modeling pairs of semantically related sentences. To efficiently extract sequences of edits, we employ a tree kernel as a heuristic in a greedy search routine. We describe a logistic regression model that uses 33 syntactic features of edit sequences to classify the sentence pairs. The approach leads to competitive performance in recognizing textual entailment, paraphrase identification, and answer selection for question answering. | Tree Edit Models for Recognizing Textual Entailments, Paraphrases, and Answers to Questions |
d8140780 | Interest in neural machine translation has grown rapidly as its effectiveness has been demonstrated across language and data scenarios.New research regularly introduces architectural and algorithmic improvements that lead to significant gains over "vanilla" NMT implementations. However, these new techniques are rarely evaluated in the context of previously published techniques, specifically those that are widely used in state-of-theart production and shared-task systems. As a result, it is often difficult to determine whether improvements from research will carry over to systems deployed for real-world use. In this work, we recommend three specific methods that are relatively easy to implement and result in much stronger experimental systems. Beyond reporting significantly higher BLEU scores, we conduct an in-depth analysis of where improvements originate and what inherent weaknesses of basic NMT models are being addressed. We then compare the relative gains afforded by several other techniques proposed in the literature when starting with vanilla systems versus our stronger baselines, showing that experimental conclusions may change depending on the baseline chosen. This indicates that choosing a strong baseline is crucial for reporting reliable experimental results. | Stronger Baselines for Trustable Results in Neural Machine Translation |
d171244 | Argument labeling of explicit discourse relations is a challenging task. The state of the art systems achieve slightly above 55% F-measure but require hand-crafted features. In this paper, we propose a Long Short Term Memory (LSTM) based model for argument labeling. We experimented with multiple configurations of our model. Using the PDTB dataset, our best model achieved an F1 measure of 23.05% without any feature engineering. This is significantly higher than the 20.52% achieved by the state of the art RNN approach, but significantly lower than the feature based state of the art systems. On the other hand, because our approach learns only from the raw dataset, it is more widely applicable to multiple textual genres and languages. | Argument Labeling of Explicit Discourse Relations using LSTM Neural Networks |
d471963 | I)ATI{ is a declarative re.presentation language ti)r lex-. ical iifformation and as such, fit prin(:iple, neul;ral with resl)(;ct; 1;o i)arl;icul&r l)rocessing st,rat,egies. Previous DATR (:l)mt)iler/inl;erI)ret(!r sy,qt(!ms support only one al:l:e.4s ,%rat,egy ~hnt, closely resembles the set, of inti~r-. | Reverse Queries in DATR* |
d53244635 | This paper presents SimpleNLG-NL, an adaptation of the SimpleNLG surface realisation engine for the Dutch language. It describes a novel method for determining and testing the grammatical constructions to be implemented, using target sentences sampled from a treebank. | Going Dutch: Creating SimpleNLG-NL |
d219302330 | This book gives a short but comprehensive overview of the field of spoken dialogue systems, outlining the issues involved in building and evaluating this type of system and making liberal use of techniques and examples from a wide range of implemented systems. It provides an excellent review of the research, with particularly relevant discussions of error handling and system evaluation, and is suitable both as an introduction to this research area and as a survey of current state-of-the-art techniques.The book is structured into seven chapters. Chapter 1 provides an introduction to the research area and briefly introduces the topics covered in the book. The end of the chapter consists of a list of links to tools and components that can be used for dialogue system development, which-although currently useful-seems likely to go out of date quickly.Chapter 2 addresses the task of dialogue management, beginning by describing simple graph-and frame-based methods for dialogue control, and continuing with a discussion of VoiceXML. The chapter ends with an extended discussion of recent work in statistical approaches to dialogue control and modeling. It is unfortunate that the discussion of other methods such as the information state approach and plan-based models is postponed to Chapter 4, but otherwise this chapter provides a good summary of both classic and recent approaches.Chapter 3 discusses error handling, which is divided into three processes: error detection, error prediction (i.e., the online prediction of errors based on dialogue features), and error recovery. After surveying a range of previous approaches to these three subtasks, the authors go on to discuss several more recent, data-driven approaches. Error handling is both a vital component of any spoken dialogue system designed for realworld use and an active area of current research, so this compact summary of techniques and issues is welcome.Chapter 4 contains case studies illustrating a range of dialogue control strategies and models. It begins with a description of the information state approach, and then moves on to discuss plan-based approaches as exemplified in the TRAINS and TRIPS projects. This is followed by a discussion of two systems that employ software agents for dialogue management: the Queen's Communicator and the AthosMail system. Finally, two systems which make use of statistical models are presented: the Microsoft Bayesian receptionist, which models conversation as decision making under uncertainty, and the DIHANA system, which employs corpus-based dialogue management. The case studies in this chapter provide detailed examples of a range of techniques, along with Computational Linguistics Volume 36, Number 4 | Book Review Spoken Dialogue Systems |
d52185631 | Question Generation is the task of automatically creating questions from textual input. In this work we present a new Attentional Encoder-Decoder Recurrent Neural Network model for automatic question generation. Our model incorporates linguistic features and an additional sentence embedding to capture meaning at both sentence and word levels. The linguistic features are designed to capture information related to named entity recognition, word case, and entity coreference resolution. In addition our model uses a copying mechanism and a special answer signal that enables generation of numerous diverse questions on a given sentence. Our model achieves state of the art results of 19.98 Bleu 4 on a benchmark Question Generation dataset, outperforming all previously published results by a significant margin. A human evaluation also shows that the added features improve the quality of the generated questions. | Neural Generation of Diverse Questions using Answer Focus, Contextual and Linguistic Features |
d5270848 | It is today acknowledged that neural network language models outperform backoff language models in applications like speech recognition or statistical machine translation. However, training these models on large amounts of data can take several days. We present efficient techniques to adapt a neural network language model to new data. Instead of training a completely new model or rely on mixture approaches, we propose two new methods: continued training on resampled data or insertion of adaptation layers. We present experimental results in an CAT environment where the post-edits of professional translators are used to improve an SMT system. Both methods are very fast and achieve significant improvements without over-fitting the small adaptation data. | INCREMENTAL ADAPTATION STRATEGIES FOR NEURAL NETWORK LANGUAGE MODELS |
d11336213 | Neural machine translation is a relatively new approach to statistical machine translation based purely on neural networks. The neural machine translation models often consist of an encoder and a decoder. The encoder extracts a fixed-length representation from a variable-length input sentence, and the decoder generates a correct translation from this representation. In this paper, we focus on analyzing the properties of the neural machine translation using two models; RNN Encoder-Decoder and a newly proposed gated recursive convolutional neural network. We show that the neural machine translation performs relatively well on short sentences without unknown words, but its performance degrades rapidly as the length of the sentence and the number of unknown words increase. Furthermore, we find that the proposed gated recursive convolutional network learns a grammatical structure of a sentence automatically. | On the Properties of Neural Machine Translation: Encoder-Decoder Approaches |
d10109787 | We present a novel learning method for word embeddings designed for relation classification. Our word embeddings are trained by predicting words between noun pairs using lexical relation-specific features on a large unlabeled corpus. This allows us to explicitly incorporate relationspecific information into the word embeddings. The learned word embeddings are then used to construct feature vectors for a relation classification model. On a well-established semantic relation classification task, our method significantly outperforms a baseline based on a previously introduced word embedding method, and compares favorably to previous state-ofthe-art models without syntactic information or manually constructed external resources. Furthermore, when incorporating external resources, our method outperforms the previous state of the art. | Task-Oriented Learning of Word Embeddings for Semantic Relation Classification |
d3576631 | Most work in relation extraction forms a prediction by looking at a short span of text within a single sentence containing a single entity pair mention. This approach often does not consider interactions across mentions, requires redundant computation for each mention pair, and ignores relationships expressed across sentence boundaries. These problems are exacerbated by the document-(rather than sentence-) level annotation common in biological text. In response, we propose a model which simultaneously predicts relationships between all mention pairs in a document. We form pairwise predictions over entire paper abstracts using an efficient self-attention encoder. Allpairs mention scores allow us to perform multi-instance learning by aggregating over mentions to form entity pair representations. We further adapt to settings without mention-level annotation by jointly training to predict named entities and adding a corpus of weakly labeled data. In experiments on two Biocreative benchmark datasets, we achieve state of the art performance on the Biocreative V Chemical Disease Relation dataset for models without external KB resources. We also introduce a new dataset an order of magnitude larger than existing human-annotated biological information extraction datasets and more accurate than distantly supervised alternatives. | Simultaneously Self-Attending to All Mentions for Full-Abstract Biological Relation Extraction |
d202785231 | Metaphors allow us to convey emotion by connecting physical experiences and abstract concepts. The results of previous research in linguistics and psychology suggest that metaphorical phrases tend to be more emotionally evocative than their literal counterparts. In this paper, we investigate the relationship between metaphor and emotion within a computational framework, by proposing the first joint model of these phenomena. We experiment with several multitask learning architectures for this purpose, involving both hard and soft parameter sharing. Our results demonstrate that metaphor identification and emotion prediction mutually benefit from joint learning and our models advance the state of the art in both of these tasks. | Modelling the interplay of metaphor and emotion through multitask learning |
d388 | Sentiment analysis seeks to identify the viewpoint(s) underlying a text span; an example application is classifying a movie review as "thumbs up" or "thumbs down". To determine this sentiment polarity, we propose a novel machine-learning method that applies text-categorization techniques to just the subjective portions of the document. Extracting these portions can be implemented using efficient techniques for finding minimum cuts in graphs; this greatly facilitates incorporation of cross-sentence contextual constraints. | A Sentimental Education: Sentiment Analysis Using Subjectivity Summarization Based on Minimum Cuts |
d8446378 | Understanding the semantic relationships between terms is a fundamental task in natural language processing applications. While structured resources that can express those relationships in a formal way, such as ontologies, are still scarce, a large number of linguistic resources gathering dictionary definitions is becoming available, but understanding the semantic structure of natural language definitions is fundamental to make them useful in semantic interpretation tasks. Based on an analysis of a subset of WordNet's glosses, we propose a set of semantic roles that compose the semantic structure of a dictionary definition, and show how they are related to the definition's syntactic configuration, identifying patterns that can be used in the development of information extraction frameworks and semantic models.This work is licensed under a Creative Commons Attribution 4.0 International Licence.Licence details: | Categorization of Semantic Roles for Dictionary Definitions |
d21729712 | The iLCM project pursues the development of an integrated research environment for the analysis of structured and unstructured data in a "Software as a Service" architecture (SaaS). The research environment addresses requirements for the quantitative evaluation of large amounts of qualitative data with text mining methods as well as requirements for the reproducibility of data-driven research designs in the social sciences. For this, the iLCM research environment comprises two central components. First, the Leipzig Corpus Miner (LCM), a decentralized SaaS application for the analysis of large amounts of news texts developed in a previous Digital Humanities project. Second, the text mining tools implemented in the LCM are extended by an "Open Research Computing" (ORC) environment for executable script documents, so-called "notebooks". This novel integration allows to combine generic, high-performance methods to process large amounts of unstructured text data and with individual program scripts to address specific research requirements in computational social science and digital humanities. | iLCM -A Virtual Research Infrastructure for Large-Scale Qualitative Data |
d263313957 | Machine involvement has the potential to speed up language documentation. We assess this potential with timed annotation experiments that consider annotator expertise, example selection methods, and suggestions from a machine classifier. We find that better example selection and label suggestions improve efficiency, but effectiveness depends strongly on annotator expertise. Our expert performed best with uncertainty selection, but gained little from suggestions. Our non-expert performed best with random selection and suggestions. The results underscore the importance both of measuring annotation cost reductions with respect to time and of the need for cost-sensitive learning methods that adapt to annotators. | How well does active learning actually work? Time-based evaluation of cost-reduction strategies for language documentation |
d18694313 | 4. | AN AUTOMATIC SPEECH RECOGNITION SYSTEM FOR T! IE ITALIAN LANGUAGE |
d8700252 | This article is concerned with determining the constraints on the selection of appropriate intonation in speech generation in humanmachine information seeking dialogues. The two pillars of our system--a state-of-theart computational dialogue model and a state-of-the-art NL generator--are presented. Based on this, we determine the kinds of linguistic and pragmatic knowledge needed to sufficiently constrain choice in the intonational resources of the system. We take into consideration factors such as dialogue history, speaker's attitudes, heater's expectations, and semantic speech functions. | Matchmaking: dialogue modelling and speech generation meet* |
d252918278 | To facilitate conversational question answering (CQA) over hybrid contexts in finance, we present a new dataset, named PACIFIC. Compared with existing CQA datasets, PACIFIC exhibits three key features: (i) proactivity, (ii) numerical reasoning, and (iii) hybrid context of tables and text. A new task is defined accordingly to study Proactive Conversational Question Answering (PCQA), which combines clarification question generation and CQA. In addition, we propose a novel method, namely UniPCQA, to adapt a hybrid format of input and output content in PCQA into the Seq2Seq problem, including the reformulation of the numerical reasoning process as code generation. UniPCQA performs multi-task learning over all sub-tasks in PCQA and incorporates a simple ensemble strategy to alleviate the error propagation issue in the multi-task learning by cross-validating top-k sampled Seq2Seq outputs. We benchmark the PACIFIC dataset with extensive baselines and provide comprehensive evaluations on each sub-task of PCQA. | PACIFIC: Towards Proactive Conversational Question Answering over Tabular and Textual Data in Finance * |
d52185392 | Capturing the semantic relations of words in a vector space contributes to many natural language processing tasks. One promising approach exploits lexico-syntactic patterns as features of word pairs. In this paper, we propose a novel model of this pattern-based approach, neural latent relational analysis (NLRA). NLRA can generalize co-occurrences of word pairs and lexicosyntactic patterns, and obtain embeddings of the word pairs that do not co-occur. This overcomes the critical data sparseness problem encountered in previous pattern-based models. Our experimental results on measuring relational similarity demonstrate that NLRA outperforms the previous pattern-based models. In addition, when combined with a vector offset model, NLRA achieves a performance comparable to that of the state-of-theart model that exploits additional semantic relational data. | Neural Latent Relational Analysis to Capture Lexical Semantic Relations in a Vector Space |
d226278291 | With thousands of academic articles shared on a daily basis, it has become increasingly difficult to keep up with the latest scientific findings. To overcome this problem, we introduce a new task of disentangled paper summarization, which seeks to generate separate summaries for the paper contributions and the context of the work, making it easier to identify the key findings shared in articles. For this purpose, we extend the S2ORC corpus of academic articles, which spans a diverse set of domains ranging from economics to psychology, by adding disentangled "contribution" and "context" reference labels. Together with the dataset, we introduce and analyze three baseline approaches: 1) a unified model controlled by input code prefixes, 2) a model with separate generation heads specialized in generating the disentangled outputs, and 3) a training strategy that guides the model using additional supervision coming from inbound and outbound citations. We also propose a comprehensive automatic evaluation protocol which reports the relevance, novelty, and disentanglement of generated outputs. Through a human study involving expert annotators, we show that in 79%, of cases our new task is considered more helpful than traditional scientific paper summarization. | What's New? Summarizing Contributions in Scientific Literature |
d5836739 | Sequence labeling for extraction of medical events and their attributes from unstructured text in Electronic Health Record (EHR) notes is a key step towards semantic understanding of EHRs. It has important applications in health informatics including pharmacovigilance and drug surveillance. The state of the art supervised machine learning models in this domain are based on Conditional Random Fields (CRFs) with features calculated from fixed context windows. In this application, we explored recurrent neural network frameworks and show that they significantly outperformed the CRF models. | Bidirectional RNN for Medical Event Detection in Electronic Health Records |
d258378205 | In multi-turn dialog understanding, semantic frames are constructed by detecting intents and slots within each user utterance. However, recent works lack the capability of modeling multi-turn dynamics within a dialog in natural language understanding (NLU), instead leaving them for updating dialog states only. Moreover, humans usually associate relevant background knowledge with the current dialog contexts to better illustrate slot semantics revealed from word connotations, where previous works have explored such possibility mostly in knowledge-grounded response generation. In this paper, we propose to amend the research gap by equipping a BERT-based NLU framework with knowledge and context awareness. We first encode dialog contexts with a unidirectional context-aware transformer encoder and select relevant inter-word knowledge with the current word and previous history based on a knowledge attention mechanism. Experimental results in two complicated multi-turn dialog datasets have demonstrated significant improvements of our proposed framework. Attention visualization also demonstrates how our modules leverage knowledge across the utterance. | Infusing Context and Knowledge Awareness in Multi-turn Dialog Understanding |
d258378276 | The news subheading summarizes an article's contents in several sentences to support the headline limited to solely conveying the main contents. So, it is necessary to generate compelling news subheadings in consideration of the structural characteristics of the news. In this paper, we propose a subheading generation model using topical headline information. We introduce a discriminative learning method that utilizes the prediction result of masked headline tokens. Experiments show that the proposed model is effective and outperforms the comparative models on three news datasets written in two languages. We also show that our model performs robustly on a small dataset and various masking ratios. Qualitative analysis and human evaluations also show that the overall quality of generated subheadings improved over the comparative models. | Headline Token-based Discriminative Learning for Subheading Generation in News Article |
d248986557 | In the presented study, we discover that the socalled "transition freedom" metric appears superior for unsupervised tokenization purposes in comparison to statistical metrics such as mutual information and conditional probability, providing F-measure scores in range from 0.71 to 1.0 across explored multilingual corpora. We find that different languages require different offshoots of that metric (such as derivative, variance, and "peak values") for successful tokenization. Larger training corpora do not necessarily result in better tokenization quality, while compressing the models by eliminating statistically weak evidence tends to improve performance. The proposed unsupervised tokenization technique provides quality better than or comparable to lexicon-based ones, depending on the language. | Unsupervised Tokenization Learning |
d1862889 | Named entity recognition, and other information extraction tasks, frequently use linguistic features such as part of speech tags or chunkings. For languages where word boundaries are not readily identified in text, word segmentation is a key first step to generating features for an NER system. While using word boundary tags as features are helpful, the signals that aid in identifying these boundaries may provide richer information for an NER system. New state-of-the-art word segmentation systems use neural models to learn representations for predicting word boundaries. We show that these same representations, jointly trained with an NER system, yield significant improvements in NER for Chinese social media. In our experiments, jointly training NER and word segmentation with an LSTM-CRF model yields nearly 5% absolute improvement over previously published results. | Improving Named Entity Recognition for Chinese Social Media with Word Segmentation Representation Learning |
d1325523 | This paper presents the systems developed by LIUM and CVC for the WMT16 Multimodal Machine Translation challenge. We explored various comparative methods, namely phrase-based systems and attentional recurrent neural networks models trained using monomodal or multimodal data. We also performed a human evaluation in order to estimate the usefulness of multimodal data for human machine translation and image description generation. Our systems obtained the best results for both tasks according to the automatic evaluation metrics BLEU and ME-TEOR. | Does Multimodality Help Human and Machine for Translation and Image Captioning? |
d12998021 | In this paper, we present a novel unsupervised algorithm for word sense disambiguation (WSD) at the document level. Our algorithm is inspired by a widely-used approach in the field of genetics for whole genome sequencing, known as the Shotgun sequencing technique. The proposed WSD algorithm is based on three main steps. First, a brute-force WSD algorithm is applied to short context windows (up to 10 words) selected from the document in order to generate a short list of likely sense configurations for each window. In the second step, these local sense configurations are assembled into longer composite configurations based on suffix and prefix matching. The resulted configurations are ranked by their length, and the sense of each word is chosen based on a voting scheme that considers only the top k configurations in which the word appears. We compare our algorithm with other state-of-the-art unsupervised WSD algorithms and demonstrate better performance, sometimes by a very large margin. We also show that our algorithm can yield better performance than the Most Common Sense (MCS) baseline on one data set. Moreover, our algorithm has a very small number of parameters, is robust to parameter tuning, and, unlike other bioinspired methods, it gives a deterministic solution (it does not involve random choices). | ShotgunWSD: An unsupervised algorithm for global word sense disambiguation inspired by DNA sequencing |
d5649853 | Verbs play a critical role in the meaning of sentences, but these ubiquitous words have received little attention in recent distributional semantics research. We introduce SimVerb-3500, an evaluation resource that provides human ratings for the similarity of 3,500 verb pairs. SimVerb-3500 covers all normed verb types from the USF free-association database, providing at least three examples for every Verb-Net class. This broad coverage facilitates detailed analyses of how syntactic and semantic phenomena together influence human understanding of verb meaning. Further, with significantly larger development and test sets than existing benchmarks, SimVerb-3500 enables more robust evaluation of representation learning architectures and promotes the development of methods tailored to verbs. We hope that SimVerb-3500 will enable a richer understanding of the diversity and complexity of verb semantics and guide the development of systems that can effectively represent and interpret this meaning. | SimVerb-3500: A Large-Scale Evaluation Set of Verb Similarity |
d1984129 | In this paper I present ongoing work on the data-oriented parsing (DOP) model. In previous work, DOP was tested on a cleaned-up set of analyzed part-of-speech strings from the Penn Treebank, achieving excellent test results. This left, however, two important questions unanswered: (1) how does DOP perform if tested on unedited data, and (2) how can DOP be used for parsing word strings that contain unknown words? This paper addresses these questions. We show that parse results on unedited data are worse than on cleaned-up data, although still very competitive if compared to other models. As to the parsing of word strings, we show that the hardness of the problem does not so much depend on unknown words, but on previously unseen lexical categories of known words. We give a novel method for parsing these words by estimating the probabilities of unknown subtrees. The method is of general interest since it shows that good performance can be obtained without the use of a part-ofspeech tagger. To the best of our knowledge, our method outperforms other statistical parsers tested on Penn Treebank word strings. | Two Questions about Data-Oriented Parsing* |
d18449288 | While generative models such as Latent Dirichlet Allocation (LDA) have proven fruitful in topic modeling, they often require detailed assumptions and careful specification of hyperparameters. Such model complexity issues only compound when trying to generalize generative models to incorporate human input. We introduce Correlation Explanation (CorEx), an alternative approach to topic modeling that does not assume an underlying generative model, and instead learns maximally informative topics through an informationtheoretic framework. This framework naturally generalizes to hierarchical and semisupervised extensions with no additional modeling assumptions. In particular, word-level domain knowledge can be flexibly incorporated within CorEx through anchor words, allowing topic separability and representation to be promoted with minimal human intervention. Across a variety of datasets, metrics, and experiments, we demonstrate that CorEx produces topics that are comparable in quality to those produced by unsupervised and semisupervised variants of LDA.529 | Anchored Correlation Explanation: Topic Modeling with Minimal Domain Knowledge |
d6293901 | A LEXICAL DATABASE TOOL FOR QUANTITATIVE PHONOLOGICAL RESEARCH | |
d6078795 | We introduce Discriminative BLEU (∆BLEU), a novel metric for intrinsic evaluation of generated text in tasks that admit a diverse range of possible outputs. Reference strings are scored for quality by human raters on a scale of [−1, +1] to weight multi-reference BLEU. In tasks involving generation of conversational responses, ∆BLEU correlates reasonably with human judgments and outperforms sentence-level and IBM BLEU in terms of both Spearman's ρ and Kendall's τ . | ∆BLEU: A Discriminative Metric for Generation Tasks with Intrinsically Diverse Targets |
d245838242 | Medieval French is known to be relatively hard to parse, with several possible sources of confusion for automatic parsers, among which its flexible word order and its graphical and syntactic variation, both synchronically and diachronically. In this work, we study in particular the influence of word order, by comparing the performances of two state-of-the-art syntactic parsers trained and evaluated on two treebanks: the Syntactic Reference Corpus of Medieval French (SRCMF), a treebank of Old French (9th-13th century) and the Google Stanford Dependency treebank of contemporary French.This work is licensed under a Creative Commons Attribution 4.0 International Licence. Licence details: http:// creativecommons.org/licenses/by/4.0/. | Is Old French tougher to parse? |
d258486909 | We present a cross-lingual study of homotransphobia on Twitter, examining the prevalence and forms of homotransphobic content in tweets related to LGBT issues in seven languages. Our findings reveal that homotransphobia is a global problem that takes on distinct cultural expressions, influenced by factors such as misinformation, cultural prejudices, and religious beliefs. To aid the detection of hate speech, we also devise a taxonomy that classifies public discourse around LGBT issues. By contributing to the growing body of research on online hate speech, our study provides valuable insights for creating effective strategies to combat homotransphobia on social media.Warning: this paper contains examples of offensive language. 1 | A Cross-Lingual Study of Homotransphobia on Twitter |
d258486957 | This paper presents a general-purpose NLP pipeline for Ancient or early forms of Greek (Classical, Koine, and Medieval) that achieves a slight state-of-art improvement by training on several Universal Dependencies treebanks jointly. We measure the performance of the model against other comparable tools. We show that the selected Greek language models tend not to generalize well to out-of-training set samples. More work is necessary to ensure interoperability between the existing datasets. We identify the main issues and list suggestions for improvements. | OdyCy -A general-purpose NLP pipeline for Ancient Greek |
d9094128 | Although there have been many research projects to extract protein pathways, most such information still exists only in the scientific literature, usually written in natural languages and defying data mining efforts. We present a novel and robust approach for extracting protein-protein interactions from the literature. Our method uses a dynamic programming algorithm to compute distinguishing patterns by aligning relevant sentences and key verbs that describe protein interactions. A matching algorithm is designed to extract the interactions between proteins. Equipped only with a protein name dictionary, our system achieves a recall rate of about 80.0% and a precision rate of about 80.5%. | Discovering patterns to extract protein-protein interactions from full biomedical texts |
d2827736 | Subcategorization information is a useful feature in dependency parsing. In this paper, we explore a method of incorporating this information via Combinatory Categorial Grammar (CCG) categories from a supertagger. We experiment with two popular dependency parsers (Malt and MST) for two languages: English and Hindi. For both languages, CCG categories improve the overall accuracy of both parsers by around 0.3-0.5% in all experiments. For both parsers, we see larger improvements specifically on dependencies at which they are known to be weak: long distance dependencies for Malt, and verbal arguments for MST. The result is particularly interesting in the case of the fast greedy parser (Malt), since improving its accuracy without significantly compromising speed is relevant for large scale applications such as parsing the web. | Improving Dependency Parsers using Combinatory Categorial Grammar |
d52137147 | We introduce a novel multi-source technique for incorporating source syntax into neural machine translation using linearized parses. This is achieved by employing separate encoders for the sequential and parsed versions of the same source sentence; the resulting representations are then combined using a hierarchical attention mechanism. The proposed model improves over both seq2seq and parsed baselines by over 1 BLEU on the WMT17 English→German task. Further analysis shows that our multi-source syntactic model is able to translate successfully without any parsed input, unlike standard parsed methods. In addition, performance does not deteriorate as much on long sentences as for the baselines. | Multi-Source Syntactic Neural Machine Translation |
d3514091 | Neural sequence-to-sequence networks with attention have achieved remarkable performance for machine translation. One of the reasons for their effectiveness is their ability to capture relevant source-side contextual information at each time-step prediction through an attention mechanism. However, the target-side context is solely based on the sequence model which, in practice, is prone to a recency bias and lacks the ability to capture effectively nonsequential dependencies among words. To address this limitation, we propose a target-sideattentive residual recurrent network for decoding, where attention over previous words contributes directly to the prediction of the next word. The residual learning facilitates the flow of information from the distant past and is able to emphasize any of the previously translated words, hence it gains access to a wider context. The proposed model outperforms a neural MT baseline as well as a memory and self-attention network on three language pairs. The analysis of the attention learned by the decoder confirms that it emphasizes a wider context, and that it captures syntactic-like structures. | Self-Attentive Residual Decoder for Neural Machine Translation |
d16132433 | We present a simple, data-driven approach to generation from knowledge bases (KB).A key feature of this approach is that grammar induction is driven by the extended domain of locality principle of TAG (Tree Adjoining Grammar); and that it takes into account both syntactic and semantic information. The resulting extracted TAG includes a unification based semantics and can be used by an existing surface realiser to generate sentences from KB data. Experimental evaluation on the KBGen data shows that our model outperforms a data-driven generate-and-rank approach based on an automatically induced probabilistic grammar; and is comparable with a handcrafted symbolic approach. | Surface Realisation from Knowledge-Bases |
d264038781 | La compréhension du langage naturel et parlé (NLU/SLU) couvre le problème d'extraire et d'annoter la structure sémantique, à partir des énoncés des utilisateurs dans le contexte des interactions humain/machine, telles que les systèmes de dialogue.Elle se compose souvent de deux tâches principales : la détection des intentions et la classification des concepts.Dans cet article, différents corpora SLU sont étudiés au niveau formel et sémantique : leurs différents formats d'annotations (à plat et structuré) et leurs ontologies ont été comparés et discutés.Avec leur pouvoir expressif gardant la hiérarchie sémantique entre les intentions et les concepts, les représentations sémantiques structurées sous forme de graphe ont été mises en exergue.En se positionnant vis à vis de la littérature et pour les futures études, une projection sémantique et une modification au niveau de l'ontologie du corpus MultiWOZ ont été proposées. | Les jeux de données en compréhension du langage naturel et parlé : paradigmes d'annotation et représentations sémantiques |
d264038830 | This paper proposes a novel approach to French patent classification leveraging data-centric strategies.We compare different approaches for the two deepest levels of the IPC hierarchy : the IPC group and subgroups.Our experiments show that while simple ensemble strategies work for shallower levels, deeper levels require more sophisticated techniques such as data augmentation, clustering, and negative sampling.Our research highlights the importance of language-specific features and data-centric strategies for accurate and reliable French patent classification.It provides valuable insights and solutions for researchers and practitioners in the field of patent classification, advancing research in French patent classification.RÉSUMÉExploration des stratégies centrées sur les données pour la classification des brevets français : Une base de référence et des comparaisons Cet article propose une nouvelle approche de classification des brevets français qui s'appuie sur des stratégies centrées sur les données.Nous comparons différentes approches pour les deux niveaux les plus profonds de la hiérarchie IPC : le groupe IPC et les sous-groupes.Nos expériences montrent que les stratégies d'ensemble simples fonctionnent pour les niveaux peu profonds, mais que les niveaux profonds nécessitent des techniques plus sophistiquées telles que l'augmentation de données, le regroupement et l'échantillonnage négatif.Notre recherche met en évidence l'importance des caractéristiques spécifiques à la langue et des stratégies centrées sur les données pour une classification précise et fiable des brevets français.Elle fournit des informations et des solutions précieuses pour les chercheurs et les praticiens dans le domaine de la classification des brevets, en faisant progresser la recherche en classification des brevets en français. | Exploring Data-Centric Strategies for French Patent Classification: A Baseline and Comparisons |
d263609622 | Most languages in the world do not have sufficient data available to develop neural-networkbased natural language generation (NLG) systems.To alleviate this resource scarcity, we propose a novel challenge for the NLG community: low-resource language corpus development (LOWRECORP).We present an innovative framework to collect a single dataset with dual tasks to maximize the efficiency of data collection efforts and respect language consultant time.Specifically, we focus on a textchat-based interface for two generation tasksconversational response generation grounded in a source document and/or image and dialogue summarization (from the former task).The goal of this shared task is to collectively develop grounded datasets for local and lowresourced languages.To enable data collection, we make available web-based software that can be used to collect these grounded conversations and summaries.Submissions will be assessed for the size, complexity, and diversity of the corpora to ensure quality control of the datasets as well as any enhancements to the interface or novel approaches to grounding conversations. | LOWRECORP: the Low-Resource NLG Corpus Building Challenge |
d9877558 | We study the task of entity linking for tweets, which tries to associate each mention in a tweet with a knowledge base entry. Two main challenges of this task are the dearth of information in a single tweet and the rich entity mention variations. To address these challenges, we propose a collective inference method that simultaneously resolves a set of mentions. Particularly, our model integrates three kinds of similarities, i.e., mention-entry similarity, entry-entry similarity, and mention-mention similarity, to enrich the context for entity linking, and to address irregular mentions that are not covered by the entity-variation dictionary. We evaluate our method on a publicly available data set and demonstrate the effectiveness of our method. | Entity Linking for Tweets |
d258378356 | Hate speech detection in online platforms has been widely studied in the past. Most of these works were conducted in English and a few rich-resource languages. Recent approaches tailored for low-resource languages have explored the interests of zero-shot cross-lingual transfer learning models in resource-scarce scenarios. However, languages variations between geolects such as American English and British English, Latin-American Spanish, and European Spanish is still a problem for NLP models that often relies on (latent) lexical information for their classification tasks. More importantly, the cultural aspect, crucial for hate speech detection, is often overlooked. | Analyzing Zero-Shot transfer Scenarios across Spanish variants for Hate Speech Detection |
d14321437 | Access to word-sentiment associations is useful for many applications, including sentiment analysis, stance detection, and linguistic analysis. However, manually assigning finegrained sentiment association scores to words has many challenges with respect to keeping annotations consistent. We apply the annotation technique of Best-Worst Scaling to obtain real-valued sentiment association scores for words and phrases in three different domains: general English, English Twitter, and Arabic Twitter. We show that on all three domains the ranking of words by sentiment remains remarkably consistent even when the annotation process is repeated with a different set of annotators. We also, for the first time, determine the minimum difference in sentiment association that is perceptible to native speakers of a language. | Capturing Reliable Fine-Grained Sentiment Associations by Crowdsourcing and Best-Worst Scaling |
d259924559 | In any system that uses structured knowledge graph (KG) data as its underlying knowledge representation, KG-to-text generation is a useful tool for turning parts of the graph data into text that can be understood by humans.Recent work has shown that models that make use of pretraining on large amounts of text data can perform well on the KG-to-text task, even with relatively little training data on the specific graph-to-text task.In this paper, we build on this concept by using large language models to perform zero-shot generation based on nothing but the model's understanding of the triple structure from what it can read.We show that ChatGPT achieves near state-of-the-art performance on some measures of the WebNLG 2020 challenge, but falls behind on others.Additionally, we compare factual, counter-factual and fictional statements, and show that there is a significant connection between what the LLM already knows about the data it is parsing and the quality of the output text. | Using Large Language Models for Zero-Shot Natural Language Generation from Knowledge Graphs |
d5616520 | The idea for this special issue came up during the preparations of the International Workshop on Finite-State Methods in Natural Language Processing, that was held at Bilkent University in Ankara, Turkey in the summer of 1998. The number of the submissions had exceeded our initial expectations and we were able to select quite a good set of papers from those submitted. Further, the workshop and the preceding tutorial by Kenneth Beesley, on finite-state methods, was attended by quite a large number of participants. This led us to believe that interest in the theory and applications of finitestate machinery was alive and well, and that some of the papers from this workshop along with further additional submissions could make a very good special issue for this journal. The five papers in this issue are the result of this process.The last decade has seen a quite a substantial surge in the use of finite-state methods in all aspects of natural language applications. Fueled by the theoretical contributions of Kaplan and Kay (1994), Mohri's recent contributions on the use of finite-state techniques in various NLP problems(Mohri 1996(Mohri , 1997, the success of finite-state approaches especially in computational morphology, for example,Koskenniemi (1983),Karttunen (1983), andKarttunen, Kaplan, andZaenen (1992), and, finally, the availability of state-of-the-art tools for building and manipulating large-scale finite-state systems(Karttunen 1993;Karttunen and Beesley 1992;Karttunen et al. 1996;Mohri, Pereira, and Riley 1998;van Noord 1999), recent years have seen many successful applications of finite-state approaches in tagging, spell checking, information extraction, parsing, speech recognition, and text-to-speech applications. This is a remarkable comeback considering that in the dawn of modern linguistics (Chomsky 1957), finitestate grammars were dismissed as fundamentally inadequate. As a result, most of the work in computational linguistics in the past few decades has been focused on far more powerful formalisms.Recent publications on finite-state technology include two collections of papers (Roche and Schabes 1997; Kornai 1999) with contributions covering a wide range of these topics. This special issue, we hope, will add to these contributions.The five papers in this collection cover many aspects of finite-state theory and applications. The papers Treatment of Epsilon Moves in Subset Construction by van Noord and Incremental Construction of Minimal Acyclic Finite-State Automata and Transducers by Daciuk, Watson, Watson, and Mihov, address two fundamental aspects in the construction of finite-state recognizers. Van Noord presents results for various methods for producing a deterministic automaton with no epsilon transitions from a nondeterministic automaton with a large number of epsilon transitions, especially those resulting from finite-state approximations of context-free and more powerful formalisms. Daciuk et al. present a new method for constructing minimal, deterministic, acyclic 6, chemin de Maupertuis, | Introduction to the Special Issue on Finite-State Methods in NLP |
d11975567 | We describe robustness techniques used in the Com-mandTalk system at: the recognition level, the parsing level, and th~ dia16gue level, and how these were influenced by the lack of domain data. We used interviews with subject matter experts (SME's) to develop a single grammar for recognition, understanding, and generation, thus eliminating the need for a robust parser. We broadened the coverage of the recognition grammar by allowing word insertions and deletions, and we implemented clarification and correction subdialogues to increase robustness at the dialogue level. We discuss the applicability of these techniques to other domains. | Building a Robust Dialogue System with Limited Data * |
d44122471 | While large-scale knowledge graphs provide vast amounts of structured facts about entities, a short textual description can often be useful to succinctly characterize an entity and its type. Unfortunately, many knowledge graph entities lack such textual descriptions. In this paper, we introduce a dynamic memory-based network that generates a short open vocabulary description of an entity by jointly leveraging induced fact embeddings as well as the dynamic context of the generated sequence of words. We demonstrate the ability of our architecture to discern relevant information for more accurate generation of type description by pitting the system against several strong baselines. | Generating Fine-Grained Open Vocabulary Entity Type Descriptions |
d7726885 | Unlexicalized probabilistic context-free parsing is a general and flexible approach that sometimes reaches competitive results in multilingual dependency parsing even if a minimum of language-specific information is supplied. Furthermore, integrating parser results (good at long dependencies) and tagger results (good at short range dependencies, and more easily adaptable to treebank peculiarities) gives competitive results in all languages. | Language Independent Probabilistic Context-Free Parsing Bolstered by Machine Learning |
d201304248 | Bidirectional Encoder Representations from Transformers (BERT; Devlin et al. 2019) represents the latest incarnation of pretrained language models which have recently advanced a wide range of natural language processing tasks. In this paper, we showcase how BERT can be usefully applied in text summarization and propose a general framework for both extractive and abstractive models. We introduce a novel document-level encoder based on BERT which is able to express the semantics of a document and obtain representations for its sentences. Our extractive model is built on top of this encoder by stacking several intersentence Transformer layers. For abstractive summarization, we propose a new fine-tuning schedule which adopts different optimizers for the encoder and the decoder as a means of alleviating the mismatch between the two (the former is pretrained while the latter is not). We also demonstrate that a two-staged fine-tuning approach can further boost the quality of the generated summaries. Experiments on three datasets show that our model achieves stateof-the-art results across the board in both extractive and abstractive settings. 1 | Text Summarization with Pretrained Encoders |
d21725995 | We propose a novel data augmentation for labeled sentences called contextual augmentation. We assume an invariance that sentences are natural even if the words in the sentences are replaced with other words with paradigmatic relations. We stochastically replace words with other words that are predicted by a bi-directional language model at the word positions. Words predicted according to a context are numerous but appropriate for the augmentation of the original words. Furthermore, we retrofit a language model with a label-conditional architecture, which allows the model to augment sentences without breaking the label-compatibility. Through the experiments for six various different text classification tasks, we demonstrate that the proposed method improves classifiers based on the convolutional or recurrent neural networks. | Contextual Augmentation: Data Augmentation by Words with Paradigmatic Relations |
d8938702 | We introduce a novel sub-character architecture that exploits a unique compositional structure of the Korean language. Our method decomposes each character into a small set of primitive phonetic units called jamo letters from which character-and word-level representations are induced. The jamo letters divulge syntactic and semantic information that is difficult to access with conventional character-level units. They greatly alleviate the data sparsity problem, reducing the observation space to 1.6% of the original while increasing accuracy in our experiments. We apply our architecture to dependency parsing and achieve dramatic improvement over strong lexical baselines. | A Sub-Character Architecture for Korean Language Processing |
d218628898 | Discourse representation tree structure (DRTS) parsing is a novel semantic parsing task which has been concerned most recently. State-of-the-art performance can be achieved by a neural sequence-to-sequence model, treating the tree construction as an incremental sequence generation problem. Structural information such as input syntax and the intermediate skeleton of the partial output has been ignored in the model, which could be potentially useful for the DRTS parsing. In this work, we propose a structural-aware model at both the encoder and decoder phase to integrate the structural information, where graph attention network (GAT) is exploited for effectively modeling. Experimental results on a benchmark dataset show that our proposed model is effective and can obtain the best performance in the literature. | DRTS Parsing with Structure-Aware Encoding and Decoding |
d1354459 | We propose an end-to-end, domainindependent neural encoder-aligner-decoder model for selective generation, i.e., the joint task of content selection and surface realization. Our model first encodes a full set of over-determined database event records via an LSTM-based recurrent neural network, then utilizes a novel coarse-to-fine aligner to identify the small subset of salient records to talk about, and finally employs a decoder to generate free-form descriptions of the aligned, selected records. Our model achieves the best selection and generation results reported to-date (with 59% relative improvement in generation) on the benchmark WEATHER-GOV dataset, despite using no specialized features or linguistic resources. Using an improved k-nearest neighbor beam filter helps further. We also perform a series of ablations and visualizations to elucidate the contributions of our key model components. Lastly, we evaluate the generalizability of our model on the ROBOCUP dataset, and get results that are competitive with or better than the state-of-the-art, despite being severely data-starved. | What to talk about and how? Selective Generation using LSTMs with Coarse-to-Fine Alignment |
d226283759 | Research on hate speech classification has received increased attention. In real-life scenarios, a small amount of labeled hate speech data is available to train a reliable classifier. Semi-supervised learning takes advantage of a small amount of labeled data and a large amount of unlabeled data. In this paper, label propagation-based semi-supervised learning is explored for the task of hate speech classification. The quality of labeling the unlabeled set depends on the input representations. In this work, we show that pre-trained representations are label agnostic, and when used with label propagation yield poor results. Neural network-based fine-tuning can be adopted to learn task-specific representations using a small amount of labeled data. We show that fully fine-tuned representations may not always be the best representations for the label propagation and intermediate representations may perform better in a semi-supervised setup. | Label Propagation-Based Semi-Supervised Learning for Hate Speech Classification |
d215827207 | This paper describes the University of Edinburgh's submissions to the WMT17 shared news translation and biomedical translation tasks. We participated in 12 translation directions for news, translating between English and Czech, German, Latvian, Russian, Turkish and Chinese. For the biomedical task we submitted systems for English to Czech, German, Polish and Romanian. Our systems are neural machine translation systems trained with Nematus, an attentional encoder-decoder. We follow our setup from last year and build BPE-based models with parallel and back-translated monolingual training data.Novelties this year include the use of deep architectures, layer normalization, and more compact models due to weight tying and improvements in BPE segmentations. We perform extensive ablative experiments, reporting on the effectivenes of layer normalization, deep architectures, and different ensembling techniques. | The University of Edinburgh's Neural MT Systems for WMT17 |
d44158569 | Dynamic oracles provide strong supervision for training constituency parsers with exploration, but must be custom defined for a given parser's transition system. We explore using a policy gradient method as a parser-agnostic alternative. In addition to directly optimizing for a tree-level metric such as F1, policy gradient has the potential to reduce exposure bias by allowing exploration during training; moreover, it does not require a dynamic oracle for supervision. On four constituency parsers in three languages, the method substantially outperforms static oracle likelihood training in almost all settings. For parsers where a dynamic oracle is available (including a novel oracle which we define for the transition system of Dyer et al.(2016)), policy gradient typically recaptures a substantial fraction of the performance gain afforded by the dynamic oracle. | Policy Gradient as a Proxy for Dynamic Oracles in Constituency Parsing |
d6570134 | In this paper, we describe a system which models a set of concurrent processes that are encountered in a typical office environment, using a body of explicitly sequenced production rules. The system employs an interval-based temporal network for storing historical information. A text planning module traverses this network to search for events which need to be mentioned in a coherent report describing the current status of the system. In addition, the planner also combines similar information for succinct presentation whenever applicable. Finally, we elaborate on how we adapt an available generation module to produce wellstructured textual report for our chosen domain.1. | AUTOMATICALLY GENERATING NATURAL LANGUAGE REPORTS IN AN OFFICE ENVIRONMENT |
d29051190 | In this paper, we introduce YEDDA, a lightweight but efficient open-source tool for text span annotation. YEDDA provides a systematic solution for text span annotation, ranging from collaborative user annotation to administrator evaluation and analysis. It overcomes the low efficiency of traditional text annotation tools by annotating entities through both command line and shortcut keys, which are configurable with custom labels. YEDDA also gives intelligent recommendations by training a predictive model using the up-to-date annotated text. An administrator client is developed to evaluate annotation quality of multiple annotators and generate detailed comparison report for each annotator pair. YEDDA is developed based on Tkinter and is compatible with all major operating systems. | YEDDA: A Lightweight Collaborative Text Span Annotation Tool |
d11174813 | We propose a novel approach to learning distributed representations of variable-length text sequences in multiple languages simultaneously. Unlike previous work which often derive representations of multi-word sequences as weighted sums of individual word vectors, our model learns distributed representations for phrases and sentences as a whole. Our work is similar in spirit to the recent paragraph vector approach but extends to the bilingual context so as to efficiently encode meaning-equivalent text sequences of multiple languages in the same semantic space. Our learned embeddings achieve state-of-theart performance in the often used crosslingual document classification task (CLDC) with an accuracy of 92.7 for English to German and 91.5 for German to English. By learning text sequence representations as a whole, our model performs equally well in both classification directions in the CLDC task in which past work did not achieve. | Learning Distributed Representations for Multilingual Text Sequences |
d12294387 | Emotion cause extraction aims to identify the reasons behind a certain emotion expressed in text. It is a much more difficult task compared to emotion classification. Inspired by recent advances in using deep memory networks for question answering (QA), we propose a new approach which considers emotion cause identification as a reading comprehension task in QA. Inspired by convolutional neural networks, we propose a new mechanism to store relevant context in different memory slots to model context information. Our proposed approach can extract both word level sequence features and lexical features. Performance evaluation shows that our method achieves the state-of-the-art performance on a recently released emotion cause dataset, outperforming a number of competitive baselines by at least 3.01% in F-measure. | A Question Answering Approach to Emotion Cause Extraction |
d30601989 | In neural text generation such as neural machine translation, summarization, and image captioning, beam search is widely used to improve the output text quality. However, in the neural generation setting, hypotheses can finish in different steps, which makes it difficult to decide when to end beam search to ensure optimality. We propose a provably optimal beam search algorithm that will always return the optimal-score complete hypothesis (modulo beam size), and finish as soon as the optimality is established (finishing no later than the baseline). To counter neural generation's tendency for shorter hypotheses, we also introduce a bounded length reward mechanism which allows a modified version of our beam search algorithm to remain optimal. Experiments on neural machine translation demonstrate that our principled beam search algorithm leads to improvement in BLEU score over previously proposed alternatives. | When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size) |
d8463166 | We present the architectural design rationale of a Sanskrit computational linguistics platform, where the lexical database has a central role. We explain the structuring requirements issued from the interlinking of grammatical tools through its hypertext rendition. | Design of a Lexical Database for Sanskrit |
d6591541 | We present a discriminative, latent variable approach to syntactic parsing in which rules exist at multiple scales of refinement. The model is formally a latent variable CRF grammar over trees, learned by iteratively splitting grammar productions (not categories). Different regions of the grammar are refined to different degrees, yielding grammars which are three orders of magnitude smaller than the single-scale baseline and 20 times smaller than the split-and-merge grammars ofPetrov et al. (2006). In addition, our discriminative approach integrally admits features beyond local tree configurations. We present a multiscale training method along with an efficient CKY-style dynamic program. On a variety of domains and languages, this method produces the best published parsing accuracies with the smallest reported grammars. | Sparse Multi-Scale Grammars for Discriminative Latent Variable Parsing |
d32533948 | Tree-structured neural network architectures for sentence encoding draw inspiration from the approach to semantic composition generally seen in formal linguistics, and have shown empirical improvements over comparable sequence models by doing so. Moreover, adding multiplicative interaction terms to the composition functions in these models can yield significant further improvements. However, existing compositional approaches that adopt such a powerful composition function scale poorly, with parameter counts exploding as model dimension or vocabulary size grows. We introduce the Lifted Matrix-Space model, which uses a global transformation to map vector word embeddings to matrices, which can then be composed via an operation based on matrix-matrix multiplication. Its composition function effectively transmits a larger number of activations across layers with relatively few model parameters. We evaluate our model on the Stanford NLI corpus, the Multi-Genre NLI corpus, and the Stanford Sentiment Treebank and find that it consistently outperforms TreeLSTM(Tai et al., 2015), the previous best known composition function for treestructured models. | The Lifted Matrix-Space Model for Semantic Composition |
d29170217 | Human communication includes information, opinions and reactions. Reactions are often captured by the affective-messages in written as well as verbal communications. While there has been work in affect modeling and to some extent affective content generation, the area of affective word distributions is not well studied. Synsets and lexica capture semantic relationships across words. These models, however, lack in encoding affective or emotional word interpretations. Our proposed model, Aff2Vec, provides a method for enriched word embeddings that are representative of affective interpretations of words. Aff2Vec outperforms the state-of-the-art in intrinsic word-similarity tasks. Further, the use of Aff2Vec representations outperforms baseline embeddings in downstream natural language understanding tasks including sentiment analysis, personality detection, and frustration prediction. This work is licensed under a Creative Commons Attribution 4.0 International License. License details: http:// creativecommons.org/licenses/by/4.0/ | Aff2Vec: Affect-Enriched Distributional Word Representations |
d16878067 | Computing implicit entities and events for story understanding | |
d957235 | In Czech corpora compound verb groups are usually tagged in word-by-word manner. As a consequence, some of the morphological tags of particular components of the verb group lose their original meaning. We present a method for automatic recognition of compound verb groups in Czech. From an annotated corpus 126 definite clause grammar rules were constructed. These rules describe all compound verb groups that are frequent in Czech. Using those rules we can find compound verb groups in unannotated texts with the accuracy 93%. Tagging compound verb groups in an annotated corpus exploiting the verb rules is described. | Recognition and Tagging of Compound Verb Groups in Czech |
d44105751 | An accurate abstractive summary of a document should contain all its salient information and should be logically entailed by the input document. We improve these important aspects of abstractive summarization via multi-task learning with the auxiliary tasks of question generation and entailment generation, where the former teaches the summarization model how to look for salient questioning-worthy details, and the latter teaches the model how to rewrite a summary which is a directed-logical subset of the input document. We also propose novel multitask architectures with high-level (semantic) layer-specific sharing across multiple encoder and decoder layers of the three tasks, as well as soft-sharing mechanisms (and show performance ablations and analysis examples of each contribution). Overall, we achieve statistically significant improvements over the state-ofthe-art on both the CNN/DailyMail and Gigaword datasets, as well as on the DUC-2002 transfer setup. We also present several quantitative and qualitative analysis studies of our model's learned saliency and entailment skills. 11 Note that all our soft and layer sharing decisions were strictly made on the dev/validation set (see Sec. 5). | Soft Layer-Specific Multi-Task Summarization with Entailment and Question Generation |
d218973791 | We present a 78.8-million-tweet, 1.3-billion-word corpus aimed at studying regional variation in Canadian English with a specific focus on the dialect regions of Toronto, Montreal, and Vancouver. Our data collection and filtering pipeline reflects complex design criteria, which aim to allow for both data-intensive modeling methods and user-level variationist sociolinguistic analysis. It specifically consists in identifying Twitter users from the three cities, crawling their entire timelines, filtering the collected data in terms of user location and tweet language, and automatically excluding near-duplicate content. The resulting corpus mirrors national and regional specificities of Canadian English, it provides sufficient aggregate and user-level data, and it maintains a reasonably balanced distribution of content across regions and users. The utility of this dataset is illustrated by two example applications: the detection of regional lexical and topical variation, and the identification of contact-induced semantic shifts using vector space models. In accordance with Twitter's developer policy, the corpus will be publicly released in the form of tweet IDs. | Collecting Tweets to Investigate Regional Variation in Canadian English |
d21632466 | Multi-task learning (MTL) has recently contributed to learning better representations in service of various NLP tasks. MTL aims at improving the performance of a primary task, by jointly training on a secondary task. This paper introduces automated tasks, which exploit the sequential nature of the input data, as secondary tasks in an MTL model. We explore next word prediction, next character prediction, and missing word completion as potential automated tasks. Our results show that training on a primary task in parallel with a secondary automated task improves both the convergence speed and accuracy for the primary task. We suggest two methods for augmenting an existing network with automated tasks and establish better performance in topic prediction, sentiment analysis, and hashtag recommendation. Finally, we show that the MTL models can perform well on datasets that are small and colloquial by nature. | Deep Automated Multi-task Learning |
d44278 | Labeling of sentence boundaries is a necessary prerequisite for many natural language processing tasks, including part-ofspeech tagging and sentence alignment. End-of-sentence punctuation marks are ambiguous; to disambiguate them most systems use brittle, special-purpose regular expression grammars and exception rules. As an alternative, we have developed an efficient, trainable algorithm that uses a lexicon with part-of-speech probabilities and a feed-forward neural network. This work demonstrates the feasibility of using prior probabilities of part-of-speech assignments, as opposed to words or definite part-ofspeech assignments, as contextual information. After training for less than one minute, the method correctly labels over 98.5% of sentence boundaries in a corpus of over 27,000 sentence-boundary marks. We show the method to be efficient and easily adaptable to different text genres, including single-case texts. | Adaptive Sentence Boundary Disambiguation |
d240353697 | Gigantic pre-trained models have become central to natural language processing (NLP), serving as the starting point for fine-tuning towards a range of downstream tasks. However, two pain points persist for this paradigm: (a) as the pre-trained models grow bigger (e.g., 175B parameters for GPT-3), even the fine-tuning process can be time-consuming and computationally expensive; (b) the finetuned model has the same size as its starting point by default, which is neither sensible due to its more specialized functionality, nor practical since many fine-tuned models will be deployed in resource-constrained environments. To address these pain points, we propose a framework for resource-and parameterefficient fine-tuning by leveraging the sparsity prior in both weight updates and the final model weights. Our proposed framework, dubbed Dually Sparsity-Embedded Efficient Tuning (DSEE), aims to achieve two key objectives: (i) parameter efficient fine-tuning -by enforcing sparsity-aware low-rank updates on top of the pre-trained weights; and (ii) resourceefficient inference -by encouraging a sparse weight structure towards the final fine-tuned model. We leverage sparsity in these two directions by exploiting both unstructured and structured sparse patterns in pre-trained language models via a unified approach. Extensive experiments and in-depth investigations, with diverse network backbones (i.e., BERT, RoBERTa, and GPT-2) on dozens of datasets, consistently demonstrate impressive parameter-/inference-efficiency, while maintaining competitive downstream performance. For instance, DSEE saves about 25% inference FLOPs while achieving comparable performance, with 0.5% trainable parameters on BERT. Codes are available at https://github.com/VITA-Group/ DSEE. | DSEE: Dually Sparsity-embedded Efficient Tuning of Pre-trained Language Models |
d3265576 | ||
d10673026 | Word alignment is an important preprocessing step for machine translation. The project aims at incorporating manual alignments from Amazon Mechanical Turk (MTurk) to help improve word alignment quality. As a global crowdsourcing service, MTurk can provide flexible and abundant labor force and therefore reduce the cost of obtaining labels. An easyto-use interface is developed to simplify the labeling process. We compare the alignment results by Turkers to that by experts, and incorporate the alignments in a semi-supervised word alignment tool to improve the quality of the labels. We also compared two pricing strategies for word alignment task. Experimental results show high precision of the alignments provided by Turkers and the semi-supervised approach achieved 0.5% absolute reduction on alignment error rate. | Consensus versus Expertise : A Case Study of Word Alignment with Mechanical Turk |
d9662636 | Existing image captioning models do not generalize well to out-of-domain images containing novel scenes or objects. This limitation severely hinders the use of these models in real world applications dealing with images in the wild. We address this problem using a flexible approach that enables existing deep captioning architectures to take advantage of image taggers at test time, without re-training. Our method uses constrained beam search to force the inclusion of selected tag words in the output, and fixed, pretrained word embeddings to facilitate vocabulary expansion to previously unseen tag words. Using this approach we achieve state of the art results for out-of-domain captioning on MSCOCO (and improved results for in-domain captioning). Perhaps surprisingly, our results significantly outperform approaches that incorporate the same tag predictions into the learning algorithm. We also show that we can significantly improve the quality of generated ImageNet captions by leveraging ground-truth labels. | Guided Open Vocabulary Image Captioning with Constrained Beam Search |
d3101974 | We propose a simple, scalable, fully generative model for transition-based dependency parsing with high accuracy. The model, parameterized by Hierarchical Pitman-Yor Processes, overcomes the limitations of previous generative models by allowing fast and accurate inference. We propose an efficient decoding algorithm based on particle filtering that can adapt the beam size to the uncertainty in the model while jointly predicting POS tags and parse trees. The UAS of the parser is on par with that of a greedy discriminative baseline. As a language model, it obtains better perplexity than a n-gram model by performing semi-supervised learning over a large unlabelled corpus. We show that the model is able to generate locally and syntactically coherent sentences, opening the door to further applications in language generation. | A Bayesian Model for Generative Transition-based Dependency Parsing |
d67856005 | We introduce a novel method for multilingual transfer that utilizes deep contextual embeddings, pretrained in an unsupervised fashion. While contextual embeddings have been shown to yield richer representations of meaning compared to their static counterparts, aligning them poses a challenge due to their dynamic nature. To this end, we construct context-independent variants of the original monolingual spaces and utilize their mapping to derive an alignment for the contextdependent spaces. This mapping readily supports processing of a target language, improving transfer by context-aware embeddings. Our experimental results demonstrate the effectiveness of this approach for zero-shot and few-shot learning of dependency parsing. Specifically, our method consistently outperforms the previous state-of-the-art on 6 tested languages, yielding an improvement of 6.8 LAS points on average. 1 * Equal contribution 1 Code and models:https://github.com/ TalSchuster/CrossLingualELMo. | Cross-Lingual Alignment of Contextual Word Embeddings, with Applications to Zero-shot Dependency Parsing |
d4054187 | Training models for the automatic correction of machine-translated text usually relies on data consisting of (source, MT, human postedit) triplets providing, for each source sentence, examples of translation errors with the corresponding corrections made by a human post-editor. Ideally, a large amount of data of this kind should allow the model to learn reliable correction patterns and effectively apply them at test stage on unseen (source, MT) pairs. In practice, however, their limited availability calls for solutions that also integrate in the training process other sources of knowledge. Along this direction, state-of-the-art results have been recently achieved by systems that, in addition to a limited amount of available training data, exploit artificial corpora that approximate elements of the "gold" training instances with automatic translations. Following this idea, we present eSCAPE, the largest freely-available Synthetic Corpus for Automatic Post-Editing released so far. eSCAPE consists of millions of entries in which the MT element of the training triplets has been obtained by translating the source side of publicly-available parallel corpora, and using the target side as an artificial human post-edit. Translations are obtained both with phrase-based and neural models. For each MT paradigm, eSCAPE contains 7.2 million triplets for English-German and 3.3 millions for English-Italian, resulting in a total of 14,4 and 6,6 million instances respectively. The usefulness of eSCAPE is proved through experiments in a general-domain scenario, the most challenging one for automatic post-editing. For both language directions, the models trained on our artificial data always improve MT quality with statistically significant gains. The current version of eSCAPE can be freely downloaded from: http://hltshare.fbk.eu/QT21/eSCAPE.html. | eSCAPE: a Large-scale Synthetic Corpus for Automatic Post-Editing |
d10296349 | Building an accurate Named Entity Recognition (NER) system for languages with complex morphology is a challenging task. In this paper, we present research that explores the feature space using both gold and bootstrapped noisy features to build an improved highly accurate Arabic NER system. We bootstrap noisy features by projection from an Arabic-English parallel corpus that is automatically tagged with a baseline NER system. The feature space covers lexical, morphological, and syntactic features. The proposed approach yields an improvement of up to 1.64 F-measure (absolute). | Arabic Named Entity Recognition: Using Features Extracted from Noisy Data |
d988010 | We present a novel graph-based summarization framework (Opinosis) that generates concise abstractive summaries of highly redundant opinions. Evaluation results on summarizing user reviews show that Opinosis summaries have better agreement with human summaries compared to the baseline extractive method. The summaries are readable, reasonably well-formed and are informative enough to convey the major opinions. | Opinosis: A Graph-Based Approach to Abstractive Summarization of Highly Redundant Opinions |
d27544235 | Learning algorithms for natural language processing (NLP) tasks traditionally rely on manually defined appropriate contextual features. On the other hand, neural network models learn these features automatically and have been successfully applied for several NLP tasks. Such models only consider vector representation of words and thus do not require efforts for manual feature engineering. This makes neural models a natural choice to be used across several domains. But this flexibility comes at the cost of interpretability. The motivation of this work is to enhance understanding of neural models towards their ability to capture contextual features. In particular, we analyze the performance of bi-directional recurrent neural models for sequence tagging task by defining several measures based on word erasure technique and investigate their ability to capture relevant features. We perform a comprehensive analysis of these measures on general as well as biomedical domain datasets. Our experiments focus on important contextual words as features, which can easily be extended to analyze various other feature types. Not only this, we also investigate positional effects of context words and show how the developed methods can be used for error analysis. | Investigating how well contextual features are captured by bi-directional recurrent neural network models |
d863202 | This paper 1 presents our approach to the problem of single sentence summarisation. We investigate the use of Singular Value Decomposition (SVD) to guide the generation of a summary towards the theme that is the focus of the document to be summarised. In doing so, the intuition is that the generated summary will more accurately reflect the content of the source document. Currently, we operate in the news domain and at present, our summaries are modelled on headlines. This paper presents SVD as an alternative method to determine if a word is a suitable candidate for inclusion in the headline. The results of a recall based evaluation comparing three different strategies to word selection, indicate that thematic information does help improve recall. | Straight to the Point: Discovering Themes for Summary Generation |
d15518351 | Language models can be formalized as loglinear regression models where the input features represent previously observed contexts up to a certain length m. The complexity of existing algorithms to learn the parameters by maximum likelihood scale linearly in nd, where n is the length of the training corpus and d is the number of observed features. We present a model that grows logarithmically in d, making it possible to efficiently leverage longer contexts. We account for the sequential structure of natural language using treestructured penalized objectives to avoid overfitting and achieve better generalization. | Structured Penalties for Log-linear Language Models |
d29245285 | Attentional sequence-to-sequence models have become the new standard for machine translation, but one challenge of such models is a significant increase in training and decoding cost compared to phrase-based systems. Here, we focus on efficient decoding, with a goal of achieving accuracy close the state-of-the-art in neural machine translation (NMT), while achieving CPU decoding speed/throughput close to that of a phrasal decoder.We approach this problem from two angles: First, we describe several techniques for speeding up an NMT beam search decoder, which obtain a 4.4x speedup over a very efficient baseline decoder without changing the decoder output. Second, we propose a simple but powerful network architecture which uses an RNN (GRU/LSTM) layer at bottom, followed by a series of stacked fully-connected layers applied at every timestep. This architecture achieves similar accuracy to a deep recurrent model, at a small fraction of the training and decoding cost. By combining these techniques, our best system achieves a very competitive accuracy of 38.3 BLEU on WMT English-French NewsTest2014, while decoding at 100 words/sec on single-threaded CPU. We believe this is the best published accuracy/speed trade-off of an NMT system. | Sharp Models on Dull Hardware: Fast and Accurate Neural Machine Translation Decoding on the CPU |
d13747533 | In this paper we introduce the notion of Demand-Weighted Completeness, allowing estimation of the completeness of a knowledge base with respect to how it is used. Defining an entity by its classes, we employ usage data to predict the distribution over relations for that entity. For example, instances of person in a knowledge base may require a birth date, name and nationality to be considered complete. These predicted relation distributions enable detection of important gaps in the knowledge base, and define the required facts for unseen entities. Such characterisation of the knowledge base can also quantify how usage and completeness change over time. We demonstrate a method to measure Demand-Weighted Completeness, and show that a simple neural network model performs well at this prediction task. | Demand-Weighted Completeness Prediction for a Knowledge Base |
d13661068 | Character-level Neural Machine Translation (NMT) models have recently achieved impressive results on many language pairs. They mainly do well for Indo-European language pairs, where the languages share the same writing system. However, for translating between Chinese and English, the gap between the two different writing systems poses a major challenge because of a lack of systematic correspondence between the individual linguistic units. In this paper, we enable character-level NMT for Chinese, by breaking down Chinese characters into linguistic units similar to that of Indo-European languages. We use the Wubi encoding scheme 1 , which preserves the original shape and semantic information of the characters, while also being reversible. We show promising results from training Wubi-based models on the characterand subword-level with recurrent as well as convolutional models. | Character-level Chinese-English Translation through ASCII Encoding |
d196213784 | While most neural machine translation (NMT) systems are still trained using maximum likelihood estimation, recent work has demonstrated that optimizing systems to directly improve evaluation metrics such as BLEU can substantially improve final translation accuracy. However, training with BLEU has some limitations: it doesn't assign partial credit, it has a limited range of output values, and it can penalize semantically correct hypotheses if they differ lexically from the reference. In this paper, we introduce an alternative reward function for optimizing NMT systems that is based on recent work in semantic similarity. We evaluate on four disparate languages translated to English, and find that training with our proposed metric results in better translations as evaluated by BLEU, semantic similarity, and human evaluation, and also that the optimization procedure converges faster. Analysis suggests that this is because the proposed metric is more conducive to optimization, assigning partial credit and providing more diversity in scores than BLEU. 1 | Beyond BLEU: Training Neural Machine Translation with Semantic Similarity |
d48360450 | In this paper, we analyze several neural network designs (and their variations) for sentence pair modeling and compare their performance extensively across eight datasets, including paraphrase identification, semantic textual similarity, natural language inference, and question answering tasks. Although most of these models have claimed state-of-the-art performance, the original papers often reported on only one or two selected datasets. We provide a systematic study and show that (i) encoding contextual information by LSTM and inter-sentence interactions are critical, (ii) Tree-LSTM does not help as much as previously claimed but surprisingly improves performance on Twitter datasets, (iii) the Enhanced Sequential Inference Model (Chen et al., 2017) is the best so far for larger datasets, while the Pairwise Word Interaction Model (He and Lin, 2016) achieves the best performance when less data is available. We release our implementations as an open-source toolkit.This work is licensed under a Creative Commons Attribution 4.0 International License. License details: http:// creativecommons.org/licenses/by/4.0/. | Neural Network Models for Paraphrase Identification, Semantic Textual Similarity, Natural Language Inference, and Question Answering |
d195345388 | Attention mechanisms have seen some success for natural language processing downstream tasks in recent years and generated new stateof-the-art results. A thorough evaluation of the attention mechanism for the task of Argumentation Mining is missing. With this paper, we report a comparative evaluation of attention layers in combination with a bidirectional long short-term memory network, which is the current state-of-the-art approach for the unit segmentation task. We also compare sentencelevel contextualized word embeddings to pregenerated ones. Our findings suggest that for this task, the additional attention layer does not improve the performance. In most cases, contextualized embeddings do also not show an improvement on the score achieved by predefined embeddings. | Is It Worth the Attention? A Comparative Evaluation of Attention Layers for Argument Unit Segmentation |
d1801525 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.