_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d244464101 | In the pandemic period, the stay-at-home trend forced businesses to switch their activities to digital mode, for example, app-based payment methods, social distancing via social media platforms, and other digital means have become an integral part of our lives. Sentiment analysis of textual information in user comments is a topical task in emotion AI because user comments or reviews are not homogeneous, they contain sparse context behind, and are misleading both for human and computer. Barriers arise from the emotional language enriched with slang, peculiar spelling, transliteration, use of emoji and their symbolic counterparts, and codeswitching. | |
d8141159 | We have developed an algorithm for the automatic conversion of dictated English sentences to written text, with essentially no restriction on the nature of the material dictated. We require that speakers undergo a short training session so that the system can adapt to their individual speaking characteristics and that they leave brief pauses between words. We have tested our algorithm extensively on an 86,000 word vocabulary (the largest of any such system in the world) using nine speakers and obtained word recognition rates on the order of 93Uo. | An 86,000-Word Recognizer Based on Phonemic Models |
d247315159 | Progress in natural language processing research is catalyzed by the possibilities given by the widespread software frameworks. This paper introduces the AdaptOr library 1 that transposes the traditional model-centric approach composed of pre-training + fine-tuning steps to objective-centric approach, composing the training process by applications of selected objectives. We survey research directions that can benefit from enhanced objective-centric experimentation in multi-task training, custom objectives development, dynamic training curricula, or domain adaptation. AdaptOr aims to ease the reproducibility of these research directions in practice. Finally, we demonstrate the practical applicability of AdaptOr in selected unsupervised domain adaptation scenarios. | AdaptOr: Objective-Centric Adaptation Framework for Language Models |
d9009202 | This paper introduces an approach to representing the kinds of information that components in a natural language generation (NLG) system will need to communicate to one another. This information may be partial, may involve more than one level of analysis and may need to include information about the history of a derivation. We present a general representation scheme capable of handling these cases. In addition, we make a proposal for organising intermodule communication in an NLG system by having a central server for this information. We have validated the approach by a reanalysis of an existing NLG system and through a full implementation of a runnable specification. 1This work is supported by ESPRC grants GR/L77041 (Edinburgh) and GR/L77102 (Brighton), | A Representation for Complex and Evolving Data Dependencies in Generation |
d18186138 | IntroductionEarly work in text, structuring, such as [McK85] and [MTSS] showed tha,t texts of all types and genres seem to be COml)osed of a small mHnl)er of simple, intuitive units, wl.riously referred to as rhetorical relations and rhetorical predicates. These two bodies of work dii[hred in whether these units were best viewed as the bricks or the mortar of text structure, hut in either case a small set of primitives seemed to suffice for all texts.McKeown showed how a text generation system can make use of these sorts of primitives to produce coherent, informative texts. However, not: long after that, it became obvious that McKeown's schematized 1)lock-sta.cking apl~roach to generation compiled out too much information about a speaker's intentional goals in choosing the blocks s/he did, and that more of this information should he recorded in the process of text structuring to a.llow for such niceties as flexibility in • auswering follow up questions or requests for elaboration. Following this l'ea.sonillg, [MS91] instead utilized the more "mortar-centered" al)I)roach of RST for text gener:L, tion. | Observations and Directions in Text Structure |
d6698479 | DEALING WITH INCOMPLETENESS OF LINGUISTIC KNOWLEDGE IN LANGUAGE TRANSLATION TRANSFER AND GENERATION STAGE OF MU MACHINE TRANSLATION PROJECT | |
d7393907 | This paper addresses a specific case of the task of lexical acquisition understood as the induction of information about the linguistic characteristics of lexical items on the basis of information gathered from their occurrences in texts. Most of the recent works in the area of lexical acquisition have used methods that take as much textual data as possible as source of evidence, but their performance decreases notably when only few occurrences of a word are available. The importance of covering such low frequency items lies in the fact that a large quantity of the words in any particular collection of texts will be occurring few times, if not just once. Our work proposes to compensate the lack of information resorting to linguistic knowledge on the characteristics of lexical classes. This knowledge, obtained from a lexical typology, is formulated probabilistically to be used in a Bayesian method to maximize the information gathered from single occurrences as to predict the full set of characteristics of the word. Our results show that our method achieves better results than others for the treatment of low frequency items. | Automatic acquisition for low frequency lexical items |
d4946715 | Ambiguities are ubiquitous in natural language and pose a major challenge for the automatic interpretation of natural language expressions. In this paper we focus on different types of lexical ambiguities that play a role in the context of ontology-based question answering, and explore strategies for capturing and resolving them. We show that by employing underspecification techniques and by using ontological reasoning in order to filter out inconsistent interpretations as early as possible, the overall number of interpretations can be effectively reduced by 44 %. | Representing and resolving ambiguities in ontology-based question answering |
d219301030 | ||
d12201644 | The focus of this article is the integration of two different perspectives on lexical semantics: Discourse Representation Theory's (DRT) inferentially motivated approach and Semantic Emphasis Theory's (SET) lexical field based view. A new joined representation format is developed which is exemplified by analyses of German verbs. The benefits thereof are on both sides. DI/T gains basic entries for whole lexical fields and, furtherlnore, a systematic interface between semantic and syntactic argument structures. SET profits both from the much larger semantic coverage and from the fine grained lexical analyses which reflect inferential behaviour. | Discourse Semantics Meets Lexical Field Semantics |
d18600087 | Pivoting on bilingual parallel corpora is a popular approach for paraphrase acquisition. Although such pivoted paraphrase collections have been successfully used to improve the performance of several different NLP applications, it is still difficult to get an intrinsic estimate of the quality and coverage of the paraphrases contained in these collections. We present ParaQuery, a tool that helps a user interactively explore and characterize a given pivoted paraphrase collection, analyze its utility for a particular domain, and compare it to other popular lexical similarity resources -all within a single interface. | ParaQuery: Making Sense of Paraphrase Collections |
d5208632 | Internet users are keen on creating different kinds of morphs to avoid censorship, express strong sentiment or humor. For example, in Chinese social media, users often use the entity morph "方便面 (Instant Noodles)" to refer to "周永康 (Zhou Yongkang)" because it shares one character "康 (Kang)" with the well-known brand of instant noodles "康师傅 (Master Kang)". We developed a wide variety of novel approaches to automatically encode proper and interesting morphs, which can effectively pass decoding tests 1 . | Be Appropriate and Funny: Automatic Entity Morph Encoding |
d229923565 | Pre-trained Language Models (PLMs) have shown superior performance on various downstream Natural Language Processing (NLP) tasks. However, conventional pre-training objectives do not explicitly model relational facts in text, which are crucial for textual understanding. To address this issue, we propose a novel contrastive learning framework ERICA to obtain a deep understanding of the entities and their relations in text. Specifically, we define two novel pre-training tasks to better understand entities and relations: (1) the entity discrimination task to distinguish which tail entity can be inferred by the given head entity and relation; (2) the relation discrimination task to distinguish whether two relations are close or not semantically, which involves complex relational reasoning. Experimental results demonstrate that ERICA can improve typical PLMs (BERT and RoBERTa) on several language understanding tasks, including relation extraction, entity typing and question answering, especially under low-resource settings. | ERICA: Improving Entity and Relation Understanding for Pre-trained Language Models via Contrastive Learning |
d8149105 | Traditional sentiment classification methods often require polarity dictionaries or crafted features to utilize machine learning. However, those approaches incur high costs in the making of dictionaries and/or features, which hinder generalization of tasks. Examples of these approaches include an approach that uses a polarity dictionary that cannot handle unknown or newly invented words and another approach that uses a complex model with 13 types of feature templates. We propose a novel high performance sentiment classification method with stacked denoising auto-encoders that uses distributed word representation instead of building dictionaries or utilizing engineering features. The results of experiments conducted indicate that our model achieves state-of-the-art performance in Japanese sentiment classification tasks.PACLIC 29150 | Japanese Sentiment Classification with Stacked Denoising Auto-Encoder using Distributed Word Representation |
d9964247 | Linguistic Data Consortium has recently embarked on an effort to create integrated linguistic resources and related infrastructure for language exploitation technologies within the DARPA GALE (Global Autonomous Language Exploitation) Program. GALE targets an end-to-end system consisting of three major engines: Transcription, Translation and Distillation. Multilingual speech or text from a variety of genres is taken as input and English text is given as output, with information of interest presented in an integrated and consolidated fashion to the end user. GALE's goals requires a quantum leap in the performance of human language technology, while also demanding solutions that are more intelligent, more robust, more adaptable, more efficient and more integrated. LDC has responded to this challenge with a comprehensive approach to linguistic resource development designed to support GALE's research and evaluation needs and to provide lasting resources for the larger Human Language Technology community. | Integrated Linguistic Resources for Language Exploitation Technologies |
d252089931 | While Transformers have had significant success in paragraph generation, they treat sentences as linear sequences of tokens and often neglect their hierarchical information. Prior work has shown that decomposing the levels of granularity (e.g., word, phrase, or sentence) for input tokens has produced substantial improvements, suggesting the possibility of enhancing Transformers via more fine-grained modeling of granularity. In this work, we propose continuous decomposition of granularity for neural paraphrase generation (C-DNPG). In order to efficiently incorporate granularity into sentence encoding, C-DNPG introduces a granularity-aware attention (GA-Attention) mechanism which extends the multi-head selfattention with: 1) a granularity head that automatically infers the hierarchical structure of a sentence by neurally estimating the granularity level of each input token; and 2) two novel attention masks, namely, granularity resonance and granularity scope, to efficiently encode granularity into attention. Experiments on two benchmarks, including Quora question pairs and Twitter URLs have shown that C-DNPG outperforms baseline models by a remarkable margin and achieves the state-of-the-art results in terms of many metrics. Qualitative analysis reveals that C-DNPG indeed captures finegrained levels of granularity with effectiveness. | Continuous Decomposition of Granularity for Neural Paraphrase Generation |
d247410985 | Contrastive learning has shown great potential in unsupervised sentence embedding tasks, e.g., SimCSE (Gao et al., 2021). However, We find that these existing solutions are heavily affected by superficial features like the length of sentences or syntactic structures. In this paper, we propose a semantics-aware contrastive learning framework for sentence embeddings, termed Pseudo-Token BERT (PT-BERT), which is able to exploit the pseudotoken space (i.e., latent semantic space) representation of a sentence while eliminating the impact of superficial features such as sentence length and syntax. Specifically, we introduce an additional pseudo token embedding layer independent of the BERT encoder to map each sentence into a sequence of pseudo tokens in a fixed length. Leveraging these pseudo sequences, we are able to construct same-length positive and negative pairs based on the attention mechanism to perform contrastive learning. In addition, we utilize both the gradientupdating and momentum-updating encoders to encode instances while dynamically maintaining an additional queue to store the representation of sentence embeddings, enhancing the encoder's learning performance for negative examples. Experiments show that our model outperforms the state-of-the-art baselines on six standard semantic textual similarity (STS) tasks. Furthermore, experiments on alignments and uniformity losses, as well as hard examples with different sentence lengths and syntax, consistently verify the effectiveness of our method.He tore up the book Book tore he up A caterpillar caught meSynonymous statements to human (Our consideration)Discrete augmentation (CLEAR, etc.) Continuous augmentation (SimCSE, etc.) | A Sentence is Worth 128 Pseudo Tokens: A Semantic-Aware Contrastive Learning Framework for Sentence Embeddings |
d253082255 | Existing controllable dialogue generation work focuses on the single-attribute control and lacks generalization capability to out-of-distribution multiple attribute combinations. In this paper, we explore the compositional generalization for multi-attribute controllable dialogue generation where a model can learn from seen attribute values and generalize to unseen combinations. We propose a prompt-based disentangled controllable dialogue generation model, DCG. It learns attribute concept composition by generating attribute-oriented prompt vectors and uses a disentanglement loss to disentangle different attributes for better generalization. Besides, we design a unified reference-free evaluation framework for multiple attributes with different levels of granularities. Experiment results on two benchmarks prove the effectiveness of our method and the evaluation metric. | Seen to Unseen: Exploring Compositional Generalization of Multi-Attribute Controllable Dialogue Generation |
d199661599 | Traditional models of language processing process language by rule. This approach faces two p . oblems.First, there are difficulties in using the rules during processing, since often one rule must be pitted against another. In this case traditional approaches face the difficult problem of deciding which rule should win in such cases. Second, there are difficulties in acquiring rules, since it is often hard to know when a rule should be proposed, or when a sentence should be handled as one of many special cases. | Connectionist Models of Language |
d6202985 | Most of the recent literature on Sentiment Analysis over Twitter is tied to the idea that the sentiment is a function of an incoming tweet. However, tweets are filtered through streams of posts, so that a wider context, e.g. a topic, is always available. In this work, the contribution of this contextual information is investigated. We modeled the polarity detection problem as a sequential classification task over streams of tweets. A Markovian formulation of the Support Vector Machine discriminative model as embodied by the SVM hmm algorithm has been here employed to assign the sentiment polarity to entire sequences. The experimental evaluation proves that sequential tagging effectively embodies evidence about the contexts and is able to reach a relative increment in detection accuracy of around 20% in F1 measure. These results are particularly interesting as the approach is flexible and does not require manually coded resources.ColMustard : Amazing match yesterday!!#Bayern vs. #Freiburg 4-0 #easyvictory SergGray : @ColMustard Surely, but #Freiburg wasted lot of chances to score.. wrong substitutions by #Guardiola during the 2nd half!! ColMustard : @SergGray Yes, I totally agree with you about the substitutions! #Bayern #Freiburg This work is licenced under a Creative Commons Attribution 4.0 International License. | A context-based model for Sentiment Analysis in Twitter |
d59759934 | The first published LR algorithm for Tree Adjoining Grammars {TAGs [Joshi and Schabes, 1996)) was due to Schabes and Vij ay-Shanker [1990]. Nederhof [1998]showed that it was incorrect {after [Kinyon, 1997)), and proposed a new one. Experimenting with his new algorithm over the XTAG En glish Grammar[XTAG Research Group, 1998) he concluded that LR parsing was inadequate for use with reasonably sized grammars because the size of the generated table was unmanageable. Also the degree of conflicts is too high. In this paper we discuss issues involved with LR parsing for TA Gs and propose a new version of the algorithm that, by maintaining the degree of prediction while deferring the "subtree reduction" , dramatically reduces both the average number of conflicts per state and the size of the parser.IntroductionFor Context Free Grammars, LR parsing [Knuth, 1965, Aho et al., 1986] can be viewed as follows.If at a certain state q 0 of the LR automaton, during the parsing of a sentence, we expect to see the expansion of a certain. non-terminal A, and there is a production A ➔ X 1 X 2 X3 •.. X n in the grammar, then the automaton must have a path labeled X 1 X 2 X 3 ... X n starting at q 0 • This is usually represented by saying that each state in the path contains a "dotted item" for the proq.uction, starting with A ➔ • X 1 X 2 X 3 ... X n at q0 , with the dot moving one symbol ahead at each state in the path. We will refer to the last state of such paths as final. The dot being in front of a symbol Xi represents the fact that we expect to see the expansion of Xi in the string. If Xi is a non-terminal, then, before crossing to the next state, we first have to check that some expansion of Xi is actually next in the input.The situation is depicted inFigure 1, where paths are represented as winding lines and single arcs as straight lines. At a certain state q1 , where some possible yield of the prefix a1 of a production A ➔ a1 Ba2 has just been scanned, the corresponding dotted item is at a non-terminal B. This state turns out to be itself the beginning of other paths, like {3 in the picture, that lead to the recognition of some match for B through some of its rules. The machine, guided by the input string, could traverse this sub-path until getting to q4 • At this final state, some sort of memory is needed to get back to the previous path for A, to then cross from q1 to q2 (i.e. B has just been seen) . In LR parsing this •1 am grateful to Aravind Joshi, Joseph Rosenzweig, Anoop Sarkar, Mahesh Viswanathan, Fei Xia, and the very kind IWPT 2000 reviewers, for their contribution to the development of this paper. | AN EFFICIENT LR PARSER GENERATOR FOR TREE ADJOINING GRAMMARS |
d235624050 | Traditionally, character-level transduction problems have been solved with finite-state models designed to encode structural and linguistic knowledge of the underlying process, whereas recent approaches rely on the power and flexibility of sequence-to-sequence models with attention. Focusing on the less explored unsupervised learning scenario, we compare the two model classes side by side and find that they tend to make different types of errors even when achieving comparable performance. We analyze the distributions of different error classes using two unsupervised tasks as testbeds: converting informally romanized text into the native script of its language (for Russian, Arabic, and Kannada) and translating between a pair of closely related languages (Serbian and Bosnian). Finally, we investigate how combining finite-state and sequence-to-sequence models at decoding time affects the output quantitatively and qualitatively. 1 | Comparative Error Analysis in Neural and Finite-state Models for Unsupervised Character-level Transduction |
d253080785 | To understand what kinds of linguistic knowledge are encoded by pretrained Chinese language models (LMs), we introduce the benchmark of Sino LINGuistics (SLING), which consists of 38K minimal sentence pairs in Mandarin Chinese grouped into 9 high-level linguistic phenomena. Each pair demonstrates the acceptability contrast of a specific syntactic or semantic phenomenon (e.g., The keys are lost vs. The keys is lost), and an LM should assign lower perplexity to the acceptable sentence. In contrast to the CLiMP dataset (Xiang et al., 2021), which also contains Chinese minimal pairs and was created by translating the vocabulary of the English BLiMP dataset, the minimal pairs in SLING are derived primarily by applying syntactic and lexical transformations to naturally-occurring, linguist-annotated sentences from the Chinese Treebank 9.0, thus addressing severe issues in CLiMP's data generation process. We test 18 publicly available pretrained monolingual (e.g., BERT-base-zh, CPM) and multi-lingual (e.g., mT5, XLM) language models on SLING. Our experiments show that the average accuracy for LMs is far below human performance (69.7% vs. 97.1%), while BERT-base-zh achieves the highest accuracy (84.8%) of all tested LMs, even much larger ones. Additionally, we find that most LMs have a strong gender and number (singular/plural) bias, and they perform better on local phenomena than hierarchical ones. 1 | SLING: Sino LINGuistic Evaluation of Large Language Models |
d10362508 | This paper addresses the problem of reliably measuring productivity gains by professional translators working with a machine translation enhanced computer assisted translation tool. In particular, we report on a field test we carried out with a commercial CAT tool in which translation memory matches were supplemented with suggestions from a commercial machine translation engine. The field test was conducted with 12 professional translators working on real translation projects. Productivity of translators were measured with two indicators, post-editing speed and post-editing effort, on two translation directions, English-Italian and English-German, and two linguistic domains, legal and information technology. Besides a detailed statistical analysis of the experimental results, we also discuss issues encountered in running the test. | Measuring User Productivity in Machine Translation Enhanced Computer Assisted Translation |
d248524857 | Dialogue state tracking (DST) aims to predict the current dialogue state given the dialogue history. Existing methods generally exploit the utterances of all dialogue turns to assign value for each slot. This could lead to suboptimal results due to the information introduced from irrelevant utterances in the dialogue history, which may be useless and can even cause confusion. To address this problem, we propose LUNA, a SLot-TUrN Alignment enhanced approach. It first explicitly aligns each slot with its most relevant utterance, then further predicts the corresponding value based on this aligned utterance instead of all dialogue utterances. Furthermore, we design a slot ranking auxiliary task to learn the temporal correlation among slots which could facilitate the alignment. Comprehensive experiments are conducted on multi-domain task-oriented dialogue datasets, i.e., MultiWOZ 2.0, MultiWOZ 2.1, and MultiWOZ 2.2. The results show that LUNA achieves new state-ofthe-art results on these datasets. 1 | LUNA: Learning Slot-Turn Alignment for Dialogue State Tracking |
d233365106 | ||
d5829210 | One of the big challenges in understanding text, i.e., constructing an overall coherent representation of the text, is that much information needed in that representation is unstated (implicit). Thus, in order to "fill in the gaps" and create an overall representation, language processing systems need a large amount of world knowledge, and creating those knowledge resources remains a fundamental challenge. In our current work, we are seeking to augment WordNet as a knowledge resource for language understanding in several ways: adding in formal versions of its word sense definitions (glosses); classifying the morphosemantic links between nouns and verbs; encoding a small number of "core theories" about WordNet's most commonly used terms; and adding in simple representations of scripts. Although this is still work in progress, we describe our experiences so far with what we hope will be a significantly improved resource for the deep understanding of language. | Augmenting WordNet for Deep Understanding of Text |
d27387420 | We present a novel technique for Arabic morphological annotation. The technique utilizes diacritization to produce morphological annotations of quality comparable to human annotators. Although Arabic text is generally written without diacritics, diacritization is already available for large corpora of Arabic text in several genres. Furthermore, diacritization can be generated at a low cost for new text as it does not require specialized training beyond what educated Arabic typists know. The basic approach is to enrich the input to a state-of-the-art Arabic morphological analyzer with word diacritics (full or partial) to enhance its performance. When applied to fully diacritized text, our approach produces annotations with an accuracy of over 97% on lemma, part-of-speech, and tokenization combined. | Exploiting Arabic Diacritization for High Quality Automatic Annotation |
d9150889 | Most state-of-the-art named entity recognition (NER) systems rely on handcrafted features and on the output of other NLP tasks such as part-of-speech (POS) tagging and text chunking. In this work we propose a language-independent NER system that uses automatically learned features only. Our approach is based on the CharWNN deep neural network, which uses word-level and character-level representations (embeddings) to perform sequential classification. We perform an extensive number of experiments using two annotated corpora in two different languages: HAREM I corpus, which contains texts in Portuguese; and the SPA CoNLL-2002 corpus, which contains texts in Spanish. Our experimental results shade light on the contribution of neural character embeddings for NER. Moreover, we demonstrate that the same neural network which has been successfully applied to POS tagging can also achieve state-of-the-art results for language-independet NER, using the same hyperparameters, and without any handcrafted features. For the HAREM I corpus, CharWNN outperforms the state-of-the-art system by 7.9 points in the F1-score for the total scenario (ten NE classes), and by 7.2 points in the F1 for the selective scenario (five NE classes). | Boosting Named Entity Recognition with Neural Character Embeddings |
d248366298 | In this paper, we consider mimicking fictional characters as a promising direction for building engaging conversation models. To this end, we present a new practical task where only a few utterances of each fictional character are available to generate responses mimicking them. Furthermore, we propose a new method named Pseudo Dialog Prompting (PDP) that generates responses by leveraging the power of largescale language models with prompts containing the target character's utterances. To better reflect the style of the character, PDP builds the prompts in the form of dialog that includes the character's utterances as dialog history. Since only utterances of the characters are available in the proposed task, PDP matches each utterance with an appropriate pseudo-context from a predefined set of context candidates using a retrieval model. Through human and automatic evaluation, we show that PDP generates responses that better reflect the style of fictional characters than baseline methods. | Meet Your Favorite Character: Open-domain Chatbot Mimicking Fictional Characters with only a Few Utterances |
d253107899 | The tasks of humor understanding and generation are challenging and subjective even for humans, requiring commonsense and realworld knowledge to master. Puns, in particular, add the challenge of fusing that knowledge with the ability to interpret lexical-semantic ambiguity. In this paper, we present the Ex-PUNations (ExPUN) dataset, in which we augment an existing dataset of puns with detailed crowdsourced annotations of keywords denoting the most distinctive words that make the text funny, pun explanations describing why the text is funny, and fine-grained funniness ratings. This is the first humor dataset with such extensive and fine-grained annotations specifically for puns. Based on these annotations, we propose two tasks: explanation generation to aid with pun classification and keywordconditioned pun generation, to challenge the current state-of-the-art natural language understanding and generation models' ability to understand and generate humor. We showcase that the annotated keywords we collect are helpful for generating better novel humorous texts in human evaluation, and that our natural language explanations can be leveraged to improve both the accuracy and robustness of humor classifiers. | ExPUNations: Augmenting Puns with Keywords and Explanations |
d247849290 | A common way to combat exposure bias is by applying scores from evaluation metrics as rewards in reinforcement learning (RL). Metrics leveraging contextualized embeddings appear more flexible than those that match n-grams and thus ideal as training rewards. Yet metrics such as BERTSCORE greedily align candidate and reference tokens, which can give system outputs excess credit relative to a reference. Past systems using such semantic similarity rewards further suffer from repetitive outputs and overfitting. To address these issues, we propose metrics that replace the greedy alignments in BERTSCORE with optimized ones. Our model optimizing discrete alignment metrics consistently outperforms cross-entropy and BLEU reward baselines on AMR-to-text generation. Additionally, we find that this model enjoys stable training relative to a non-RL setting. | Rewarding Semantic Similarity under Optimized Alignments for AMR-to-Text Generation |
d5524852 | COMPUTATIONAL COMPARATIVE STUDIES ON ROMANCE LAGUAGES A linguistic comparison of lexicon-grammars | |
d226301517 | RESUMELes caractéristiques temporelles et spectrales du schwa transitionnel en tachlhit sont analysées dans cette étude. Nous avons examiné 18 items du type C1C2VC afin d'explorer comment la durée et la qualité de ce vocoïde sont affectées par le contexte consonantique et vocalique avoisinant. Les résultats obtenus à partir de la réalisation de 7 locuteurs natifs montrent que la durée du schwa est beaucoup plus court comparées aux voyelles pleines. Alors que cette durée varie peu selon le contexte, la qualité du schwa peut être affectée par une combinaison de facteurs incluant la nature de la voyelle qui suit, ainsi que le lieu et le mode d'articulation des consonnes adjacentes. Ces variations sont observées pour F1, F2 et F3, et la plupart d'entre elles peuvent être prédites selon que la consonne qui suit est une occlusive emphatique ou une sonante battue.ABSTRACTVariations of transitional schwa in Tashlhiyt: an acoustic analysisThe temporal and spectral characteristics of the Tashlhiyt transitional schwa are analyzed in this study. We examined 18 C1C2VC type items aiming to explore how the duration and quality of this vowel are affected by the consonantal context and surrounding vowels. The results obtained from the realization of 7 native speakers show that the duration of the schwa is much shorter compared to full vowels. While this duration varies little according to context, the quality of the schwa can be affected by a combination of factors including the nature of the vowel that follows, and the place and mode of articulation of adjacent consonants. These variations are observed for F1, F2 and F3, and most of them can be predicted by whether the following consonant is an emphatic stop or a rhotic sonorant. | Les variations du schwa transitionnel en tachlhit : une analyse acoustique |
d4867671 | BackgroundThe University of Michigan's natural language processing system, called LINK, is a unificationbased system which we have developed over the last four years . Prior to MUC-4, LINK ha d been used to extract information from free-form texts in two narrow application domains . One application corpus contained terse descriptions of symptoms displayed by malfunctioning automobiles, and the repairs which fixed them . The other corpus described sequences of activities to be performed on an assembly line . In empirical testing in these two domains, LINK correctly processed 70% of previously unseen descriptions . A template was counted as correct only if all of the fillers in the template were filled correctly . In addition, LINK generated incomplete (but not incorrect) templates for another 15% of the descriptions .These previous domains were much narrower than the MUC-4 terrorism domain . As a comparison, the lexicons for the previous domains contained only 300-500 words, compared wit h 6700 words in our MUC-4 test configuration . Previous grammar size ranged from 75-100 rules , compared with over 500 rules in the MUC-4 knowledge base . In addition, the previous application domains consisted only of single-sentence inputs . Thus, the integration of information from multiple sentences was not an issue in our previous work . | Description of the LINK System Used for MUC-4 |
d237485549 | The majority of language domains require prudent use of terminology to ensure clarity and adequacy of information conveyed. While the correct use of terminology for some languages and domains can be achieved by adapting general-purpose MT systems on large volumes of in-domain parallel data, such quantities of domain-specific data are seldom available for less-resourced languages and niche domains. Furthermore, as exemplified by COVID-19 recently, no domain-specific parallel data is readily available for emerging domains. However, the gravity of this recent calamity created a high demand for reliable translation of critical information regarding pandemic and infection prevention. This work is part of WMT2021 Shared Task: Machine Translation using Terminologies, where we describe Tilde MT systems that are capable of dynamic terminology integration at the time of translation. Our systems achieve up to 94% COVID-19 term use accuracy on the test set of the EN-FR language pair without having access to any form of in-domain information during system training. We conclude our work with a broader discussion considering the Shared Task itself and terminology translation in MT. | Dynamic Terminology Integration for COVID-19 and other Emerging Domains |
d250179945 | Les approches de compréhension automatique de la parole ont récemment bénéficié de l'apport de modèles préappris par autosupervision sur de gros corpus de parole. Pour le français, le projet LeBenchmark a rendu disponibles de tels modèles et a permis des évolutions impressionnantes sur plusieurs tâches dont la compréhension automatique de la parole. Ces avancées ont un coût non négligeable en ce qui concerne le temps de calcul et la consommation énergétique. Dans cet article, nous comparons plusieurs stratégies d'apprentissage visant à réduire le coût énergétique tout en conservant des performances compétitives. Les expériences sont effectuées sur le corpus MEDIA, et montrent qu'il est possible de réduire significativement le coût d'apprentissage tout en conservant des performances à l'état de l'art.ABSTRACTTowards automatic end-to-end speech understanding with less effortRecent advances in spoken language understanding benefited from Self-Supervised models trained on large speech corpora. For French, the LeBenchmark project has made such models available and has led to impressive progress on several tasks including spoken language understanding. These advances have a non-negligible cost in terms of computation time and energy consumption. In this paper, we compare several learning strategies aiming at reducing such cost while keeping competitive performances. The experiments are performed on the MEDIA corpus, and show that it is possible to reduce the learning cost while maintaining state-of-the-art performances.MOTS-CLÉS : compréhension de la parole, apprentissage autosupervisé, apprentissage par transfert. | Vers la compréhension automatique de la parole bout-en-bout à moindre effort |
d252090014 | Machine-learning-as-a-service (MLaaS) has attracted millions of users to their splendid largescale models. Although published as black-box APIs, the valuable models behind these services are still vulnerable to imitation attacks. Recently, a series of works have demonstrated that attackers manage to steal or extract the victim models. Nonetheless, none of the previous stolen models can outperform the original black-box APIs. In this work, we conduct unsupervised domain adaptation and multi-victim ensemble to showing that attackers could potentially surpass victims, which is beyond previous understanding of model extraction. Extensive experiments on both benchmark datasets and real-world APIs validate that the imitators can succeed in outperforming the original blackbox models on transferred domains. We consider our work as a milestone in the research of imitation attack, especially on NLP APIs, as the superior performance could influence the defense or even publishing strategy of API providers. | Student Surpasses Teacher: Imitation Attack for Black-Box NLP APIs |
d21694889 | We propose in this paper to add to the captions of the Flickr30k Entities corpus some syntactic annotations in order to study the joint processing of image and language features for the Preposition-Phrase attachment disambiguation task. The annotation has been performed on the English version of the captions and automatically projected on their French and German translations. | Adding Syntactic Annotations to Flickr30k Entities Corpus for Multimodal Ambiguous Prepositional-Phrase Attachment Resolution |
d493888 | Conversational Agents have been shown to be effective tutors in a wide range of educational domains. However, these agents are often ignored and abused in collaborative learning scenarios involving multiple students. In our work presented here, we design and evaluate interaction strategies motivated from prior research in small group communication. We will discuss how such strategies can be implemented in agents. As a first step towards evaluating agents that can interact socially, we report results showing that human tutors employing these strategies are able to cover more concepts with the students besides being rated as better integrated, likeable and friendlier. | Engaging learning groups using Social Interaction Strategies |
d247158434 | Multilingual pre-trained language models have shown impressive performance on crosslingual tasks. It greatly facilitates the applications of natural language processing on lowresource languages. However, there are still some languages that the current multilingual models do not perform well on. In this paper, we propose CINO (Chinese Minority Pretrained Language Model), a multilingual pretrained language model for Chinese minority languages. It covers Standard Chinese, Yue Chinese, and six other ethnic minority languages. To evaluate the cross-lingual ability of the multilingual model on ethnic minority languages, we collect documents from Wikipedia and news websites, and construct two text classification datasets, WCM (Wiki-Chinese-Minority) and CMNews (Chinese-Minority-News). We show that CINO notably outperforms the baselines on various classification tasks. The CINO model and the datasets are publicly available at http:// cino.hfl-rc.com. | CINO: A Chinese Minority Pre-trained Language Model |
d21689303 | Building a Knowledge Base from text corpora is useful for many applications such as question answering and web search. Since 2012, the Cold Start Knowledge Base Population (KBP) evaluation at the Text Analysis Conference (TAC) has attracted many participants. Despite the popularity, the Cold Start KBP evaluation has several problems including but not limited to the following two: first, each year's assessment dataset is a pooled set of query-answer pairs, primarily generated by participating systems. It is well known to participants that there is pooling bias: a system developed outside of the official evaluation period is not rewarded for finding novel answers, but rather is penalized for doing so. Second, the assessment dataset, constructed with lots of human effort, offers little help in training information extraction algorithms which are crucial ingredients for the end-to-end KBP task. To address these problems, we propose a new unbiased evaluation methodology that uses existing component-level annotation such as the Automatic Content Extraction (ACE) dataset, to evaluate Cold Start KBP. We also propose bootstrap resampling to provide statistical significance to the results reported. We will then present experimental results and analysis. | When ACE met KBP 1 : End-to-End Evaluation of Knowledge Base Population with Component-level Annotation |
d252918635 | Meaning of words constantly change given the events in modern civilization. Large Language Models use word embeddings, which are often static and thus cannot cope with this semantic change. Thus,it is important to resolve ambiguity in word meanings. This paper is an effort in this direction, where we explore methods for word sense disambiguation for the EvoNLP shared task. We conduct rigorous ablations for two solutions to this problem. We see that an approach using time-aware language models helps this task. Furthermore, we explore possible future directions to this problem. | Temporal Word Meaning Disambiguation using TimeLMs |
d63689630 | Dans cet article, nous décrivons la méthode que nous avons développée pour la résolution de métonymie des entités nommées dans le cadre de la compétition SemEval 2007. Afin de résoudre les métonymies sur les noms de lieux et noms d'organisation, tel que requis pour cette tâche, nous avons mis au point un système hybride basé sur l'utilisation d'un analyseur syntaxique robuste combiné avec une méthode d'analyse distributionnelle. Nous décrivons cette méthode ainsi que les résultats obtenus par le système dans le cadre de la compétition SemEval 2007.Abstract. In this paper, we describe the method we develop in order to solve Named entity metonymy in the framework of the SemEval 2007 competition. In order to perform Named Entity metonymy resolution on location names and company names, as required for this task, we developed a hybrid system based on the use of a robust parser that extracts deep syntactic relations combined with a non supervised distributional approach, also relying on the relations extracted by the parser. We describe this methodology as well as the results obtained at SemEval 2007 Mots-clés : Entités Nommées, métonymie, méthode hybride, analyse syntaxique robuste, approche distributionnelle. | Résolution de Métonymie des Entités Nommées : proposition d'une méthode hybride |
d850921 | Learning entailment rules is fundamental in many semantic-inference applications and has been an active field of research in recent years. In this paper we address the problem of learning transitive graphs that describe entailment rules between predicates (termed entailment graphs). We first identify that entailment graphs exhibit a "tree-like" property and are very similar to a novel type of graph termed forest-reducible graph. We utilize this property to develop an iterative efficient approximation algorithm for learning the graph edges, where each iteration takes linear time. We compare our approximation algorithm to a recently-proposed state-of-the-art exact algorithm and show that it is more efficient and scalable both theoretically and empirically, while its output quality is close to that given by the optimal solution of the exact algorithm. | Efficient Tree-based Approximation for Entailment Graph Learning |
d242955 | This paper attempts to bridge the gap between FrameNet frames and inference. We describe a computational formalism that captures structural relationships among participants in a dynamic scenario. This representation is used to describe the internal structure of FrameNet frames in terms of parameters for event simulations. We apply our formalism to the commerce domain and show how it provides a flexible means of accounting for linguistic perspective and other inferential effects. | Putting Frames in Perspective |
d9821946 | This paper describes USAAR's submission to the the metrics shared task of the Workshop on Statistical Machine Translation (WMT) in 2015. The goal of our submission is to take advantage of the semantic overlap between hypothesis and reference translation for predicting MT output adequacy using language independent document embeddings. The approach presented here is learning a Bayesian Ridge Regressor using document skip-gram embeddings in order to automatically evaluate Machine Translation (MT) output by predicting semantic adequacy scores. The evaluation of our submission -measured by the correlation with human judgements -shows promising results on system-level scores. | Predicting Machine Translation Adequacy with Document Embeddings |
d51989481 | We perform a fine-grained large-scale analysis of coreference projection. By projecting gold coreference from Czech to English and vice versa on Prague Czech-English Dependency Treebank 2.0 Coref, we set an upper bound of a proposed projection approach for these two languages. We undertake a detailed thorough analysis that combines the analysis of projection's subtasks with analysis of performance on individual mention types. The findings are accompanied with examples from the corpus. | A Fine-grained Large-scale Analysis of Coreference Projection |
d31080548 | Automatic generation of multiple-choice questions is an emerging topic in application of natural language processing. Particularly, applying it to language testing has been proved to be useful(Sumita et al., 2005).This demo presents an novel approach of question generation using machine learning we have introduced in(Hoshino and Nakagawa, 2005). Our study aims to generate TOEIC-like 1 multiple choice, fillin-the-blank questions from given text using a classifier trained on a set of human-made questions. The system comprises of a question pool, which is a database of questions, an instance converter which does feature extraction, etc. for machine learning and a question generator. Each step of learning and generation is conducted through a web-browser. Figure 1: A system diagram The demo serves for the following three purposes; To facilitates repeating the experiment with different 1 TOEIC: Test of English for International Communication parameters, to demonstrate our method of question generation by showing the result of each steps, and to collect the data (training data and the students' answers) from multiple users in possibly different places.ProcessesAn experiment is performed in a sequence of processes in each of which the system allows the user to change input/parameters and shows the result. The demo follows the processes described in the following.Input QuestionsThe questions in the question pool are listed on the browser. The user can modify those questions or add new ones.Convert to InstancesEach question in the question pool is automatically converted into instances each of which represents a possible blank position.A sentence is [ ] to instances. 1.convert 2. converted 3. converts 4. conversion Above question sentence is converted into the following instances, then, features such as POS 2 , lemma, POS of the previous word, POS of the next word, position-in-sentence, sentence length are assigned to each instance in a totally automatic fashion.We decide a blank position for a question by classifying an instance into true or false. Temporally, 2 Part-of-speech tags are tagged by a modified version of the Tree Tagger by the University of Stuttgart.18 | WebExperimenter for multiple-choice question generation |
d235658105 | Research on overlapped and discontinuous named entity recognition (NER) has received increasing attention. The majority of previous work focuses on either overlapped or discontinuous entities. In this paper, we propose a novel span-based model that can recognize both overlapped and discontinuous entities jointly. The model includes two major steps. First, entity fragments are recognized by traversing over all possible text spans, thus, overlapped entities can be recognized. Second, we perform relation classification to judge whether a given pair of entity fragments to be overlapping or succession. In this way, we can recognize not only discontinuous entities, and meanwhile doubly check the overlapped entities. As a whole, our model can be regarded as a relation extraction paradigm essentially. Experimental results on multiple benchmark datasets (i.e., CLEF, GENIA and ACE05) show that our model is highly competitive for overlapped and discontinuous NER. * Corresponding author. 1 We consider "nested" as a special case of "overlapped". At issue is the liability of a [[Pennsylvania] 1 radio station] 2 under the federal wiretap statute. Example 1 The [mitral] 1 valve [leaflets] 1 are mildly [thickened] 1 . Sequence Labeling Model Our Proposed Model | A Span-Based Model for Joint Overlapped and Discontinuous Named Entity Recognition |
d16758264 | The process engine for pattern recognition and information fusion tasks, the pepr framework, aims to empower the researcher to develop novel solutions in the field of pattern recognition and information fusion tasks in a timely manner, by supporting reuse and combination of well tested and established components in an environment, that eases the wiring of distinct algorithms and description of the control flow through graphical tooling. The framework, not only consisting of the runtime environment, comes with several highly useful components that can be leveraged as a starting point in creating new solutions, as well as a graphical process builder that allows for easy development of pattern recognition processes in a graphical, modeled manner. Additionally, numerous work has been invested in order to keep the entry barrier with regards to extending the framework as low as possible, enabling developers to add additional functionality to the framework in as less time as possible. | An open source process engine framework for realtime pattern recognition and information fusion tasks |
d7039907 | SemEval-2010 Task 11: Event detection in Chinese news sentences | |
d13412990 | Answer Validation is a topic of significant interest within the Question Answering community. In this paper, we propose the use of language modeling methodologies for Answer Validation, using corpus-based methods that do not require the use of external sources. Specifically, we propose a model for Answer Credibility which quantifies the reliability of a source document that contains a candidate answer and the Question's Context Model.157 | Answer Credibility: A Language Modeling Approach to Answer Validation |
d239016988 | In practical applications of semantic parsing, we often want to rapidly change the behavior of the parser, such as enabling it to handle queries in a new domain, or changing its predictions on certain targeted queries. While we can introduce new training examples exhibiting the target behavior, a mechanism for enacting such behavior changes without expensive model re-training would be preferable. To this end, we propose ControllAble Semantic Parser via Exemplar Retrieval (CASPER). Given an input query, the parser retrieves related exemplars from a retrieval index, augments them to the query, and then applies a generative seq2seq model to produce an output parse. The exemplars act as a control mechanism over the generic generative model: by manipulating the retrieval index or how the augmented query is constructed, we can manipulate the behavior of the parser. On the MTOP dataset, in addition to achieving stateof-the-art on the standard setup, we show that CASPER can parse queries in a new domain, adapt the prediction toward the specified patterns, or adapt to new semantic schemas without having to further re-train the model. | Controllable Semantic Parsing via Retrieval Augmentation |
d5239843 | State-of-the-art statistical part-of-speech taggers mainly use information on tag bi-or trigrams, depending on the size of the training corpus. Some also use lexical emission probabilities above unigrams with beneficial results. In both cases, a wider context usually gives better accuracy for a large training corpus, which in turn gives better accuracy than a smaller one. Large corpora with validated tags, however, are scarce, so a bootstrap technique can be used. As the corpus grows, it is probable that a widened context would improve results even further.In this paper, we looked at the contribution to accuracy of such an extended view for both tag transitions and lexical emissions, applied to both a validated Swedish source corpus and a raw bootstrap corpus. We found that the extended view was more important for tag transitions, in particular if applied to the bootstrap corpus. For lexical emission, it was also more important if applied to the bootstrap corpus than to the source corpus, although it was beneficial for both. The overall best tagger had an accuracy of 98.05%. | Extending the View Explorations in Bootstrapping a Swedish PoS Tagger |
d253265138 | Bi-encoder architectures for distantlysupervised relation extraction are designed to make use of the complementary information found in text and knowledge graphs (KG). However, current architectures suffer from two drawbacks. They either do not allow any sharing between the text encoder and the KG encoder at all, or, in case of models with KG-to-text attention, only share information in one direction. Here, we introduce cross-stitch bi-encoders, which allow full interaction between the text encoder and the KG encoder via a cross-stitch mechanism. The cross-stitch mechanism allows sharing and updating representations between the two encoders at any layer, with the amount of sharing being dynamically controlled via cross-attention-based gates. Experimental results on two relation extraction benchmarks from two different domains show that enabling full interaction between the two encoders yields strong improvements. https://github.com/cl-tohoku/xbe 60.2 † 47.9 † JointE 26.3 ⋆ 70.0 ⋆ 61.4 ⋆ 46.4 ⋆ 30.0 ⋆ 38.5 ⋆ 74.0 ⋆ 71.5 ⋆ 69.0 ⋆ 65.4 ⋆ 55.9 ⋆ 43.6 ⋆ RELE 25.6 ⋆ 78.7 ⋆ 66.8 ⋆ 44.7 ⋆ 27.5 ⋆ 40.5 ⋆ 79.0 ⋆ 77.0 ⋆ 77.0 ⋆ 71.2 ⋆ 59.3 ⋆ 44.7 ⋆ BRE+KA 50.3 ⋆ 79.7 ⋆ 79.2 ⋆ 70.3 ⋆ 51.2 ⋆ 48.8 ⋆ 68.0 ⋆ 68.0 ⋆ 67.0 ⋆ 66.0 ⋆ 63.7 ⋆ 52.4 ⋆ | Cross-stitching Text and Knowledge Graph Encoders for Distantly Supervised Relation Extraction |
d218973696 | ||
d2124551 | In this paper we motivate the need for a corpus for the development and testing of summarisation systems for evidencebased medicine. We describe the corpus which we are currently creating, and show its applicability by evaluating several simple query-based summarisation techniques using a small fragment of the corpus. | A Corpus for Evidence Based Medicine Summarisation |
d14100450 | We present our correction annotation guidelines to create a manually corrected nonnative (L2) Arabic corpus. We develop our approach by extending an L1 large-scale Arabic corpus and its manual corrections, to include manually corrected non-native Arabic learner essays. Our overarching goal is to use the annotated corpus to develop components for automatic detection and correction of language errors that can be used to help Standard Arabic learners (native and non-native) improve the quality of the Arabic text they produce. The created corpus of L2 text manual corrections is the largest to date. We evaluate our guidelines using inter-annotator agreement and show a high degree of consistency. | Correction Annotation for Non-Native Arabic Texts: Guidelines and Corpus |
d253116838 | A wide range of control perspectives have been explored in controllable text generation. Structure-controlled summarization is recently proposed as a useful and interesting research direction. However, current structure-controlling methods have limited effectiveness in enforcing the desired structure. To address this limitation, we propose a sentence-level beam search generation method (SentBS), where evaluation is conducted throughout the generation process to select suitable sentences for subsequent generations. We experiment with different combinations of decoding methods to be used as subcomponents by SentBS and evaluate results on the structure-controlled dataset MReD. Experiments show that all explored combinations for SentBS can improve the agreement between the generated text and the desired structure, with the best method significantly reducing the structural discrepancies suffered by the existing model, by approximately 68%. 1 | SentBS: Sentence-level Beam Search for Controllable Summarization |
d235266157 | Open-domain dialog systems have a usercentric goal: to provide humans with an engaging conversation experience. User engagement is one of the most important metrics for evaluating open-domain dialog systems, and could also be used as real-time feedback to benefit dialog policy learning. Existing work on detecting user disengagement typically requires hand-labeling many dialog samples. We propose HERALD, an efficient annotation framework that reframes the training data annotation process as a denoising problem. Specifically, instead of manually labeling training samples, we first use a set of labeling heuristics to label training samples automatically. We then denoise the weakly labeled data using the Shapley algorithm. Finally, we use the denoised data to train a user engagement detector. Our experiments show that HERALD improves annotation efficiency significantly and achieves 86% user disengagement detection accuracy in two dialog corpora. Our implementation is available at https:// github.com/Weixin-Liang/HERALD/. | HERALD: An Annotation Efficient Method to Detect User Disengagement in Social Conversations |
d257622986 | While modern masked language models (LMs) are trained on ever larger corpora, we here explore the effects of down-scaling training to a modestly-sized but representative, wellbalanced, and publicly available English text source -the British National Corpus. We show that pre-training on this carefully curated corpus can reach better performance than the original BERT model. We argue that this type of corpora has great potential as a language modeling benchmark. To showcase this potential, we present fair, reproducible and data-efficient comparative studies of LMs, in which we evaluate several training objectives and model architectures and replicate previous empirical results in a systematic way. We propose an optimized LM architecture called LTG-BERT. | Trained on 100 million words and still in shape: BERT meets British National Corpus |
d235253817 | Aspect sentiment classification (ASC) aims at determining sentiments expressed towards different aspects in a sentence. While state-ofthe-art ASC models have achieved remarkable performance, they are recently shown to suffer from the issue of robustness. Particularly in two common scenarios: when domains of test and training data are different (out-of-domain scenario) or test data is adversarially perturbed (adversarial scenario), ASC models may attend to irrelevant words and neglect opinion expressions that truly describe diverse aspects. To tackle the challenge, in this paper, we hypothesize that position bias (i.e., the words closer to a concerning aspect would carry a higher degree of importance) is crucial for building more robust ASC models by reducing the probability of mis-attending. Accordingly, we propose two mechanisms for capturing position bias, namely position-biased weight and position-biased dropout, which can be flexibly injected into existing models to enhance representations for classification. Experiments conducted on out-of-domain and adversarial datasets demonstrate that our proposed approaches largely improve the robustness and effectiveness of current models. 1 | Exploiting Position Bias for Robust Aspect Sentiment Classification |
d7592530 | The Tree Adjoining Grammar formalism, both its single-as well as multiple-component versions, has recently received attention as a basis for the description and explication of natural language. We show in this paper that the number-name system of Chinese is generated neither by this formalism nor by any other equivalent or weaker ones, suggesting that such a task might require the use of the more powerful Indexed Grammar formalism. Given that our formal results apply only to a proper subset of Chinese, we extensively discuss the issue of whether they have any implications for the whole of that natural language. We conclude that our results bear directly either on the syntax of Chinese or on the interface between Chinese and the cognitive component responsible for arithmetic reasoning. Consequently, either Tree Adjoining Grammars, as currently defined, fail to generate the class of natural languages in a way that discriminates between linguistically warranted sublanguages, or formalisms with generative power equivalent to Tree Adjoining Grammar cannot serve as a basis for the interface between the human linguistic and mathematical faculties. | Chinese Number-Names, Tree Adjoining Languages, and Mild Context-Sensitivity |
d15241522 | In this paper we describe our efforts on POS annotation of a code-switching corpus created from Turkish-German tweets. We use Universal Dependencies (UD) POS tags as our tag set. While the German parts of the corpus employ UD specifications, for the Turkish parts we propose annotation guidelines that adopt UD's language-general rules when it is applicable and adapt its principles to Turkishspecific phenomena when it is not. The resulting corpus has POS annotation of 1029 tweets, which is aligned with existing language identification annotation. | Part of Speech Annotation of a Turkish-German Code-Switching Corpus |
d22109805 | We present a fast query-based multi-document summarizer called FastSum based solely on word-frequency features of clusters, documents and topics. Summary sentences are ranked by a regression SVM. The summarizer does not use any expensive NLP techniques such as parsing, tagging of names or even part of speech information. Still, the achieved accuracy is comparable to the best systems presented in recent academic competitions (i.e., Document Understanding Conference (DUC)). Because of a detailed feature analysis using Least Angle Regression (LARS), FastSum can rely on a minimal set of features leading to fast processing times: 1250 news documents in 60 seconds. | FastSum: Fast and accurate query-based multi-document summarization |
d11159467 | This paper presents a workbench for Tree Adjoining Grammars that we are currently developing. This workbench includes several tools and resources based on the markup language XML, used as a convenient language to format and exchange linguistic resources. | Tools and resources for Tree Adjoining Grammars |
d247446927 | Understanding causal narratives communicated in clinical notes can help make strides towards personalized healthcare. Extracted causal information from clinical notes can be combined with structured EHR data such as patients' demographics, diagnoses, and medications. This will enhance healthcare providers' ability to identify aspects of a patient's story communicated in the clinical notes and help make more informed decisions. * Corresponding Author † Equal contribution ‡ Contributed during an internship at Accenture labs, SF 1 MIMICause dataset will be available under the "Community Annotations Downloads" at https://portal.dbmi.hms.harvard.edu/projects/n2c2-nlp/ | MIMICause: Representation and automatic extraction of causal relation types from clinical notes |
d35948347 | This study primarily aims to build a Turkish psycholinguistic database including three variables: word frequency, age of acquisition (AoA), and imageability, where AoA and imageability information are limited to nouns. We used a corpus-based approach to obtain information about the AoA variable. We built two corpora: a child literature corpus (CLC) including 535 books written for 3-12 years old children, and a corpus of transcribed children's speech (CSC) at ages 1;4-4;8. A comparison between the word frequencies of CLC and CSC gave positive correlation results, suggesting the usability of the CLC to extract AoA information. We assumed that frequent words of the CLC would correspond to early acquired words whereas frequent words of a corpus of adult language would correspond to late acquired words. To validate AoA results from our corpus-based approach, a rated AoA questionnaire was conducted on adults. Imageability values were collected via a different questionnaire conducted on adults. We conclude that it is possible to deduce AoA information for high frequency words with the corpus-based approach. The results about low frequency words were inconclusive, which is attributed to the fact that corpus-based AoA information is affected by the strong negative correlation between corpus frequency and rated AoA. | A Turkish Database for Psycholinguistic Studies Based on Frequency, Age of Acquisition, and Imageability |
d248218772 | Word ordering is a constrained language generation task taking unordered words as input.Existing work uses linear models and neural networks for the task, yet pre-trained language models have not been studied in word ordering, let alone why they help.We use BART as an instance and show its effectiveness in the task.To explain why BART helps word ordering, we extend analysis with probing and empirically identify that syntactic dependency knowledge in BART is a reliable explanation.We also report performance gains with BART in the related partial tree linearization task, which readily extends our analysis. 1 | On the Role of Pre-trained Language Models in Word Ordering: A Case Study with BART |
d253097695 | Our goal is a teachable reasoning system for question-answering (QA), where a user can interact with faithful answer explanations, and correct its errors so that the system improves over time. Our approach is to augment a QA model with a dynamic memory of user feedback, containing user-supplied corrections to erroneous model beliefs that users identify during interaction. Retrievals from memory are used as additional context for QA, to help avoid previous mistakes in similar new situations -a novel application of memory-based continuous learning. With simulated feedback, we find that our system (called TeachMe 1 ) continually improves with time, and without model retraining, requiring feedback on only 25% of training examples to reach within 1% of the upper-bound (feedback on all examples). Similarly, in experiments with real users, we observe a similar trend, with performance improving by over 15% on a hidden test set after teaching. This suggests new opportunities for using frozen language models in an interactive setting where users can inspect, debug, and correct the model's beliefs, leading to improved system's performance over time. | Towards Teachable Reasoning Systems: Using a Dynamic Memory of User Feedback for Continual System Improvement |
d8526022 | Both greedy and domain-oriented REG algorithms have significant strengths but tend to perform poorly according to humanlikeness criteria as measured by, e.g., Dice scores. In this work we describe an attempt to combine both perspectives into a single attribute selection strategy to be used as part of the Dale & Reiter Incremental algorithm in the REG Challenge 2008, and the results in both Furniture and People domains. | USP-EACH Frequency-based Greedy Attribute Selection for Referring Expressions Generation |
d218974309 | ||
d248227500 | Conventionally, generation of natural language for dialogue agents may be viewed as a statistical learning problem: determine the patterns in human-provided data and generate appropriate responses with similar statistical properties. However, dialogue can also be regarded as a goal directed process, where speakers attempt to accomplish a specific task. Reinforcement learning (RL) algorithms are designed specifically for solving such goal-directed problems, but the most direct way to apply RLthrough trial-and-error learning in human conversations, -is costly. In this paper, we study how offline reinforcement learning can instead be used to train dialogue agents entirely using static datasets collected from human speakers. Our experiments show that recently developed offline RL methods can be combined with language models to yield realistic dialogue agents that better accomplish task goals. | CHAI: A CHatbot AI for Task-Oriented Dialogue with Offline Reinforcement Learning |
d258762869 | Dense retrieval has shown promise in the firststage retrieval process when trained on indomain labeled datasets. However, previous studies have found that dense retrieval is hard to generalize to unseen domains due to its weak modeling of domain-invariant and interpretable feature (i.e., matching signal between two texts, which is the essence of information retrieval). In this paper, we propose a novel method to improve the generalization of dense retrieval via capturing matching signal called BERM. Fully fine-grained expression and query-oriented saliency are two properties of the matching signal. Thus, in BERM, a single passage is segmented into multiple units and two unit-level requirements are proposed for representation as the constraint in training to obtain the effective matching signal. One is semantic unit balance and the other is essential matching unit extractability. Unit-level view and balanced semantics make representation express the text in a fine-grained manner. Essential matching unit extractability makes passage representation sensitive to the given query to extract the pure matching information from the passage containing complex context. Experiments on BEIR show that our method can be effectively combined with different dense retrieval training methods (vanilla, hard negatives mining and knowledge distillation) to improve its generalization ability without any additional inference overhead and target domain data. | BERM: Training the Balanced and Extractable Representation for Matching to Improve Generalization Ability of Dense Retrieval |
d257255577 | Identifying the difference between two versions of the same article is useful to update knowledge bases and to understand how articles evolve. Paired texts occur naturally in diverse situations: reporters write similar news stories and maintainers of authoritative websites must keep their information up to date. We propose representing factual changes between paired documents as question-answer pairs, where the answer to the same question differs between two versions. We find that question-answer pairs can flexibly and concisely capture the updated contents. Provided with paired documents, annotators identify questions that are answered by one passage but answered differently or cannot be answered by the other. We release DIFFQG which consists of 759 QA pairs and 1153 examples of paired passages with no factual change. These questions are intended to be both unambiguous and information-seeking and involve complex edits, pushing beyond the capabilities of current question generation and factual change detection systems. Our dataset summarizes the changes between two versions of the document as questions and answers, studying automatic update summarization in a novel way. | DIFFQG: Generating Questions to Summarize Factual Changes |
d226283562 | ||
d15795846 | We present a three-part bilingual specialized dictionary Mexican Sign Language-Spanish / Spanish-Mexican Sign Language. This dictionary will be the outcome of a three-years agreement between the Italian "Consiglio Nazionale delle Ricerche" and the Mexican Conacyt. Although many other sign language dictionaries have been provided to deaf communities, there are no Mexican Sign Language dictionaries in Mexico, yet. We want to stress on the specialized feature of the proposed dictionary: the bilingual dictionary will contain frequently used general Spanish forms along with scholastic course specific specialized words whose meanings warrant comprehension of school curricula. We emphasize that this aspect of the bilingual dictionary can have a deep social impact, since we will furnish to deaf people the possibility to get competence in official language, which is necessary to ensure access to school curriculum and to become full-fledged citizens. From a technical point of view, the dictionary consists of a relational database, where we have saved the sign parameters and a graphical user interface especially designed to allow deaf children to retrieve signs using the relevant parameters and,thus, the meaning of the sign in Spanish. | A bilingual dictionary Mexican Sign Language-Spanish/Spanish-Mexican Sign Language |
d260495679 | We present a large scale collection of diverse natural language inference (NLI) datasets that help provide insight into how well a sentence representation captures distinct types of reasoning. The collection results from recasting 13 existing datasets from 7 semantic phenomena into a common NLI structure, resulting in over half a million labeled context-hypothesis pairs in total. We refer to our collection as the DNC: Diverse Natural Language Inference Collection. The DNC is available online at http://www.decomp.net, and will grow over time as additional resources are recast and added from novel sources. | Collecting Diverse Natural Language Inference Problems for Sentence Representation Evaluation |
d2535099 | Narrated Animation: A Case for Generation | |
d233365248 | ||
d252090194 | Event argument extraction (EAE) has been well studied at the sentence level but under- | Few-Shot Document-Level Event Argument Extraction |
d236145057 | ||
d245424821 | Emotion recognition in conversation (ERC) has attracted much attention in recent years for its necessity in widespread applications. Existing ERC methods mostly model the self and inter-speaker context separately, posing a major issue for lacking enough interaction between them. In this paper, we propose a novel Speaker and Position-Aware Graph neural network model for ERC (S+PAGE), which contains three stages to combine the benefits of both Transformer and relational graph convolution network (R-GCN) for better contextual modeling. Firstly, a two-stream conversational Transformer is presented to extract the coarse self and inter-speaker contextual features for each utterance. Then, a speaker and position-aware conversation graph is constructed, and we propose an enhanced R-GCN model, called PAG, to refine the coarse features guided by a relative positional encoding. Finally, both of the features from the former two stages are input into a conditional random field layer to model the emotion transfer. Extensive experiments demonstrate that our model achieves state-of-the-art performance on three ERC datasets. | S+PAGE: A Speaker and Position-Aware Graph Neural Network Model for Emotion Recognition in Conversation |
d4386013 | In the paper we presented a new Russian wordnet, RuWordNet, which was semiautomatically obtained by transformation of the existing Russian thesaurus RuThes. At the first step, the basic structure of wordnets was reproduced: synsets' hierarchy for each part of speech and the basic set of relations between synsets (hyponym-hypernym, partwhole, antonyms). At the second stage, we added causation, entailment and domain relations between synsets. Also derivation relations were established for single words and the component structure for phrases included in RuWordNet. The described procedure of transformation highlights the specific features of each type of thesaurus representations. | Comparing Two Thesaurus Representations for Russian |
d39170387 | Coling 2008: Educational Natural Language Processing -Tutorial notes Educational Natural Language Processing Tutorial at COLING'08 Educational Natural Language Processing Iryna Gurevych Presenters | |
d259991841 | Domain adaptation is an important and widely studied problem in natural language processing. A large body of literature tries to solve this problem by adapting models trained on the source domain to the target domain. In this paper, we instead solve this problem from a dataset perspective. We modify the source domain dataset with simple lexical transformations to reduce the domain shift between the source dataset distribution and the target dataset distribution. We find that models trained on the transformed source domain dataset performs significantly better than zeroshot models. Using our proposed transformations to convert standard English to tweets, we reach an unsupervised part-of-speech (POS) tagging accuracy of 92.14% (from 81.54% zero shot accuracy), which is only slightly below the supervised performance of 94.45%. We also use our proposed transformations to synthetically generate tweets and augment the Twitter dataset to achieve state-of-the-art performance for POS tagging. | Unsupervised Domain Adaptation using Lexical Transformations and Label Injection for Twitter Data |
d796558 | In this paper, we show that Reileralion and Collocation relations as introduced I)y Ilalliday and llasan may function as lexieally I)iased discourse structure relations and that these relations are well represented by sequences of Mel'&flds Lexical Funclions (Ll,'s). We propose to use Lie sequences for tl,e final determination and realization of discourse organization during lexical choice in text generation. | ON LEXICALLY BIASED DISCOURSE ORGANIZATION IN TEXT GENERATION |
d235097222 | ||
d16503100 | Named Entity Extraction is a mature task in the NLP field that has yielded numerous services gaining popularity in the Semantic Web community for extracting knowledge from web documents. These services are generally organized as pipelines, using dedicated APIs and different taxonomy for extracting, classifying and disambiguating named entities. Integrating one of these services in a particular application requires to implement an appropriate driver. Furthermore, the results of these services are not comparable due to different formats. This prevents the comparison of the performance of these services as well as their possible combination. We address this problem by proposing NERD, a framework which unifies 10 popular named entity extractors available on the web, and the NERD ontology which provides a rich set of axioms aligning the taxonomies of these tools. | NERD: A Framework for Unifying Named Entity Recognition and Disambiguation Extraction Tools |
d227230618 | ||
d219302365 | ||
d27889503 | This paper discusses an innovative approach to the computer assisted scoring of student responses in WebLAS (web-based language assessment system)-a language assessment system delivered entirely over the web. Expected student responses are limited production free response questions.The portions of WebLAS with which we are concerned are the task creation and scoring modules. Within the task creation module, instructors and language experts do not only provide the task input and prompt. More importantly, they interactively inform the system how and how much to score student responses. This interaction consists of WebLAS' natural language processing (NLP) modules searching for alternatives of the provided "gold standard" (Hirschman et al, 2000) answer and asking for confirmation of score assignment. WebLAS processes and stores all this information within its database, to be used in the task delivery and scoring phases. | A reliable approach to automatic assessment of short answer free responses |
d220047065 | Visual question answering aims to answer the natural language question about a given image. Existing graph-based methods only focus on the relations between objects in an image and neglect the importance of the syntactic dependency relations between words in a question. To simultaneously capture the relations between objects in an image and the syntactic dependency relations between words in a question, we propose a novel dual channel graph convolutional network (DC-GCN) for better combining visual and textual advantages. The DC-GCN model consists of three parts: an I-GCN module to capture the relations between objects in an image, a Q-GCN module to capture the syntactic dependency relations between words in a question, and an attention alignment module to align image representations and question representations. Experimental results show that our model achieves comparable performance with the state-of-theart approaches. | Aligned Dual Channel Graph Convolutional Network for Visual Question Answering |
d15743676 | This paper presents a novel approach to automatic captioning of toponym-referenced images. The automatic captioning procedure works by summarizing multiple web-documents that contain information related to an image's location. Our summarizer can generate both query-based and language model-biased multidocument summaries. The models are created from large numbers of existing articles pertaining to places of the same "object type". Evaluation relative to human written captions shows that when language models are used to bias the summarizer the summaries score more highly than the non-biased ones. | Summary Generation for Toponym-Referenced Images using Object Type Language Models |
d1708669 | Word alignment models form an important part of building statistical machine translation systems. Semi-supervised word alignment aims to improve the accuracy of automatic word alignment by incorporating full or partial alignments acquired from humans. Such dedicated elicitation effort is often expensive and depends on availability of bilingual speakers for the language-pair. In this paper we study active learning query strategies to carefully identify highly uncertain or most informative alignment links that are proposed under an unsupervised word alignment model. Manual correction of such informative links can then be applied to create a labeled dataset used by a semi-supervised word alignment model. Our experiments show that using active learning leads to maximal reduction of alignment error rates with reduced human effort. | Active Semi-Supervised Learning for Improving Word Alignment |
d14754940 | Computational models of infant word segmentation have not been tested on a wide range of languages. This paper applies a phonotactic segmentation model to Korean. In contrast to the undersegmentation pattern previously found in English and Russian, the model exhibited more oversegmentation errors and more errors overall. Despite the high error rate, analysis suggested that lexical acquisition might not be problematic, provided that infants attend only to frequently segmented items. | Does Korean defeat phonotactic word segmentation? |
d227230548 | This paper analyses the challenge of working with dialectal variation when semi-automatically normalising and analysing historical Basque texts. This work is part of a more general ongoing project for the construction of a morphosyntactically annotated historical corpus of Basque called Basque in the Making (BIM): A Historical Look at a European Language Isolate, whose main objective is the systematic and diachronic study of a number of grammatical features. This will be not only the first tagged corpus of historical Basque, but also a means to improve language processing tools by analysing historical Basque varieties more or less distant from present-day standard Basque.This work is licensed under a Creative Commons Attribution 4.0 International Licence. Licence details: | Dealing with dialectal variation in the construction of the Basque historical corpus |
d252819090 | Generating temporally-ordered event sequences in texts is important to natural language processing. Two emerging tasks in this direction are temporal event ordering (rearranging the set of events to correct order) and event infilling (generating an event at a specified position). To tackle the two related tasks, the existing method adopts a vanilla sequence-to-sequence model via maximum likelihood estimation (MLE). However, applying this approach to these tasks will cause two issues. One issue is that the MLE loss emphasizes strict local alignment and ignores the global semantics of the event. The other issue is that the model adopts a word-level objective to model events in texts, failing to evaluate the predicted results of the model from the perspective of event sequence. To alleviate these issues, we present a novel model to tackle the generation of temporally-ordered event sequences via Event Optimal Transport (EOT). First, we treat the events in the sequence as modeling units and explicitly extract the semantics of the events. Second, to provide event sequence-level evaluation of the predicted results of the model, we directly match events in sequences. Extensive experimental results show that our approach outperforms previous models on all evaluation datasets. In particular, the accuracy is improved by 7.7%, and the Macro F1 is improved by 7.2% on one of the datasets. | Generating Temporally-ordered Event Sequences via Event Optimal Transport |
d258378255 | With the rise in larger language models, researchers started exploiting them by pivoting the downstream tasks as language modeling tasks using prompts. In this work, we convert the Named Entity Recognition task into a seq2seq task by generating the synthetic sentences using templates. Our main contribution is the conversion framework which provides faster inference. In addition, we test our method's performance in resource-rich, low resource and domain transfer settings. Results show that our method achieves comparable results in the resource-rich setting and outperforms the current seq2seq paradigm state-ofthe-art approach in few-shot settings. Through the experiments, we observed that the negative examples play an important role in model's performance. We applied our approach over BART and T5-base models, and we notice that the T5 architecture aligns better with our task. The work is performed on the datasets in English language. | Improving and Simplifying Template-Based Named Entity Recognition |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.