_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d237393105 | Automated Essay Assessment (AEA) aims to judge students' writing proficiency in an automatic way. This paper presents a Chinese AEA system IFlyEssayAssess (IFlyEA), targeting on evaluating essays written by native Chinese students from primary and junior schools. IFlyEA provides multi-level and multi-dimension analytical modules for essay assessment. It has state-of-the-art grammar level analysis techniques, and also integrates components for rhetoric and discourse level analysis, which are important for evaluating native speakers' writing ability, but still challenging and less studied in previous work. Based on the comprehensive analysis, IFlyEA provides application services for essay scoring, review generation, recommendation, and explainable analytical visualization. These services can benefit both teachers and students during the process of writing teaching and learning. | IFlyEA: A Chinese Essay Assessment System with Automated Rating, Review Generation, and Recommendation |
d234681873 | Recent studies on neural networks with pretrained weights (i.e., BERT) have mainly focused on a low-dimensional subspace, where the embedding vectors computed from input words (or their contexts) are located. In this work, we propose a new approach, called OoMMix, to finding and regularizing the remainder of the space, referred to as out-ofmanifold, which cannot be accessed through the words. Specifically, we synthesize the outof-manifold embeddings based on two embeddings obtained from actually-observed words, to utilize them for fine-tuning the network. A discriminator is trained to detect whether an input embedding is located inside the manifold or not, and simultaneously, a generator is optimized to produce new embeddings that can be easily identified as out-of-manifold by the discriminator. These two modules successfully collaborate in a unified and end-to-end manner for regularizing the out-of-manifold. Our extensive evaluation on various text classification benchmarks demonstrates the effectiveness of our approach, as well as its good compatibility with existing data augmentation techniques which aim to enhance the manifold. | OoMMix: Out-of-manifold Regularization in Contextual Embedding Space for Text Classification |
d222090799 | While extensive popularity of online social media platforms has made information dissemination faster, it has also resulted in widespread online abuse of different types like hate speech, offensive language, sexist and racist opinions, etc. Detection and curtailment of such abusive content is critical for avoiding its psychological impact on victim communities, and thereby preventing hate crimes. Previous works have focused on classifying user posts into various forms of abusive behavior. But there has hardly been any focus on estimating the severity of abuse and the target. In this paper, we present a first of the kind dataset with 7,601 posts from Gab 1 which looks at online abuse from the perspective of presence of abuse, severity and target of abusive behavior. We also propose a system to address these tasks, obtaining an accuracy of ∼80% for abuse presence, ∼82% for abuse target prediction, and ∼65% for abuse severity prediction. * | AbuseAnalyzer: Abuse Detection, Severity and Target Prediction for Gab Posts |
d7621282 | We present approaches for the identification of sentences understandable by second language learners of Swedish, which can be used in automatically generated exercises based on corpora. In this work we merged methods and knowledge from machine learning-based readability research, from rule-based studies of Good Dictionary Examples and from second language learning syllabuses. The proposed selection methods have also been implemented as a module in a free web-based language learning platform. Users can use different parameters and linguistic filters to personalize their sentence search with or without a machine learning component assessing readability. The sentences selected have already found practical use as multiple-choice exercise items within the same platform. Out of a number of deep linguistic indicators explored, we found mainly lexical-morphological and semantic features informative for second language sentence-level readability. We obtained a readability classification accuracy result of 71%, which approaches the performance of other models used in similar tasks. Furthermore, during an empirical evaluation with teachers and students, about seven out of ten sentences selected were considered understandable, the rulebased approach slightly outperforming the method incorporating the machine learning model. | Rule-based and machine learning approaches for second language sentence-level readability |
d256461165 | The bloom of the Internet and the recent breakthroughs in deep learning techniques open a new door to AI for E-commence, with a trend evolved from using a few financial factors such as liquidity and profitability to using more advanced AI techniques to process complex and multi-modal data. In this paper, we tackle the practical problem of restaurant survival prediction. We argue that traditional methods ignore two essential aspects, which are very helpful for the task: 1) modeling customer reviews and 2) jointly considering status prediction and result explanation. Thus, we propose a novel joint learning framework for explainable restaurant survival prediction based on the multi-modal data of user-restaurant interactions and users' textual reviews. Moreover, we design a graph neural network to capture the high-order interactions and design a co-attention mechanism to capture the most informative and meaningful signal from noisy textual reviews. Our results on two datasets show a significant and consistent improvement over the SOTA techniques (average 6.8% improvement in prediction and 45.3% improvement in explanation). | A Joint Learning Framework for Restaurant Survival Prediction and Explanation |
d233289631 | We propose a multi-task, probabilistic approach to facilitate distantly supervised relation extraction by bringing closer the representations of sentences that contain the same Knowledge Base pairs. To achieve this, we bias the latent space of sentences via a Variational Autoencoder (VAE) that is trained jointly with a relation classifier. The latent code guides the pair representations and influences sentence reconstruction. Experimental results on two datasets created via distant supervision indicate that multi-task learning results in performance benefits. Additional exploration of employing Knowledge Base priors into the VAE reveals that the sentence space can be shifted towards that of the Knowledge Base, offering interpretability and further improving results 1 . | Distantly Supervised Relation Extraction with Sentence Reconstruction and Knowledge Base Priors |
d233296012 | The existence of multiple datasets for sarcasm detection prompts us to apply transfer learning to exploit their commonality. The adversarial neural transfer (ANT) framework utilizes multiple loss terms that encourage the source-domain and the target-domain feature distributions to be similar while optimizing for domain-specific performance. However, these objectives may be in conflict, which can lead to optimization difficulties and sometimes diminished transfer. We propose a generalized latent optimization strategy that allows different losses to accommodate each other and improves training dynamics. The proposed method outperforms transfer learning and meta-learning baselines. In particular, we achieve 10.02% absolute performance gain over the previous state of the art on the iSarcasm dataset. | Latent-Optimized Adversarial Neural Transfer for Sarcasm Detection |
d51874610 | We propose a novel approach to OCR post-correction that exploits repeated texts in large corpora both as a source of noisy target outputs for unsupervised training and as a source of evidence when decoding. A sequence-to-sequence model with attention is applied for single-input correction, and a new decoder with multi-input attention averaging is developed to search for consensus among multiple sequences. We design two ways of training the correction model without human annotation, either training to match noisily observed textual variants or bootstrapping from a uniform error model. On two corpora of historical newspapers and books, we show that these unsupervised techniques cut the character and word error rates nearly in half on single inputs and, with the addition of multi-input decoding, can rival supervised methods. | Multi-Input Attention for Unsupervised OCR Correction |
d215745354 | Despite the continuing efforts to improve the engagingness and consistency of chit-chat dialogue systems, the majority of current work simply focus on mimicking human-like responses, leaving understudied the aspects of modeling understanding between interlocutors. The research in cognitive science, instead, suggests that understanding is an essential signal for a high-quality chit-chat conversation. Motivated by this, we propose P 2 BOT, a transmitter-receiver based framework with the aim of explicitly modeling understanding. Specifically, P 2 BOT incorporates mutual persona perception to enhance the quality of personalized dialogue generation. Experiments on a large public dataset, PERSONA-CHAT, demonstrate the effectiveness of our approach, with a considerable boost over the state-of-theart baselines across both automatic metrics and human evaluations. | You Impress Me: Dialogue Generation via Mutual Persona Perception |
d202542757 | Contextual word representations, typically trained on unstructured, unlabeled text, do not contain any explicit grounding to real world entities and are often unable to remember facts about those entities. We propose a general method to embed multiple knowledge bases (KBs) into large scale models, and thereby enhance their representations with structured, human-curated knowledge. For each KB, we first use an integrated entity linker to retrieve relevant entity embeddings, then update contextual word representations via a form of word-to-entity attention. In contrast to previous approaches, the entity linkers and selfsupervised language modeling objective are jointly trained end-to-end in a multitask setting that combines a small amount of entity linking supervision with a large amount of raw text. After integrating WordNet and a subset of Wikipedia into BERT, the knowledge enhanced BERT (KnowBert) demonstrates improved perplexity, ability to recall facts as measured in a probing task and downstream performance on relationship extraction, entity typing, and word sense disambiguation. KnowBert's runtime is comparable to BERT's and it scales to large KBs. | Knowledge Enhanced Contextual Word Representations |
d220047754 | Abstractive summarisation is notoriously hard to evaluate since standard word-overlap-based metrics are biased towards specific words in the human reference. We introduce a new evaluation metric which abstracts away from the word-level and instead is based on factlevel content weighting, i.e. relating the facts of the document to the facts of the summary. We follow the assumption that a good summary will reflect all relevant facts, i.e. the ones present in the ground truth (human-generated reference summary). We confirm this hypothesis by showing that our weightings are highly correlated to human perception and compare favourably to the recent manual highlightbased metric ofHardy et al. (2019). | Fact-based Content Weighting for Evaluating Abstractive Summarisation |
d207853300 | In this paper, we present Hierarchical Graph Network (HGN) for multi-hop question answering. To aggregate clues from scattered texts across multiple paragraphs, a hierarchical graph is created by constructing nodes from different levels of granularity (i.e., questions, paragraphs, sentences, and entities), the representations of which are initialized with BERTbased context encoders. By weaving heterogeneous nodes in an integral unified graph, this characteristic hierarchical differentiation of node granularity enables HGN to support different question answering sub-tasks simultaneously (e.g., paragraph selection, supporting facts extraction, and answer prediction). Given a constructed hierarchical graph for each question, the initial node representations are updated through graph propagation; and for each sub-task, multi-hop reasoning is performed by traversing through graph edges. Extensive experiments on the HotpotQA benchmark demonstrate that the proposed HGN approach significantly outperforms prior state-ofthe-art methods by a large margin in both Distractor and Fullwiki settings. | Hierarchical Graph Network for Multi-hop Question Answering |
d3019829 | In human translation, translators first make draft translations and then modify and edit them. In the case of experienced translators, this process involves the use of wide-ranging expert knowledge, which has mostly remained implicit so far. Describing the difference between draft and final translations, therefore, should contribute to making this knowledge explicit. If we could clarify the expert knowledge of translators, hopefully in a computationally tractable way, we would be able to contribute to the automatic notification of awkward translations to assist inexperienced translators, improving the quality of MT output, etc. Against this backdrop, we have started constructing a corpus that indicates patterns of modification between draft and final translations made by human translators. This paper reports on our progress to date. | Constructing a corpus that indicates patterns of modification between draft and final translations by human translators |
d256461367 | The popularity of multimodal dialogue has stimulated the need for a new generation of dialogue agents with multimodal interactivity. When users communicate with customer service, they may express their requirements by means of text, images, or even videos. Visual information usually acts as discriminators for product models, or indicators of product failures, which play an important role in the Ecommerce scenario. On the other hand, detailed information provided by the images is limited, and typically, customer service systems cannot understand the intent of users without the input text. Thus, bridging the gap between the image and text is crucial for communicating with customers. In this paper, we construct JDDC 2.1, a large-scale multimodal multi-turn dialogue dataset collected from a mainstream Chinese E-commerce platform 1 , containing about 246K dialogue sessions, 3M utterances, and 507K images, along with product knowledge bases and image category annotations. Over our dataset, we jointly define four tasks: the multimodal dialogue response generation task, the multimodal query rewriting task, the multimodal dialogue discourse parsing task, and the multimodal dialogue summarization task. JDDC 2.1 is the first corpus with annotations for all the above tasks over the same dialogue sessions, which facilitates the comprehensive research around the dialogue. In addition, we present several text-only and multimodal baselines and show the importance of visual information for these tasks. Our dataset and implements will be publicly available. * Equal contribution. | JDDC 2.1: A Multimodal Chinese Dialogue Dataset with Joint Tasks of Query Rewriting, Response Generation, Discourse Parsing, and Summarization |
d18855442 | In this paper, we present an unsupervized segmentation system tested on Mandarin Chinese. Following Harris's Hypothesis inKempe (1999)andTanaka-Ishii's (2005)reformulation, we base our work on the Variation of Branching Entropy. We improve on (Jin and Tanaka-Ishii, 2006) by adding normalization and viterbidecoding. This enable us to remove most of the thresholds and parameters from their model and to reach near state-of-the-art results (Wang et al., 2011) with a simpler system. We provide evaluation on different corpora available from the Segmentation bake-off II(Emerson, 2005)and define a more precise topline for the task using cross-trained supervized system available off-the-shelf(Zhang and Clark, 2010;Zhao and Kit, 2008;Huang and Zhao, 2007) | Unsupervized Word Segmentation: the case for Mandarin Chinese |
d7988049 | We have developed a word sense disambiguation algorithm, followingCheng and Wilensky (1997), to disambiguate among WordNet synsets. This algorithm is to be used in a cross-language information retrieval system, CINDOR, which indexes queries and documents in a language-neutral concept representation based on WordNet synsets. Our goal is to improve retrieval precision through word sense disambiguation. An evaluation against human disambiguation judgements suggests promise for our approach. | Word Sense Disambiguation for Cross-Language Information Retrieval |
d9726397 | The sociolinguistic situation in Arabic countries is characterized by diglossia (Ferguson, 1959) : whereas one variant Modern Standard Arabic (MSA) is highly codified and mainly used for written communication, other variants coexist in regular everyday's situations (dialects). Similarly, while a number of resources and tools exist for MSA (lexica, annotated corpora, taggers, parsers . . . ), very few are available for the development of dialectal Natural Language Processing tools. Taking advantage of the closeness of MSA and its dialects, one way to solve the problem of the lack of resources for dialects consists in exploiting available MSA resources and NLP tools in order to adapt them to process dialects. This paper adopts this general framework: we propose a method to build a lexicon of deverbal nouns for Tunisian (TUN) using MSA tools and resources as starting material.This work is licenced under a Creative Commons Attribution 4.0 International License. Page numbers and proceedings footer are added by the organizers. License details: http://creativecommons.org/licenses/by/4.0/ | Automatically building a Tunisian Lexicon for Deverbal Nouns |
d1258901 | State-of-the-art statistical machine translation systems use hypotheses from several maximum a posteriori inference steps, including word alignments and parse trees, to identify translational structure and estimate the parameters of translation models. While this approach leads to a modular pipeline of independently developed components, errors made in these "single-best" hypotheses can propagate to downstream estimation steps that treat these inputs as clean, trustworthy training data. In this work we integrate N -best alignments and parses by using a probability distribution over these alternatives to generate posterior fractional counts for use in downstream estimation. Using these fractional counts in a DOPinspired syntax-based translation system, we show significant improvements in translation quality over a single-best trained baseline. | Wider Pipelines: N -Best Alignments and Parses in MT Training |
d247158648 | Entity linking (EL) is the task of linking entity mentions in a document to referent entities in a knowledge base (KB). Many previous studies focus on Wikipedia-derived KBs. There is little work on EL over Wikidata, even though it is the most extensive crowdsourced KB. The scale of Wikidata can open up many new real-world applications, but its massive number of entities also makes EL challenging. To effectively narrow down the search space, we propose a novel candidate retrieval paradigm based on entity profiling. Wikidata entities and their textual fields are first indexed into a text search engine (e.g., Elasticsearch). During inference, given a mention and its context, we use a sequence-tosequence (seq2seq) model to generate the profile of the target entity, which consists of its title and description. We use the profile to query the indexed search engine to retrieve candidate entities. Our approach complements the traditional approach of using a Wikipedia anchor-text dictionary, enabling us to further design a highly effective hybrid method for candidate retrieval. Combined with a simple cross-attention reranker, our complete EL framework achieves state-of-the-art results on three Wikidata-based datasets and strong performance on TACKBP-2010 1 . | Improving Candidate Retrieval with Entity Profile Generation for Wikidata Entity Linking |
d250390691 | This paper describes a method to quantify the amount of information H(t|s) added by the target sentence t that is not present in the source s in a neural machine translation system. We do this by providing the model the target sentence in a highly compressed form (a "cheat code"), and exploring the effect of the size of the cheat code. We find that the model is able to capture extra information from just a single float representation of the target and nearly reproduces the target with two 32-bit floats per target token. | Cheat Codes to Quantify Missing Source Information in Neural Machine Translation |
d6412912 | Sentence compression is the task of producing a summary at the sentence level. This paper focuses on three aspects of this task which have not received detailed treatment in the literature: training requirements, scalability, and automatic evaluation. We provide a novel comparison between a supervised constituentbased and an weakly supervised wordbased compression algorithm and examine how these models port to different domains (written vs. spoken text). To achieve this, a human-authored compression corpus has been created and our study highlights potential problems with the automatically gathered compression corpora currently used. Finally, we assess whether automatic evaluation measures can be used to determine compression quality. | Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures |
d7190753 | Topic segmentation plays an important role for discourse parsing and information retrieval. Due to the absence of training data, previous work mainly adopts unsupervised methods to rank semantic coherence between paragraphs for topic segmentation. In this paper, we present an intuitive and simple idea to automatically create a "quasi" training dataset, which includes a large amount of text pairs from the same or different documents with different semantic coherence. With the training corpus, we design a symmetric CNN neural network to model text pairs and rank the semantic coherence within the learning to rank framework. Experiments show that our algorithm is able to achieve competitive performance over strong baselines on several real-world datasets. | Learning to Rank Semantic Coherence for Topic Segmentation |
d1564278 | The past 10 years of event ordering research has focused on learning partial orderings over document events and time expressions. The most popular corpus, the TimeBank, contains a small subset of the possible ordering graph. Many evaluations follow suit by only testing certain pairs of events (e.g., only main verbs of neighboring sentences). This has led most research to focus on specific learners for partial labelings. This paper attempts to nudge the discussion from identifying some relations to all relations. We present new experiments on strongly connected event graphs that contain ∼10 times more relations per document than the TimeBank. We also describe a shift away from the single learner to a sieve-based architecture that naturally blends multiple learners into a precision-ranked cascade of sieves. Each sieve adds labels to the event graph one at a time, and earlier sieves inform later ones through transitive closure. This paper thus describes innovations in both approach and task. We experiment on the densest event graphs to date and show a 14% gain over state-of-the-art. | Dense Event Ordering with a Multi-Pass Architecture |
d2646100 | We compare and contrast the strengths and weaknesses of a syntax-based machine translation model with a phrase-based machine translation model on several levels. We briefly describe each model, highlighting points where they differ. We include a quantitative comparison of the phrase pairs that each model has to work with, as well as the reasons why some phrase pairs are not learned by the syntax-based model. We then evaluate proposed improvements to the syntax-based extraction techniques in light of phrase pairs captured. We also compare the translation accuracy for all variations.↔ i felt obliged to do my, is beyond the size limit of 4, so it is not extracted in this example.↔ felt | What Can Syntax-based MT Learn from Phrase-based MT? |
d2920304 | Computational linguistic approaches to sign languages could benefit from investigating how complexity influences structure. We investigate whether morphological complexity has an effect on the order of Verb (V) and Object (O) in Swedish Sign Language (SSL), on the basis of elicited data from five Deaf signers. We find a significant difference in the distribution of the orderings OV vs. VO, based on an analysis of morphological weight. While morphologically heavy verbs exhibit a general preference for OV, humanness seems to affect the ordering in the opposite direction, with [+human] Objects pushing towards a preference for VO. | Morphological Complexity Influences Verb-Object Order in Swedish Sign Language |
d1414264 | This paper proposes a Hidden Markov Model (HMM) and an HMM-based chunk tagger, from which a named entity (NE) recognition (NER) system is built to recognize and classify names, times and numerical quantities. Through the HMM, our system is able to apply and integrate four types of internal and external evidences: 1) simple deterministic internal feature of the words, such as capitalization and digitalization; 2) internal semantic feature of important triggers; 3) internal gazetteer feature; 4) external macro context feature. In this way, the NER problem can be resolved effectively. Evaluation of our system on MUC-6 and MUC-7 English NE tasks achieves F-measures of 96.6% and 94.1% respectively. It shows that the performance is significantly better than reported by any other machine-learning system. Moreover, the performance is even consistently better than those based on handcrafted rules. | Named Entity Recognition using an HMM-based Chunk Tagger |
d1716598 | We present the first domain adaptation model for authorship attribution to leverage unlabeled data. The model includes extensions to structural correspondence learning needed to make it appropriate for the task. For example, we propose a median-based classification instead of the standard binary classification used in previous work. Our results show that punctuation-based character n-grams form excellent pivot features. We also show how singular value decomposition plays a critical role in achieving domain adaptation, and that replacing (instead of concatenating) non-pivot features with correspondence features yields better performance. | Domain Adaptation for Authorship Attribution: Improved Structural Correspondence Learning |
d1756650 | A typical knowledge-based question answering (KB-QA) system faces two challenges: one is to transform natural language questions into their meaning representations (MRs); the other is to retrieve answers from knowledge bases (KBs) using generated MRs. Unlike previous methods which treat them in a cascaded manner, we present a translation-based approach to solve these two tasks in one unified framework. We translate questions to answers based on CYK parsing. Answers as translations of the span covered by each CYK cell are obtained by a question translation method, which first generates formal triple queries as MRs for the span based on question patterns and relation expressions, and then retrieves answers from a given KB based on triple queries generated. A linear model is defined over derivations, and minimum error rate training is used to tune feature weights based on a set of question-answer pairs. Compared to a KB-QA system using a state-of-the-art semantic parser, our method achieves better results. | Knowledge-Based Question Answering as Machine Translation |
d202542357 | We introduce a novel scheme for parsing a piece of text into its Abstract Meaning Representation (AMR): Graph Spanning based Parsing (GSP). One novel characteristic of GSP is that it constructs a parse graph incrementally in a top-down fashion. Starting from the root, at each step, a new node and its connections to existing nodes will be jointly predicted. The output graph spans the nodes by the distance to the root, following the intuition of first grasping the main ideas then digging into more details. The core semantic first principle emphasizes capturing the main ideas of a sentence, which is of great interest. We evaluate our model on the latest AMR sembank and achieve the state-of-the-art performance in the sense that no heuristic graph re-categorization is adopted. More importantly, the experiments show that our parser is especially good at obtaining the core semantics. | Core Semantic First: A Top-down Approach for AMR Parsing * |
d202542534 | Many tasks in natural language processing can be viewed as multi-label classification problems. However, most of the existing models are trained with the standard cross-entropy loss function and use a fixed prediction policy (e.g., a threshold of 0.5) for all the labels, which completely ignores the complexity and dependencies among different labels. In this paper, we propose a meta-learning method to capture these complex label dependencies. More specifically, our method utilizes a metalearner to jointly learn the training policies and prediction policies for different labels. The training policies are then used to train the classifier with the cross-entropy loss function, and the prediction policies are further implemented for prediction. Experimental results on fine-grained entity typing and text classification demonstrate that our proposed method can obtain more accurate multi-label classification results. | Learning to Learn and Predict: A Meta-Learning Approach for Multi-Label Classification |
d21707078 | Predicting emotion categories (e.g. anger, joy, sadness) expressed by a sentence is challenging due to inherent multi-label smaller pieces such as phrases and clauses. To date, emotion has been studied in single genre, while models of human behaviors or situational awareness in the event of disasters require emotion modeling in multi-genres. In this paper, we expand and unify existing annotated data in different genres (emotional blog post, news title, and movie reviews) using an inventory of 8 emotions from Plutchik's Wheel of Emotions tags. We develop systems for automatically detecting and classifying emotions in text, in different textual genres and granularity levels, namely, sentence and clause levels in a supervised setting. We explore the effectiveness of clause annotation in sentence-level emotion detection and classification (EDC). To our knowledge, our EDC system is the first to target the clause level; further we provide emotion annotation for movie reviews dataset for the first time. | Sentence and Clause Level Emotion Annotation, Detection, and Classification in a Multi-Genre Corpus |
d17676483 | Computer-based generation of natural language requires consideration of two different types of problems: i) determining the content and textual shape of what is to be said, and 2) transforming that message into English.A computational solution to the problems of deciding what to say and how to organize it effectively is proposed that relies on an interaction between structural and semantic processes.Schemas, which encode aspects of discourse structure, are used to guide the generation process.A focusing mechanism monitors the use of the schemas, providing constraints on what can be said at any point. These mechanisms have been implemented as part of a generation method within the context of a natural language database system, addressing the specific problem of responding to questions about database structure. | THE TEXT SYSTEM FO~NATURAL LANGUAGE GENERATION: AN OVERVIEW* |
d140101838 | This paper presents an algorithm for tagging words whose part-of-speech properties are unknown. Unlike previous work, the algorithm categorizes word tokens in con$ezt instead of word ~ypes. The algorithm is evaluated on the Brown Corpus.*Although Biber (1993) classifies collocations, these can also be ambiguous. For example, "for certain" has both senses of "certain": "particular" and "sure". | Distributional Part-of-Speech Tagging |
d252819441 | Named entity recognition (e.g., disease mention extraction) is one of the most relevant tasks for data mining in the medical field. Although it is a well-known challenge, the bulk of the efforts to tackle this task have been made using clinical texts commonly written in English. In this work, we present our contribution to the SocialDisNER competition, which consists of a transfer learning approach to extracting disease mentions in a corpus from Twitter written in Spanish. We fine-tuned a model based on mBERT and applied post-processing using regular expressions to propagate the entities identified by the model and enhance disease mention extraction. Our system achieved a competitive strict F1 of 0.851 on the testing data set. | NLP-CIC-WFU at SocialDisNER: Disease Mention Extraction in Spanish Tweets Using Transfer Learning and Search by Propagation |
d825928 | The research focus of computational coreference resolution has exhibited a shift from heuristic approaches to machine learning approaches in the past decade. This paper surveys the major milestones in supervised coreference research since its inception fifteen years ago. | Supervised Noun Phrase Coreference Research: The First Fifteen Years |
d248779864 | Coreference resolution over semantic graphs like AMRs aims to group the graph nodes that represent the same entity. This is a crucial step for making document-level formal semantic representations. With annotated data on AMR coreference resolution, deep learning approaches have recently shown great potential for this task, yet they are usually data hungry and annotating data is costly. We propose a general pretraining method using variational graph autoencoder (VGAE) for AMR coreference resolution, which can leverage any general AMR corpus and even automatically parsed AMR data. Experiments on benchmarks show that the pretraining approach achieves performance gains of up to 6% absolute F1 points. Moreover, our model significantly improves on the previous state-of-theart model by up to 11% F1 points. | Variational Graph Autoencoding as Cheap Supervision for AMR Coreference Resolution |
d5808228 | This paper describes an effort to rapidly develop language resources and component technology to support searching Cebuano news stories using English queries. Results from the first 60 hours of the exercise are presented. | Desparately Seeking Cebuano |
d237513406 | Minimal sentence pairs are frequently used to analyze the behavior of language models. It is often assumed that model behavior on contrastive pairs is predictive of model behavior at large. We argue that two conditions are necessary for this assumption to hold: First, a tested hypothesis should be well-motivated, since experiments show that contrastive evaluation can lead to false positives. Secondly, test data should be chosen such as to minimize distributional discrepancy between evaluation time and deployment time. For a good approximation of deployment-time decoding, we recommend that minimal pairs are created based on machine-generated text, as opposed to humanwritten references. We present a contrastive evaluation suite for English-German MT that implements this recommendation. 1 | On the Limits of Minimal Pairs in Contrastive Evaluation |
d2661260 | The emergence of dialogue on social medial neccessitates the development of new dialogue processing models. We argue that to address coherence and to infer the implicatures of social dialogue it is vital to understand the social aspirations of the dialogue participants.One key aspect of understanding social dialogue is to understand the intentions and goals of participants. In this paper, we present 11 social acts that capture a broad number of social intentions and goals. We define social acts as pragmatic speech acts designed to give insight into the socio-cognitive processes that individuals unconsciously go through when communicating in dialogue. Identification of the social acts is done using a combination of a generative model in which utterances are generated from gappy patterns, which define a given social act, and a series of binary classifiers. Our experimentation shows that we can capture these social acts with an overall F-measure of 50.4%. | Identification of Social Acts in Dialogue |
d252818961 | Learning word embeddings is an essential topic in natural language processing. Most existing works use a vast corpus as a primary source while training, but this requires massive time and space for data pre-processing and model training. We propose a new model, HG2Vec, that learns word embeddings utilizing only dictionaries and thesauri. Our model reaches the state-of-art on multiple word similarity and relatedness benchmarks. We demonstrate that dictionaries and thesauri are effective resources to learn word embeddings. In addition, we exploit a new context-focused loss that models transitive relationships between word pairs and balances the performance between similarity and relatedness benchmarks, yielding superior results. | HG2Vec: Improved Word Embeddings from Dictionary and Thesaurus Based Heterogeneous Graph |
d216641774 | Most of the successful and predominant methods for bilingual lexicon induction (BLI) are mapping-based, where a linear mapping function is learned with the assumption that the word embedding spaces of different languages exhibit similar geometric structures (i.e., approximately isomorphic). However, several recent studies have criticized this simplified assumption showing that it does not hold in general even for closely related languages. In this work, we propose a novel semi-supervised method to learn cross-lingual word embeddings for BLI. Our model is independent of the isomorphic assumption and uses nonlinear mapping in the latent space of two independently trained auto-encoders. Through extensive experiments on fifteen (15) different language pairs (in both directions) comprising resource-rich and low-resource languages from two different datasets, we demonstrate that our method outperforms existing models by a good margin. Ablation studies show the importance of different model components and the necessity of non-linear mapping. | LNMAP: Departures from Isomorphic Assumption in Bilingual Lexicon Induction Through Non-Linear Mapping in Latent Space |
d231632841 | Dialogue State Tracking (DST) forms a core component of automated chatbot based systems designed for specific goals like hotel, taxi reservation, tourist information etc. With the increasing need to deploy such systems in new domains, solving the problem of zero/fewshot DST has become necessary. There has been a rising trend for learning to transfer knowledge from resource-rich domains to unknown domains with minimal need for additional data. In this work, we explore the merits of meta-learning algorithms for this transfer and hence, propose a meta-learner D-REPTILE specific to the DST problem. With extensive experimentation, we provide clear evidence of benefits over conventional approaches across different domains, methods, base models and datasets with significant (5-25%) improvement over the baseline in lowdata setting. Our proposed meta-learner is agnostic of the underlying model and hence any existing state-of-the-art DST system can improve its performance on unknown domains using our training strategy. | Few Shot Dialogue State Tracking using Meta-learning |
d218581602 | Learning to follow instructions is of fundamental importance to autonomous agents for vision-and-language navigation (VLN). In this paper, we study how an agent can navigate long paths when learning from a corpus that consists of shorter ones. We show that existing state-of-the-art agents do not generalize well. To this end, we propose BabyWalk, a new VLN agent that is learned to navigate by decomposing long instructions into shorter ones (BabySteps) and completing them sequentially. A special design memory buffer is used by the agent to turn its past experiences into contexts for future steps. The learning process is composed of two phases. In the first phase, the agent uses imitation learning from demonstration to accomplish BabySteps. In the second phase, the agent uses curriculum-based reinforcement learning to maximize rewards on navigation tasks with increasingly longer instructions. We create two new benchmark datasets (of long navigation tasks) and use them in conjunction with existing ones to examine BabyWalk's generalization ability. Empirical results show that BabyWalk achieves state-of-the-art results on several metrics, in particular, is able to follow long instructions better. The codes and the datasets are released on our project page https://github.com/ Sha-Lab/babywalk. | BabyWalk: Going Farther in Vision-and-Language Navigation by Taking Baby Steps |
d218581436 | Fake news detection is a very prominent and essential task in the field of journalism. This challenging problem is seen so far in the field of politics, but it could be even more challenging when it is to be determined in the multi-domain platform. In this paper, we propose two effective models based on deep learning for solving fake news detection problem in online news contents of multiple domains. We evaluate our techniques on the two recently released datasets, namely Fake-News AMT and Celebrity for fake news detection. The proposed systems yield encouraging performance, outperforming the current handcrafted feature engineering based state-of-theart system with a significant margin of 3.08% and 9.3% by the two models, respectively. In order to exploit the datasets, available for the related tasks, we perform cross-domain analysis (i.e. model trained on FakeNews AMT and tested on Celebrity and vice versa) to explore the applicability of our systems across the domains. | A Deep Learning Approach for Automatic Detection of Fake News |
d259376828 | In this paper we present and discuss the results achieved by the "Augustine of Hippo" team at SemEval-2023 Task 4 about human value detection. In particular, we provide a quantitative and qualitative reviews of the results obtained by SuperASKE, discussing respectively performance metrics and classification errors. Finally, we present our main contribution: an explainable and unsupervised approach mapping arguments to concepts, followed by a supervised classification model mapping concepts to human values. | Augustine Of Hippo at SemEval-2023 Task 4: An Explainable Knowledge Extraction Method to Identify Human Values in Arguments with SuperASKE |
d219301328 | ||
d11177932 | Language processing tools suffer from significant performance drops in social media domain due to its continuously evolving language. Transforming non-standard words into their standard forms has been studied as a step towards proper processing of ill-formed texts. This work describes a normalization system that considers contextual and lexical similarities between standard and non-standard words for removing noise in texts. A bipartite graph that represents contexts shared by words in a large unlabeled text corpus is utilized for exploring normalization candidates via random walks. Input context of a non-standard word in a given sentence is tailored in cases where a direct match to shared contexts is not possible. The performance of the system was evaluated on Turkish social media texts. | Context Tailoring for Text Normalization |
d248798649 | Recently, the structural reading comprehension (SRC) task on web pages has attracted increasing research interests. Although previous SRC work has leveraged extra information such as HTML tags or XPaths, the informative topology of web pages is not effectively exploited. In this work, we propose a Topological Information Enhanced model (TIE), which transforms the token-level task into a tag-level task by introducing a two-stage process (i.e. node locating and answer refining). Based on that, TIE integrates Graph Attention Network (GAT) and Pre-trained Language Model (PLM) to leverage the topological information of both logical structures and spatial structures. Experimental results demonstrate that our model outperforms strong baselines and achieves state-ofthe-art performances on the web-based SRC benchmark WebSRC at the time of writing. The code of TIE will be publicly available at https://github.com/X-LANCE/TIE. | TIE: Topological Information Enhanced Structural Reading Comprehension on Web Pages |
d215548225 | Large pre-trained language models (LMs) are known to encode substantial amounts of linguistic information. However, high-level reasoning skills, such as numerical reasoning, are difficult to learn from a language-modeling objective only. Consequently, existing models for numerical reasoning have used specialized architectures with limited flexibility. In this work, we show that numerical reasoning is amenable to automatic data generation, and thus one can inject this skill into pre-trained LMs, by generating large amounts of data, and training in a multi-task setup. We show that pre-training our model, GENBERT, on this data, dramatically improves performance on DROP (49.3 → 72.3 F 1 ), reaching performance that matches state-of-the-art models of comparable size, while using a simple and general-purpose encoder-decoder architecture. Moreover, GENBERT generalizes well to math word problem datasets, while maintaining high performance on standard RC tasks. Our approach provides a general recipe for injecting skills into large pre-trained LMs, whenever the skill is amenable to automatic data augmentation. * These authors contributed equally. (b) fine-tuning pre-trained LM numerical reasoning reading compr.Numerical Data (ND) | Injecting Numerical Reasoning Skills into Language Models |
d3141720 | We present dependency-based n-gram models for general-purpose, widecoverage, probabilistic sentence realisation. Our method linearises unordered dependencies in input representations directly rather than via the application of grammar rules, as in traditional chartbased generators. The method is simple, efficient, and achieves competitive accuracy and complete coverage on standard English (Penn-II, 0.7440 BLEU, 0.05 sec/sent) and Chinese (CTB6, 0.7123 BLEU, 0.14 sec/sent) test data. | Dependency-Based N-Gram Models for General Purpose Sentence Realisation |
d220047315 | We propose a novel text editing task, referred to as fact-based text editing, in which the goal is to revise a given document to better describe the facts in a knowledge base (e.g., several triples). The task is important in practice because reflecting the truth is a common requirement in text editing. First, we propose a method for automatically generating a dataset for research on fact-based text editing, where each instance consists of a draft text, a revised text, and several facts represented in triples. We apply the method into two public tableto-text datasets, obtaining two new datasets consisting of 233k and 37k instances, respectively. Next, we propose a new neural network architecture for fact-based text editing, called FACTEDITOR, which edits a draft text by referring to given facts using a buffer, a stream, and a memory. A straightforward approach to address the problem would be to employ an encoder-decoder model. Our experimental results on the two datasets show that FACTE-DITOR outperforms the encoder-decoder approach in terms of fidelity and fluency. The results also show that FACTEDITOR conducts inference faster than the encoder-decoder approach. | Fact-based Text Editing |
d235390653 | Automatic speech recognition (ASR) in Sanskrit is interesting, owing to the various linguistic peculiarities present in the language. The Sanskrit language is lexically productive, undergoes euphonic assimilation of phones at the word boundaries and exhibits variations in spelling conventions and in pronunciations. In this work, we propose the first large scale study of automatic speech recognition (ASR) in Sanskrit, with an emphasis on the impact of unit selection in Sanskrit ASR. In this work, we release a 78 hour ASR dataset for Sanskrit, which faithfully captures several of the linguistic characteristics expressed by the language. We investigate the role of different acoustic model and language model units in ASR systems for Sanskrit. We also propose a new modelling unit, inspired by the syllable level unit selection, that captures character sequences from one vowel in the word to the next vowel. We also highlight the importance of choosing graphemic representations for Sanskrit and show the impact of this choice on word error rates (WER). Finally, we extend these insights from Sanskrit ASR for building ASR systems in two other Indic languages, Gujarati and Telugu. For both these languages, our experimental results show that the use of phonetic based graphemic representations in ASR results in performance improvements as compared to ASR systems that use native scripts. | Automatic Speech Recognition in Sanskrit: A New Speech Corpus and Modelling Insights |
d249538440 | Recent work has demonstrated the successful extraction of training data from generative language models. However, it is not evident whether such extraction is feasible in text classification models since the training objective is to predict the class label as opposed to next-word prediction. This poses an interesting challenge and raises an important question regarding the privacy of training data in text classification settings. Therefore, we study the potential privacy leakage in the text classification domain by investigating the problem of unintended memorization of training data that is not pertinent to the learning task. We propose an algorithm to extract missing tokens of a partial text by exploiting the likelihood of the class label provided by the model. We test the effectiveness of our algorithm by inserting canaries into the training set and attempting to extract tokens in these canaries post-training. In our experiments, we demonstrate that successful extraction is possible to some extent. This can also be used as an auditing strategy to assess any potential unauthorized use of personal data without consent. | Privacy Leakage in Text Classification: A Data Extraction Approach |
d218551245 | We propose a sentence-level language model which selects the next sentence in a story from a finite set of fluent alternatives. Since it does not need to model fluency, the sentence-level language model can focus on longer range dependencies, which are crucial for multisentence coherence. Rather than dealing with individual words, our method treats the story so far as a list of pre-trained sentence embeddings and predicts an embedding for the next sentence, which is more efficient than predicting word embeddings. Notably this allows us to consider a large number of candidates for the next sentence during training. We demonstrate the effectiveness of our approach with state-of-the-art accuracy on the unsupervised Story Cloze task and with promising results on larger-scale next sentence prediction tasks. | Toward Better Storylines with Sentence-Level Language Models |
d233029500 | Natural Language Processing for Digital Humanities | |
d250179898 | La Communication Alternative et Augmentée (CAA) prend une place importante chez les personnes en situation de handicap ainsi que leurs proches à cause de la difficulté de son utilisation. Pour réduire ce poids, l'utilisation d'outils de traduction de la parole en pictogrammes est pertinente. De plus, ils peuvent être d'une grande aide pour l'accessibilité communicative dans le milieu hospitalier. Dans cet article, nous présentons un projet de recherche visant à développer un système de traduction de la parole vers des pictogrammes. Il met en jeu une chaîne de traitement comportant plusieurs axes relevant du traitement automatique des langues et de la parole, tels que la reconnaissance automatique de la parole, l'analyse syntaxique, la simplification de texte et la traduction automatique vers les pictogrammes. Nous présentons les difficultés liées à chacun de ces axes ainsi que, pour certains, les pistes de résolution.ABSTRACTSimplification and automatic translation of speech into pictogramsAlternative and Augmentative Communication (AAC) is becoming an important issue among people with disabilities and their relatives because of the difficulty of its use. To reduce this burden, using speech translation tools in pictograms is relevant. In addition, they can be of great help for communicative accessibility in the hospital environment. Developing such tools requires in-depth research on several axes of automatic language processing. In this article, we present a research project aiming at developing a system for translating speech into pictograms. It involves a processing chain with several axes related to automatic language and speech processing, such as automatic speech recognition, syntactic analysis, sentence simplification, and automatic translation to pictogram units. We present the difficulties related to each of these axes as well as, for some, the avenues of resolution. | Une chaîne de traitements pour la simplification automatique de la parole et sa traduction automatique vers des pictogrammes |
d259002935 | Obtaining information about loan words and irregular morphological patterns can be difficult for low-resource languages. Using Sakha as an example, we show that it is possible to exploit known phonemic regularities such as vowel harmony and consonant distributions to identify loan words and irregular patterns, which can be helpful in rule-based downstream tasks such as parsing and POS-tagging. We evaluate phonemically inspired methods for loanword detection, combined with bigram vowel transition probabilities to inspect irregularities in the morphology of loanwords. We show that both these techniques can be useful for the detection of such patterns. Finally, we inspect the plural suffix -ЛАр [-LAr] to observe some of the variation in morphology between native and foreign words. | Phonotactics as an Aid in Low Resource Loan Word Detection and Morphological Analysis in Sakha |
d15268382 | We present an approach to perform external plagiarism analysis by applying several similarity detection techniques, such as lexical measures and a textual entailment recognition system developed by our research group. Some of the least expensive features of this system are applied to all corpus documents to detect those that are likely to be plagiarized. After this is done, the whole system is applied over this subset of documents to extract the exact n-grams that have been plagiarized, given that we now have less data to process and therefore can use a more complex and costly function. Apart from the application of strictly lexical measures, we also experiment with a textual entailment recognition system to detect plagiarisms with a high level of obfuscation. In addition, we experiment with the application of a spell corrector and a machine translation system to handle misspellings and plagiarisms translated into different languages, respectively. | Investigating Advanced Techniques for Document Content Similarity Applied to External Plagiarism Analysis |
d9977148 | At least in the realm of fast parsing, the mass-count distinction has led the life of a wallflower. We argue in this paper that this should not be so. In particular, we argue, both theoretical linguistics and computational linguistics can gain by a corpus-based investigation of this distinction: Computational linguists get more accurate parses; the knowledge extracted from these parses becomes more reliable; theoretical linguists are presented with new data in a field that has been intensely discussed and yet remains in a state that is not satisfactory from a practical point of view. | The Mass-Count Distinction: Acquisition and Disambiguation |
d14685368 | In this paper we present an active learning approach used to create an annotated corpus of literal and nonliteral usages of verbs. The model uses nearly unsupervised word-sense disambiguation and clustering techniques. We report on experiments in which a human expert is asked to correct system predictions in different stages of learning: (i) after the last iteration when the clustering step has converged, or (ii) during each iteration of the clustering algorithm. The model obtains an f-score of 53.8% on a dataset in which literal/nonliteral usages of 25 verbs were annotated by human experts. In comparison, the same model augmented with active learning obtains 64.91%. We also measure the number of examples required when model confidence is used to select examples for human correction as compared to random selection. The results of this active learning system have been compiled into a freely available annotated corpus of literal/nonliteral usage of verbs in context. | Active Learning for the Identification of Nonliteral Language * |
d221853543 | Dans ce travail nous avons recours aux variations de f0 et d'intensité de 44 locuteurs francophones à partir de séquences de 4 secondes de parole spontanée pour comprendre comment ces paramètres prosodiques peuvent être utilisés pour caractériser des locuteurs. Une classification automatique est effectuée avec un réseau de neurones convolutifs, fournissant comme réponse des scores de probabilité pour chacun des 44 locuteurs modélisés. Une représentation par spectrogrammes a été utilisée comme référence pour le même système de classification. Nous avons pu mettre en avant la pertinence de l'intensité, et lorsque les deux paramètres prosodiques sont combinés pour représenter les locuteurs nous observons un score qui atteint en moyenne 59 % de bonnes classifications.ABSTRACTCNN speaker characterisation through prosody : spectrogram comparisonIn this study we focused on f0 and intensity variation in four-second spontaneous speech sequences by 44 French speakers in order to evaluate the strength of prosodic parameters for speaker characterisation. We used a deep Convolutional Neural Network (CNN) with f0 and/or intensity values as input in a classification task, where the system had to classify 44 speakers in a closed dataset. Spectrograms were also used as input with the same CNN architecture as a benchmark for maximum possible performance. Results show that f0 and intensity are complementary as together they yield 59 % classification precision. MOTS-CLÉS : Caractérisation du locuteur, intensité, f 0 , prosodie, réseaux de neurones convolutifs. | Caractérisation du locuteur par CNN à l'aide des contours d'intensité et d'intonation : comparaison avec le spectrogramme |
d221265974 | This paper presents our systems for SemEval 2020 Shared Task 11: Detection of Propaganda Techniques in News Articles. We participate in both the span identification and technique classification subtasks and report on experiments using different BERT-based models along with handcrafted features. Our models perform well above the baselines for both tasks, and we contribute ablation studies and discussion of our results to dissect the effectiveness of different features and techniques with the goal of aiding future studies in propaganda detection. | Team DoNotDistribute at SemEval-2020 Task 11: Features, Finetuning, and Data Augmentation in Neural Models for Propaganda Detection in News Articles |
d222177226 | With the growth of computing power neural machine translation (NMT) models also grow accordingly and become better. However, they also become harder to deploy on edge devices due to memory constraints. To cope with this problem, a common practice is to distill knowledge from a large and accurately-trained teacher network (T ) into a compact student network (S). Although knowledge distillation (KD) is useful in most cases, our study shows that existing KD techniques might not be suitable enough for deep NMT engines, so we propose a novel alternative. In our model, besides matching T and S predictions we have a combinatorial mechanism to inject layer-level supervision from T to S. In this paper, we target low-resource settings and evaluate our translation engines for Portuguese→English, Turkish→English, and English→German directions. Students trained using our technique have 50% fewer parameters and can still deliver comparable results to those of 12-layer teachers. | Why Skip If You Can Combine: A Simple Knowledge Distillation Technique for Intermediate Layers |
d222208886 | Response generation for task-oriented dialogues implicitly optimizes two objectives at the same time: task completion and language quality. Conditioned response generation serves as an effective approach to separately and better optimize these two objectives. Such an approach relies on system action annotations which are expensive to obtain. To alleviate the need of action annotations, latent action learning is introduced to map each utterance to a latent representation. However, this approach is prone to over-dependence on the training data, and the generalization capability is thus restricted. To address this issue, we propose to learn natural language actions that represent utterances as a span of words. This explicit action representation promotes generalization via the compositional structure of language. It also enables an explainable generation process. Our proposed unsupervised approach learns a memory component to summarize system utterances into a short span of words. To further promote a compact action representation, we propose an auxiliary task that restores state annotations as the summarized dialogue context using the memory component. Our proposed approach outperforms latent action baselines on MultiWOZ, a benchmark multi-domain dataset. | Generalizable and Explainable Dialogue Generation via Explicit Action Learning |
d174801388 | Recent approaches to data-to-text generation have shown great promise thanks to the use of large-scale datasets and the application of neural network architectures which are trained end-to-end. These models rely on representation learning to select content appropriately, structure it coherently, and verbalize it grammatically, treating entities as nothing more than vocabulary tokens. In this work we propose an entity-centric neural architecture for data-to-text generation. Our model creates entity-specific representations which are dynamically updated. Text is generated conditioned on the data input and entity memory representations using hierarchical attention at each time step. We present experiments on the ROTOWIRE benchmark and a (five times larger) new dataset on the baseball domain which we create. Our results show that the proposed model outperforms competitive baselines in automatic and human evaluation. 1 | Data-to-text Generation with Entity Modeling |
d782758 | This paper describes a hybrid Chinese word segmenter that is being developed as part of a larger Chinese unknown word resolution system. The segmenter consists of two components: a tagging component that uses the transformation-based learning algorithm to tag each character with its position in a word, and a merging component that transforms a tagged character sequence into a word-segmented sentence. In addition to the position-of-character tags assigned to the characters, the merging component makes use of a number of heuristics to handle non-Chinese characters, numeric type compounds, and long words. The segmenter achieved a 92.8% F-score and a 72.8% recall for OOV words in the closed track of the Peking University Corpus in the Second International Chinese Word Segmentation Bakeoff. | Towards a Hybrid Model for Chinese Word Segmentation |
d6378898 | In this work, we propose a novel, implicitly-defined neural network architecture and describe a method to compute its components. The proposed architecture forgoes the causality assumption used to formulate recurrent neural networks and instead couples the hidden states of the network, allowing improvement on problems with complex, long-distance dependencies. Initial experiments demonstrate the new architecture outperforms both the Stanford Parser and baseline bidirectional networks on the Penn Treebank Part-of-Speech tagging task and a baseline bidirectional network on an additional artificial random biased walk task. | Implicitly-Defined Neural Networks for Sequence Labeling * |
d237491889 | We introduce EfficientCL, a memory-efficient continual pretraining method that applies contrastive learning with novel data augmentation and curriculum learning. For data augmentation, we stack two types of operation sequentially: cutoff and PCA jittering. While pretraining steps proceed, we apply curriculum learning by incrementing the augmentation degree for each difficulty step. After data augmentation, we apply contrastive learning on projected embeddings of original and augmented examples. When fine-tuned on GLUE benchmark, our model outperforms baseline models, especially for sentence-level tasks. Additionally, this improvement is achieved with only 70% of computational memory compared to the baseline model. 1 | Efficient Contrastive Learning via Novel Data Augmentation and Curriculum Learning |
d236459941 | Transformer-based language models pretrained on large amounts of text data have proven remarkably successful in learning generic transferable linguistic representations. Here we study whether structural guidance leads to more human-like systematic linguistic generalization in Transformer language models without resorting to pre-training on very large amounts of data. We explore two general ideas. The "Generative Parsing" idea jointly models the incremental parse and word sequence as part of the same sequence modeling task. The "Structural Scaffold" idea guides the language model's representation via additional structure loss that separately predicts the incremental constituency parse. We train the proposed models along with a vanilla Transformer language model baseline on a 14 million-token and a 46 million-token subset of the BLLIP dataset, and evaluate models' syntactic generalization performances on SG Test Suites and sized BLiMP. Experiment results across two benchmarks suggest converging evidence that generative structural supervisions can induce more robust and humanlike linguistic generalization in Transformer language models without the need for data intensive pre-training. | Structural Guidance for Transformer Language Models |
d6890211 | This paper presents a language model and its application to sentence structure manipulations for various natural language applications including human-computer communications. Building a working natural language dialog systems requires the integration of solutions to many of the important subproblems of natural language processing. In order to materialize any of these subproblems, handling of natural language expressions plays a central role; natural language manipulation facilities axe indispensable for any natural language dialog systems. Concept Compound Manipulation Language (CCML) proposed in this paper is intended to provide a practical means to manipulate sentences by means of formal uniform operations. | Language Model and Sentence Structure Manipulations for Natural Language Application Systems |
d8071352 | In this paper we present a tool to automatically determine the semantic interpretation of some ambiguous sentences in Spanish, the so-called "se-constructions". These sentences are syntactically similar, but their argument structure is different. The performance of the disambiguation procedure has been evaluated against a manually annotated corpus, achieving 95% precision. | Semantic categorization of Spanish se-constructions |
d220045419 | Evaluating image captions is very challenging partially due to the fact that there are multiple correct captions for every single image. Most of the existing one-to-one metrics operate by penalizing mismatches between reference and generative caption without considering the intrinsic variance between ground truth captions. It usually leads to over-penalization and thus a bad correlation to human judgment. Recently, the latest one-to-one metric BERTScore can achieve high human correlation in system-level tasks while some issues can be fixed for better performance. In this paper, we propose a novel metric based on BERTScore that could handle such a challenge and extend BERTScore with a few new features appropriately for image captioning evaluation. The experimental results show that our metric achieves state-of-the-art human judgment correlation. | Improving Image Captioning Evaluation by Considering Inter References Variance |
d6111717 | This paper deals with the problem of recognizing and extracting acronymdefinition pairs in Swedish medical texts. This project applies a rule-based method to solve the acronym recognition task and compares and evaluates the results of different machine learning algorithms on the same task. The method proposed is based on the approach that acronym-definition pairs follow a set of patterns and other regularities that can be usefully applied for the acronym identification task. Supervised machine learning was applied to monitor the performance of the rule-based method, using Memory Based Learning (MBL). The rule-based algorithm was evaluated on a hand tagged acronym corpus and performance was measured using standard measures recall, precision and fscore. The results show that performance could further improve by increasing the training set and modifying the input settings for the machine learning algorithms. An analysis of the errors produced indicates that further improvement of the rulebased method requires the use of syntactic information and textual pre-processing. | Automatic Acronym Recognition |
d194380963 | Notre travail porte sur la détection automatique des segments en relation de reformulation paraphrastique dans les corpus oraux. L'approche proposée est une approche syntagmatique qui tient compte des marqueurs de reformulation paraphrastique et des spécificités de l'oral. Les données de référence sont consensuelles. Une méthode automatique fondée sur l'apprentissage avec les CRF est proposée afin de détecter les segments paraphrasés. Différents descripteurs sont exploités dans une fenêtre de taille variable. Les tests effectués montrent que les segments en relation de paraphrase sont assez difficiles à détecter, surtout avec leurs frontières correctes. Les meilleures moyennes atteignent 0,65 de Fmesure, 0,75 de précision et 0,63 de rappel. Nous avons plusieurs perspectives à ce travail pour améliorer la détection des segments en relation de paraphrase et pour étudier les données depuis d'autres points de vue.Abstract....des conférences enfin disons des causeries... Automatic detection of segments with paraphrase relation in spoken corpora rephrasings. Our work addresses automatic detection of segments with paraphrastic rephrasing relation in spoken corpus. The proposed approach is syntagmatic. It is based on paraphrastic rephrasing markers and the specificities of the spoken language. The reference data used are consensual. Automatic method based on machine learning using CRFs is proposed in order to detect the segments that are paraphrased. Different descriptors are exploited within a window with various sizes. The tests performed indicate that the segments that are in paraphrastic relation are quite difficult to detect. Our best average reaches up to 0.65 F-measure, 0.75 precision, and 0.63 recall. We have several perspectives to this work for improving the detection of segments that are in paraphrastic relation and for studying the data from other points of view.Mots-clés : Corpus oraux, Paraphrase, Reformulation, Marqueur de reformulation paraphrastique, Apprentissage supervisé. ing. NATALIA GRABAR, IRIS ESHKOL-TARAVELLA relève également du paraphrasage : entre les différentes versions d'une oeuvre littéraire (Fuchs, 1982), d'un article de Wikipédia (Vila et al., 2014) ou d'un article scientifique, les auteurs peuvent réécrire plusieurs fois leur texte avant de produire celui qui les satisfait enfin mieux. En dehors des situations communes de la vie, la paraphrase joue aussi un rôle important pour différentes applications de TAL (Androutsopoulos & Malakasiotis, 2010; Madnani & Dorr, 2010; Bouamor et al., 2012). L'objectif est alors de détecter les expressions linguistiques formellement différentes mais véhiculant une sémantique similaire ou proche : -En recherche et extraction d'information, et dans les systèmes de questions-réponses, les paraphrases permettent d'augmenter la couverture des résultats grâce aux expressions équivalentes entre les requêtes ou les questions et les textes dans lesquelles les réponses doivent être trouvées. Par exemple, les paires {infarctus du myocarde; crise cardiaque} et {maladie d'Alzheimer; maladie neurodégénérative} contiennent des expressions différentes qui véhiculent cependant une sémantique identique ou proche. Si le système automatique dispose de telles connaissances, la couverture et la qualité de ses résultats peuvent être améliorées ; -En traduction automatique, les paraphrases permettent d'éviter des répétitions lexicales et peuvent introduire ainsi une légèreté du texte cible. Par exemple, le segment original en anglaisFigure 10.2 shows money growth and output growth. There is a strong, but not absolute, link between money growth and output growth est traduit en français de manière à éviter la répétition money growth and output growth : Le graphique 10.2 montre une augmentation des fonds et de la production. Il existe entre ces éléments un lien étroit, bien que non absolu (Scarpa, 2010). Différentes langues ont en effet une tolérance variable vis-à-vis des répétitions (Hatim & Mason, 1990; Baker, 1992) ; -L'inférence textuelle (Dagan et al., 2013) consiste à établir une relation entre deux segments textuels, appelés Texte et Hypothèse. L'inférence textuelle est une relation directionnelle, dans laquelle la vérité de l'Hypothèse peut être inférée à partir du sens du Texte, ou, en d'autres mots, il est possible de vérifier si l'Hypothèse est subsumée par le Texte. Par exemple, le Texte The drugs that slow down or halt Alzheimer's disease work best the earlier you administer them permet d'inférer que l'Hypothèse Alzheimer's disease is treated by drugs est vraie ; par contre, l'Hypothèse Alzheimer's disease is cured by drugs ne peut pas être inférée à partir de ce Texte. Dans cet exemple, le lien de paraphrase entre les paires {administer drugs; treated by drugs} et {slow down or halt; cured by drugs} permet justement d'établir ce lien d'inférence entre le Texte et l'Hypothèse. Comme ces quelques exemples montrent, en fonction des objectifs poursuivis, la paraphrase requiert des phénomènes linguistiques plus ou moins nombreux. L'étendue des classifications de paraphrases proposées peut être ainsi plus ou moins couvrante, et aller de 25 catégories de paraphrases (Bhagat & Hovy, 2013) à 67 fonctions lexicales pour le paraphrasage(Melčuk, 1988). Le plus souvent, ces classifications focalisent sur un aspect donné, comme par exemple les moyens linguistiques mis en oeuvre(Melčuk, 1988;Vila et al., 2011;Bhagat & Hovy, 2013), la taille de l'unité paraphrasée(Flottum, 1995;Fujita, 2010;Bouamor, 2012), les connaissances requises(Milicevic, 2007), le registre de langue. À notre connaissance, la seule classification multidimensionnelle est celle de(Milicevic, 2007): elle couvre plusieurs des dimensions mentionnées. Dans notre travail précédent, nous avons également proposé de travailler sur une classification à plusieurs dimensions(Eshkol-Taravella & Grabar, 2014). Elle prend en compte les dimensions suivantes : -la catégorie syntaxique des segments paraphrasés, -le type de la relation lexicale entre ces segments (e.g. hyperonyme, synonyme, antonyme, instance, méronyme), -le type de la modification lexicale (e.g. remplacement, suppression, ajout), -le type de la modification morphologique (i.e. flexion, dérivation, composition), -le type de la modification syntaxique (e.g. passif/actif), -le type de la relation pragmatique liée aux fonctionnalités de la paraphrase et de la reformulation (e.g. définition, explication, précision, résultat, correction linguistique, correction référentielle, équivalence). Dans notre travail, nous adoptons donc une acception large de la paraphrase.Travaux existants en acquisition automatique de paraphrasesPlusieurs approches sont proposées pour la détection automatique de la paraphrase. De manière générale, ces approches reposent sur les propriétés paradigmatiques des mots et sur leur capacité de se substituer mutuellement dans un contexte donné. Ces approches dépendent du type de corpus exploités. Quatre types de corpus sont généralement distingués : | 22 ème Traitement Automatique des Langues Naturelles |
d216553636 | Most general-purpose extractive summarization models are trained on news articles, which are short and present all important information upfront. As a result, such models are biased by position and often perform a smart selection of sentences from the beginning of the document. When summarizing long narratives, which have complex structure and present information piecemeal, simple position heuristics are not sufficient. In this paper, we propose to explicitly incorporate the underlying structure of narratives into general unsupervised and supervised extractive summarization models. We formalize narrative structure in terms of key narrative events (turning points) and treat it as latent in order to summarize screenplays (i.e., extract an optimal sequence of scenes). Experimental results on the CSI corpus of TV screenplays, which we augment with scene-level summarization labels, show that latent turning points correlate with important aspects of a CSI episode and improve summarization performance over general extractive algorithms, leading to more complete and diverse summaries. | Screenplay Summarization Using Latent Narrative Structure |
d235097329 | We propose MULTIOPED 1 , an open-domain news editorial corpus that supports various tasks pertaining to the argumentation structure in news editorials, focusing on automatic perspective discovery. News editorial is a genre of persuasive text, where the argumentation structure is usually implicit. However, the arguments presented in an editorial typically center around a concise, focused thesis, which we refer to as their perspective. MULTIOPED aims at supporting the study of multiple tasks relevant to automatic perspective discovery, where a system is expected to produce a singlesentence thesis statement summarizing the arguments presented. We argue that identifying and abstracting such natural language perspectives from editorials is a crucial step toward studying the implicit argumentation structure in news editorials.We first discuss the challenges and define a few conceptual tasks towards our goal. To demonstrate the utility of MULTIOPED and the induced tasks, we study the problem of perspective summarization in a multi-task learning setting, as a case study. We show that, with the induced tasks as auxiliary tasks, we can improve the quality of the perspective summary generated. We hope that MULTIOPED will be a useful resource for future studies on argumentation in the news editorial domain. | MULTIOPED: A Corpus of Multi-Perspective News Editorials |
d9324974 | Several semi-supervised learning methods have been proposed to leverage unlabeled data, but imbalanced class distributions in the data set can hurt the performance of most algorithms. In this paper, we adapt the new approach of contrast classifiers for semi-supervised learning. This enables us to exploit large amounts of unlabeled data with a skewed distribution. In experiments on a speech act (agreement/disagreement) classification problem, we achieve better results than other semi-supervised methods. We also obtain performance comparable to the best results reported so far on this task and outperform systems with equivalent feature sets. | Agreement/Disagreement Classification: Exploiting Unlabeled Data using Contrast Classifiers |
d1831149 | We propose three enhancements to the treeto-string (TTS) transducer for machine translation: first-level expansion-based normalization for TTS templates, a syntactic alignment framework integrating the insertion of unaligned target words, and subtree-based ngram model addressing the tree decomposition probability. Empirical results show that these methods improve the performance of a TTS transducer based on the standard BLEU-4 metric. We also experiment with semantic labels in a TTS transducer, and achieve improvement over our baseline system. | Improved Tree-to-string Transducer for Machine Translation |
d203610580 | We present our Generative Enhanced Model (GEM) that we used to create samples awarded the first prize on the FEVER 2.0 Breakers Task. GEM is the extended language model developed upon GPT-2 architecture. The addition of novel target vocabulary input to the already existing context input enabled controlled text generation. The training procedure resulted in creating a model that inherited the knowledge of pretrained GPT-2, and therefore was ready to generate natural-like English sentences in the task domain with some additional control. As a result, GEM generated malicious claims that mixed facts from various articles, so it became difficult to classify their truthfulness. | TMLab: Generative Enhanced Model (GEM) for adversarial attacks |
d16007831 | We present a new compositional tense-aspect deindexing mechanism that makes use of tense trees as components of discourse contexts. The mechanism allows reference episodes to be correctly identified even for embedded clauses and for discourse that involves shifts in temporal perspective, and permits deindexed logical forms to be automatically computed with a small number of deindexing rules. | TENSE TREES AS THE "FINE STRUCTURE" OF DISCOURSE |
d252819502 | Generating synthetic data for supervised learning from large-scale pre-trained language models has enhanced performances across several NLP tasks, especially in low-resource scenarios. In particular, many studies of data augmentation employ masked language models to replace words with other words in a sentence. However, most of them are evaluated on sentence classification tasks and cannot immediately be applied to tasks related to the sentence structure. In this paper, we propose a simple yet effective approach to generating sentences with a coordinate structure in which the boundaries of its conjuncts are explicitly specified. For a given span in a sentence, our method embeds a mask with a coordinating conjunction in two ways ("X and <mask>", "<mask> and X") and forces masked language models to fill the two blanks with an identical text. To achieve this, we introduce decoding methods for BERT and T5 models with the constraint that predictions for different masks are synchronized. Furthermore, we develop a training framework that effectively selects synthetic examples for the supervised coordination disambiguation task. We demonstrate that our method produces promising coordination instances that provide gains for the task in low-resource settings. | Coordination Generation via Synchronized Text-Infilling |
d252819180 | In this paper, we explore the relation between gestures and language. Using a multimodal dataset, consisting of TED talks where the language is aligned with the gestures made by the speakers, we adapt a semi-supervised multimodal model to learn gesture embeddings. We show that gestures are predictive of the native language of the speaker, and that gesture embeddings further improve language prediction result. In addition, gesture embeddings might contain some linguistic information, as we show by probing embeddings for psycholinguistic categories. Finally, we analyze the words that lead to the most expressive gestures and find that function words drive the expressiveness of gestures. Our code is available at https://github.com/ MichiganNLP/gestures-language. | Towards Understanding the Relation between Gestures and Language |
d7786274 | This paper describes the work on automated Information Extraction that accepts arbitrary text and extracts information from the text. A new approach to implement Information Extraction system is proposed in this paper. Firstly, the article will be decomposed according to paragraph, sentence and phrase. Every sentence will be compared with the knowledge node, and then append the information extracted to the knowledge model. Finally, the answers are generated to the questions about the input text. With the experimental corpus the accuracy rate of knowledge matching is 63.5%, and accuracy rate of question answering is 65.0% with the system knowledge model. | Improving Information Extraction using Knowledge Model |
d28384449 | ||
d237353191 | Dialogue summarization has drawn much attention recently. Especially in the customer service domain, agents could use dialogue summaries to help boost their works by quickly knowing customer's issues and service progress. These applications require summaries to contain the perspective of a single speaker and have a clear topic flow structure, while neither are available in existing datasets. Therefore, in this paper, we introduce a novel Chinese dataset for Customer Service Dialogue Summarization (CSDS). CSDS improves the abstractive summaries in two aspects: (1) In addition to the overall summary for the whole dialogue, role-oriented summaries are also provided to acquire different speakers' viewpoints. (2) All the summaries sum up each topic separately, thus containing the topic-level structure of the dialogue. We define tasks in CSDS as generating the overall summary and different role-oriented summaries for a given dialogue. Next, we compare various summarization methods on CSDS, and experiment results show that existing methods are prone to generate redundant and incoherent summaries. Besides, the performance becomes much worse when analyzing the performance on role-oriented summaries and topic structures. We hope that this study could benchmark Chinese dialogue summarization and benefit further studies. | CSDS: A Fine-Grained Chinese Dataset for Customer Service Dialogue Summarization |
d219179245 | As part of the NLP Scholar project, we created a single unified dataset of NLP papers and their meta-information (including citation numbers), by extracting and aligning information from the ACL Anthology and Google Scholar. In this paper, we describe several interconnected interactive visualizations (dashboards) that present various aspects of the data. Clicking on an item within a visualization or entering query terms in the search boxes filters the data in all visualizations in the dashboard. This allows users to search for papers in the area of their interest, published within specific time periods, published by specified authors, etc. The interactive visualizations presented here, and the associated dataset of papers mapped to citations, have additional uses as well including understanding how the field is growing (both overall and across sub-areas), as well as quantifying the impact of different types of papers on subsequent publications. | NLP Scholar: An Interactive Visual Explorer for Natural Language Processing Literature |
d222310161 | Neural language models are often trained with maximum likelihood estimation (MLE), where the next word is generated conditioned on the ground-truth word tokens.During testing, however, the model is instead conditioned on previously generated tokens, resulting in what is termed exposure bias.To reduce this gap between training and testing, we propose using optimal transport (OT) to match the sequences generated in these two modes.An extension is further proposed to improve the OT learning, based on the structural and contextual information of the text sequences.The effectiveness of the proposed method is validated on machine translation, text summarization, and text generation tasks. | Improving Text Generation with Student-Forcing Optimal Transport |
d155100205 | In this work, we propose to model the interaction between visual and textual features for multi-modal neural machine translation (MMT) through a latent variable model. This latent variable can be seen as a multi-modal stochastic embedding of an image and its description in a foreign language. It is used in a target-language decoder and also to predict image features. Importantly, our model formulation utilises visual and textual inputs during training but does not require that images be available at test time. We show that our latent variable MMT formulation improves considerably over strong baselines, including a multi-task learning approach (Elliott and Kádár, 2017) and a conditional variational auto-encoder approach(Toyama et al., 2016). Finally, we show improvements due to (i) predicting image features in addition to only conditioning on them, (ii) imposing a constraint on the KL term to promote models with nonnegligible mutual information between inputs and latent variable, and (iii) by training on additional target-language image descriptions (i.e. synthetic data). | Latent Variable Model for Multi-modal Translation |
d237532539 | We present a simple and effective pretraining strategy -bidirectional training (BiT) for neural machine translation. Specifically, we bidirectionally update the model parameters at the early stage and then tune the model normally. To achieve bidirectional updating, we simply reconstruct the training samples from "src→tgt" to "src+tgt→tgt+src" without any complicated model modifications. Notably, our approach does not increase any parameters or training steps, requiring the parallel data merely. Experimental results show that BiT pushes the SOTA neural machine translation performance across 15 translation tasks on 8 language pairs (data sizes range from 160K to 38M) significantly higher. Encouragingly, our proposed model can complement existing data manipulation strategies, i.e. back translation, data distillation and data diversification. Extensive analyses show that our approach functions as a novel bilingual code-switcher, obtaining better bilingual alignment. | Improving Neural Machine Translation by Bidirectional Training |
d238215458 | We propose a novel multitask learning method for diacritization which trains a model to both diacritize and translate. Our method addresses data sparsity by exploiting large, readily avail able bitext corpora. Furthermore, transla tion requires implicit linguistic and seman tic knowledge, which is helpful for resolving ambiguities in diacritization. We apply our method to the Penn Arabic Treebank and re port a new stateoftheart word error rate of 4.79%. We also conduct manual and automatic analysis to better understand our method and highlight some of the remaining challenges in diacritization. Our method has applications in texttospeech, speechtospeech translation, and other NLP tasks. | Improving Arabic Diacritization by Learning to Diacritize and Translate |
d8131267 | Corpus-based Natural Language Processing (NLP) tasks for such popular languages as English, French, etc. have been well studied with satisfactory achievements. In contrast, corpus-based NLP tasks for unpopular languages (e.g. Vietnamese) are at a deadlock due to absence of annotated training data for these languages. Furthermore, hand-annotation of even reasonably well-determined features such as part-ofspeech (POS) tags has proved to be labor intensive and costly. In this paper, we suggest a solution to partially overcome the annotated resource shortage in Vietnamese by building a POS-tagger for an automatically word-aligned English-Vietnamese parallel Corpus (named EVC). This POS-tagger made use of the Transformation-Based Learning (or TBL) method to bootstrap the POS-annotation results of the English POS-tagger by exploiting the POS-information of the corresponding Vietnamese words via their wordalignments in EVC. Then, we directly project POSannotations from English side to Vietnamese via available word alignments. This POS-annotated Vietnamese corpus will be manually corrected to become an annotated training data for Vietnamese NLP tasks such as POS-tagger, Phrase-Chunker, Parser, Word-Sense Disambiguator, etc. | POS-Tagger for English-Vietnamese Bilingual Corpus |
d222142647 | Cross-lingual text classification alleviates the need for manually labeled documents in a target language by leveraging labeled documents from other languages. Existing approaches for transferring supervision across languages require expensive cross-lingual resources, such as parallel corpora, while less expensive cross-lingual representation learning approaches train classifiers without target labeled documents. In this work, we propose a cross-lingual teacher-student method, CLTS, that generates "weak" supervision in the target language using minimal cross-lingual resources, in the form of a small number of word translations. Given a limited translation budget, CLTS extracts and transfers only the most important task-specific seed words across languages and initializes a teacher classifier based on the translated seed words. Then, CLTS iteratively trains a more powerful student that also exploits the context of the seed words in unlabeled target documents and outperforms the teacher. CLTS is simple and surprisingly effective in 18 diverse languages: by transferring just 20 seed words, even a bag-of-words logistic regression student outperforms state-of-theart cross-lingual methods (e.g., based on multilingual BERT). Moreover, CLTS can accommodate any type of student classifier: leveraging a monolingual BERT student leads to further improvements and outperforms even more expensive approaches by up to 12% in accuracy. Finally, CLTS addresses emerging tasks in low-resource languages using just a small number of word translations. | Cross-Lingual Text Classification with Minimal Resources by Transferring a Sparse Teacher |
d237454564 | Temporal knowledge graph (TKG) reasoning is a crucial task that has gained increasing research interest in recent years. Most existing methods focus on reasoning at past timestamps to complete the missing facts, and there are only a few works of reasoning on known TKGs to forecast future facts. Compared with the completion task, the forecasting task is more difficult and faces two main challenges:(1) how to effectively model the time information to handle future timestamps? (2) how to make inductive inference to handle previously unseen entities that emerge over time? To address these challenges, we propose the first reinforcement learning method for forecasting. Specifically, the agent travels on historical knowledge graph snapshots to search for the answer. Our method defines a relative time encoding function to capture the timespan information, and we design a novel time-shaped reward based on Dirichlet distribution to guide the model learning. Furthermore, we propose a novel representation method for unseen entities to improve the inductive inference ability of the model. We evaluate our method for this link prediction task at future timestamps. Extensive experiments on four benchmark datasets demonstrate substantial performance improvement meanwhile with higher explainability, less calculation, and fewer parameters when compared with existing stateof-the-art methods. | TimeTraveler: Reinforcement Learning for Temporal Knowledge Graph Forecasting |
d235293883 | Inferring social relations from dialogues is vital for building emotionally intelligent robots to interpret human language better and act accordingly. We model the social network as an And-or Graph, named SocAoG, for the consistency of relations among a group and leveraging attributes as inference cues. Moreover, we formulate a sequential structure prediction task, and propose an α-β-γ strategy to incrementally parse SocAoG for the dynamic inference upon any incoming utterance: (i) an α process predicting attributes and relations conditioned on the semantics of dialogues, (ii) a β process updating the social relations based on related attributes, and (iii) a γ process updating individual's attributes based on interpersonal social relations. Empirical results on Di-alogRE and MovieGraph show that our model infers social relations more accurately than the state-of-the-art methods. Moreover, the ablation study shows the three processes complement each other, and the case study demonstrates the dynamic relational inference. 1 | SocAoG: Incremental Graph Parsing for Social Relation Inference in Dialogues |
d201633883 | Shipibo-Konibo is a low-resource language from Peru with prior results in statistical machine translation; however, it is challenging to enhance them mainly due to the expensiveness of building more parallel corpora. Thus, we aim for a continuous improvement framework of the Spanish-Shipibo-Konibo language-pair by taking advantage of more advanced strategies and crowd-sourcing. Besides the introduction of a new domain for translation based on language learning flashcards, our main contributions are the extension of the machine translation experiments for Shipibo-Konibo to neural architectures with transfer and active learning; and the building of a conversational agent prototype to retrieve new translations through a social media platform. | A Continuous Improvement Framework of Machine Translation for Shipibo-Konibo |
d5856074 | Given a parallel parsed corpus, statistical treeto-tree alignment attempts to match nodes in the syntactic trees for a given sentence in two languages. We train a probabilistic tree transduction model on a large automatically parsed Chinese-English corpus, and evaluate results against human-annotated word level alignments. We find that a constituent-based model performs better than a similar probability model trained on the same trees converted to a dependency representation. | Dependencies vs. Constituents for Tree-Based Alignment |
d235624320 | Open Knowledge Graphs (OpenKG) refer to a set of (head noun phrase, relation phrase, tail noun phrase) triples such as (tesla, return to, new york) extracted from a corpus using Ope-nIE tools. While OpenKGs are easy to bootstrap for a domain, they are very sparse and far from being directly usable in an end task. Therefore, the task of predicting new facts, i.e., link prediction, becomes an important step while using these graphs in downstream tasks such as text comprehension, question answering, and web search query recommendation. Learning embeddings for OpenKGs is one approach for link prediction that has received some attention lately. However, on careful examination, we found that current OpenKG link prediction algorithms often predict noun phrases (NPs) with incompatible types for given noun and relation phrases. We address this problem in this work and propose OKGIT that improves OpenKG link prediction using novel type compatibility score and type regularization. With extensive experiments on multiple datasets, we show that the proposed method achieves state-of-the-art performance while producing type compatible NPs in the link prediction task. | OKGIT: Open Knowledge Graph Link Prediction with Implicit Types |
d239009473 | Recent work has raised concerns about the inherent limitations of text-only pretraining. In this paper, we first demonstrate that reporting bias, the tendency of people to not state the obvious, is one of the causes of this limitation, and then investigate to what extent multimodal training can mitigate this issue. To accomplish this, we 1) generate the Color Dataset (CoDa), a dataset of human-perceived color distributions for 521 common objects; 2) use CoDa to analyze and compare the color distribution found in text, the distribution captured by language models, and a human's perception of color; and 3) investigate the performance differences between text-only and multimodal models on CoDa. Our results show that the distribution of colors that a language model recovers correlates more strongly with the inaccurate distribution found in text than with the ground-truth, supporting the claim that reporting bias negatively impacts and inherently limits text-only training. We then demonstrate that multimodal models can leverage their visual training to mitigate these effects, providing a promising avenue for future research. 2 We calculate this number using version 3 from February 2020. | The World of an Octopus: How Reporting Bias Influences a Language Model's Perception of Color |
d253098671 | Pre-trained language models (PLMs) have shown their effectiveness in multiple scenarios. However, KBQA remains challenging, especially regarding coverage and generalization settings. This is due to two main factors: i) understanding the semantics of both questions and relevant knowledge from the KB; ii) generating executable logical forms with both semantic and syntactic correctness. In this paper, we present a new KBQA model, TIARA, which addresses those issues by applying multi-grained retrieval to help the PLM focus on the most relevant KB contexts, viz., entities, exemplary logical forms, and schema items. Moreover, constrained decoding is used to control the output space and reduce generation errors. Experiments over important benchmarks demonstrate the effectiveness of our approach. TIARA outperforms previous SOTA, including those using PLMs or oracle entity annotations, by at least 4.1 and 1.1 F1 points on GrailQA and WebQuestionsSP, respectively. Specifically on GrailQA, TIARA outperforms previous models in all categories, with an improvement of 4.7 F1 points in zeroshot generalization. 1Question what napa county wine is 13.9 percent alcohol by volume?(AND wine.wine (AND (JOIN (R wine.wine_sub_region.wines) m.0l2l_) (JOIN wine.wine.percentage_alcohol 13.9^^float)))Schema Retrieval Entity RetrievalMention Detection napa county Candidate Generation (napa valley, m.0l2l_) (napa airport, m.0dlb8x) Entity Disambiguation m.0l2l_ m.0dlb8x Exemplary Logical Form Retrieval (AND wine.wine (JOIN wine.wine.percent_new_oak 13.9^^float)) (AND wine.wine (JOIN wine.wine.wine_sub_region m.0l2l_)) … Class wine.wine wine.wine_type wine.vineyard wine.wine_region food.beverage wine.wine_color … Relation wine.wine.percentage_alcohol wine.wine_region wine.wine_region.wines wine.wine_sub_region.wines wine.wine_wine_sub_region wine.wine_country … Class Trie Relation Trie Decoding Constraints PLMFigure 1: Overview of TIARA. 1) Entity retrieval grounds the mention to entity m.0l2l_. 2) Exemplary logical form retrieval enumerates logical forms starting from the entity m.0l2l_ or the number 13.9, and ranks them. 3) Schema retrieval independently grounds the most related schema items. 4) Retrieved multi-grained contexts are then fed to the PLM for generation. 5) Constrained decoding controls the schema search space during logical form generation.ReferencesNikita Bhutani, Xinyi Zheng, and H. V. Jagadish. 2019.Learning to answer complex questions over knowledge bases with query composition. . 2008. Freebase: a collaboratively created graph database for structuring human knowledge. In | TIARA: Multi-grained Retrieval for Robust Question Answering over Large Knowledge Bases |
d253098848 | Crises such as the COVID-19 pandemic continuously threaten our world and emotionally affect billions of people worldwide in distinct ways. Understanding the triggers leading to people's emotions is of crucial importance. Social media posts can be a good source of such analysis, yet these texts tend to be charged with multiple emotions, with triggers scattering across multiple sentences. This paper takes a novel angle, namely, emotion detection and trigger summarization, aiming to both detect perceived emotions in text, and summarize events and their appraisals that trigger each emotion. To support this goal, we introduce COVIDET (Emotions and their Triggers during Covid-19), a dataset of~1, 900 English Reddit posts related to COVID-19, which contains manual annotations of perceived emotions and abstractive summaries of their triggers described in the post. We develop strong baselines to jointly detect emotions and summarize emotion triggers. Our analyses show that COVIDET presents new challenges in emotion-specific summarization, as well as multi-emotion detection in long social media posts. * Hongli Zhan and Tiberiu Sosea contributed equally.Reddit Post1: My sibling is 19 and she constantly goes places with her friends and to there houses and its honestly stressing me out. 2: Our grandfather lives with us and he has dementia along with other health issues and my mom has diabetes and heart problems and I have autoimmune diseases & chronic health issues. 3: She also has asthma. 4: Its stressing me out because despite this she seems to not care about how badly it would affect all of us if we were to get the virus. 5: And sadly I feel like its not much I can do she literally doesn't respect my mom and though I'm older she doesn't respect me either. 6: Its so frustrating.Emotions and Abstractive Summaries of TriggersEmotion: anger Abstractive Summary of Trigger: My sister having absolutely no regard for any of our family's health coupled with the fact that I can't do anything about it is so aggravating to me.Emotion: fear Abstractive Summary of Trigger: My sibling, who, in spite of our family's myriad of issues that all make us high-risk people, continuously goes out and about, which makes her likely to get infected. I am scared for all of us right now. | Why Do You Feel This Way? Summarizing Triggers of Emotions in Social Media Posts |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.