_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d236486164 | ||
d3011971 | Understanding the connotation of words plays an important role in interpreting subtle shades of sentiment beyond denotative or surface meaning of text, as seemingly objective statements often allude nuanced sentiment of the writer, and even purposefully conjure emotion from the readers' minds. The focus of this paper is drawing nuanced, connotative sentiments from even those words that are objective on the surface, such as "intelligence", "human", and "cheesecake". We propose induction algorithms encoding a diverse set of linguistic insights (semantic prosody, distributional similarity, semantic parallelism of coordination) and prior knowledge drawn from lexical resources, resulting in the first broad-coverage connotation lexicon. POSITIVE NEGATIVE | Connotation Lexicon: A Dash of Sentiment Beneath the Surface Meaning |
d2425693 | The primary goal of our effort is the development of robust and portable language processin g capabilities for information extraction applications. The system under evaluation here is based on language processing components that have demonstrated strong performance capabilities in previous evaluation s [Lehnert et al . 1992a] . Having demonstrated the general viability of these techniques, we are no w concentrating on the practicality of our technology by creating trainable system components to replac e hand-coded data and manually-engineered software.Our general strategy is to automate the construction of domain-specific dictionaries and other languagerelated resources so that information extraction can be customized for specific applications with a minimal amount of human assistance. We employ a hybrid system architecture that combines selective concept extraction [Lehnert 1991] technologies developed at UMass with trainable classifier technologies develope d at Hughes [Dolan et al . 1991] . Our MUC-5 system incorporates seven trainable language components to handle (1) lexical recognition and part-of-speech tagging, (2) knowledge of semantic/syntactic interactions ,(3) semantic feature tagging, (4) noun phrase analysis, (5) limited coreference resolution, (6) domain objec t recognition, and (7) relational link recognition . Our trainable components have been developed so domai n experts who have no background in natural language or machine learning can train individual syste m components in the space of a few hours.Many critical aspects of a complete information extraction are not appropriate for customization o r trainable knowledge acquisition . For example, our system uses low-level text specialists designed t o recognize dates, locations, revenue objects, and other common constructions that involve knowledge o f conventional language . Resources of this type are portable across domains (although not all domains require all specialists) and should be developed as shamble language resources. The UMass/Hughes focus has been on other aspects of information extraction that can benefit from corpus-based knowledge acquisition . For example, in any given information extraction application, some sentences are more important than others , and within a single sentence some phrases are more important than others. When a dictionary is customized for a specific application, vocabulary coverage can be sensitive to the fact that a lot of words contribut e little or no information to the final extraction task : full dictionary coverage is not needed for informatio n extraction applications.In this paper we will overview our hybrid architecture and trainable system components . We will look at examples taken from our official test runs, discuss the test results obtained in our official and optional test runs, and identify promising opportunities for additional research . | UMASS/HUGHES : DESCRIPTION OF THE CIRCUS SYSTE M USED FOR MUC-5 1 INTRODUCTIO N |
d7374530 | In this study, a learning device based on the PATtree data structures was developed. The original PAT-trees were enhanced with the deletion function to emulate human learning competence. The learning process worked as follows. The linguistic patterns from the text corpus are inserted into the PAT-tree one by one. Since the memory was limited, hopefully, the important and new patterns would be retained in the PAT-tree and the old and unimportant patterns would be released from the tree automatically. The proposed PAT-trees with the deletion function have the following advantages. 1) They are easy to construct and maintain. 2) Any prefix substring and its frequency count through PAT-tree can be searched very quickly. 3) The space requirement for a PAT-tree is linear with respect to the size of the input text. 4) The insertion of a new element can be carried out at any time without being blocked by the memory constraints because the free space is released through the deletion of unimportant elements.Experiments on learning high frequency bigrams were carried out under different memory size constraints. High recall rates were achieved. The results show that the proposed PAT-trees can be used as on-line learning devices. | PAT-Trees with the Deletion Function as the Learning Device for Linguistic Patterns |
d219309806 | ||
d11887118 | This paper explores how to automatically generate cross-language links between resources in large document collections. The paper presents new methods for Cross-Lingual Link Discovery (CLLD) based on Explicit Semantic Analysis (ESA). The methods are applicable to any multilingual document collection. In this report, we present their comparative study on the Wikipedia corpus and provide new insights into the evaluation of link discovery systems. In particular, we measure the agreement of human annotators in linking articles in different language versions of Wikipedia, and compare it to the results achieved by the presented methods. | Using Explicit Semantic Analysis for Cross-Lingual Link Discovery Conference or Workshop Item How to cite: Using Explicit Semantic Analysis for Cross-Lingual Link Discovery |
d1686688 | This paper reports a critical analysis of the ISO TimeML standard, in the light of several experiences of temporal annotation that were conducted on spoken French. It shows that the norm suffers from weaknesses that should be corrected to fit a larger variety of needs in NLP and in corpus linguistics. We present our proposition of some improvements of the norm before it will be revised by the ISO Committee in 2017. These modifications concern mainly (1) Enrichments of well identified features of the norm: temporal function of TIMEX time expressions, additional types for TLINK temporal relations; (2) Deeper modifications concerning the units or features annotated: clarification between time and tense for EVENT units, coherence of representation between temporal signals (the SIGNAL unit) and TIMEX modifiers (the MOD feature); (3) A recommendation to perform temporal annotation on top of a syntactic (rather than lexical) layer (temporal annotation on a treebank). | Covering various Needs in Temporal Annotation: a Proposal of Extension of ISO TimeML that Preserves Upward Compatibility |
d259145191 | Voice assistants help users make phone calls, send messages, create events, navigate and do a lot more. However assistants have limited capacity to understand their users' context. In this work, we aim to take a step in this direction. Our work dives into a new experience for users to refer to phone numbers, addresses, email addresses, urls, and dates on their phone screens. Our focus lies in reference understanding, which becomes particularly interesting when multiple similar texts are present on screen, similar to visual grounding. We collect a dataset and propose a lightweight general purpose model for this novel experience. Due to the high cost of consuming pixels directly, our system is designed to rely on the extracted text from the UI. Our model is modular, thus offering flexibility, improved interpretability, and efficient runtime memory utilization. | Referring to Screen Texts with Voice Assistants |
d1345645 | Cross-lingual information retrieval (CLIR) involving the Chinese language has been thoroughly studied in the general language domain, but rarely in the biomedical domain, due to the lack of suitable linguistic resources and parsing tools. In this paper, we describe a Chinese-English CLIR system for biomedical literature, which exploits a bilingual ontology, the "eCMeSH Tree". This is an extension of the Chinese Medical Subject Headings (CMeSH) Tree, based on Medical Subject Headings (MeSH). Using the 2006 and 2007 TREC Genomics track data, we have evaluated the performance of the eCMeSH Tree in expanding queries. We have compared our results to those obtained using two other approaches, i.e. pseudo-relevance feedback (PRF) and document translation (DT). Subsequently, we evaluate the performance of different combinations of these three retrieval methods. Our results show that our method of expanding queries using the eCMeSH Tree can outperform the PRF method. Furthermore, combining this method with PRF and DT helps to smooth the differences in query expansion, and consequently results in the best performance amongst all experiments reported. All experiments compare the use of two different retrieval models, i.e. Okapi BM25 and a query likelihood language model. In general, the former performs slightly better. | Biomedical Chinese-English CLIR Using an Extended CMeSH Resource to Expand Queries |
d252819479 | We explore the task of generating long-form technical questions from textbooks. Semistructured metadata of a textbook -the table of contents and the index -provide rich cues for technical question generation. Existing literature for long-form question generation focuses mostly on reading comprehension assessment, and does not use semi-structured metadata for question generation. We design unsupervised template based algorithms for generating questions based on structural and contextual patterns in the index and ToC. We evaluate our approach on textbooks on diverse subjects and show that our approach generates high quality questions of diverse types. We show that, in comparison, zero-shot question generation using pre-trained LLMs on the same meta-data has much poorer quality. | Unsupervised Generation of Long-form Technical Questions from Textbook Metadata using Structured Templates |
d7170202 | We describe a parser used in the CoNLL 2006 Shared Task, "Multingual Dependency Parsing." The parser first identifies syntactic dependencies and then labels those dependencies using a maximum entropy classifier. We consider the impact of feature engineering and the choice of machine learning algorithm, with particular focus on Slovene, Spanish and Swedish. | Dependency Parsing with Reference to Slovene, Spanish and Swedish |
d2610731 | We report on the construction of the Webis text reuse corpus 2012 for advanced research on text reuse. The corpus compiles manually written documents obtained from a completely controlled, yet representative environment that emulates the web. Each of the 297 documents in the corpus is about one of the 150 topics used at the TREC Web Tracks 2009-2011, thus forming a strong connection with existing evaluation efforts. Writers, hired at the crowdsourcing platform oDesk, had to retrieve sources for a given topic and to reuse text from what they found. Part of the corpus are detailed interaction logs that consistently cover the search for sources as well as the creation of documents. This will allow for in-depth analyses of how text is composed if a writer is at liberty to reuse texts from a third party-a setting which has not been studied so far. In addition, the corpus provides an original resource for the evaluation of text reuse and plagiarism detectors, where currently only less realistic resources are employed. | Crowdsourcing Interaction Logs to Understand Text Reuse from the Web |
d53584863 | We introduce the contemporary Amharic corpus, which is automatically tagged for morphosyntactic information. Texts are collected from 25,199 documents from different domains and about 24 million orthographic words are tokenized. Since it is partly a web corpus, we made some automatic spelling error corrections. We have also modified the existing morphological analyzer, HornMorpho, to use it for automatic tagging. | Contemporary Amharic Corpus: Automatically Morpho-Syntactically Tagged Amharic Corpus |
d14834892 | A maximum entropy (ME) model is usually estimated so that it conforms to equality constraints on feature expectations. However, the equality constraint is inappropriate for sparse and therefore unreliable features. This study explores an ME model with box-type inequality constraints, where the equality can be violated to reflect this unreliability. We evaluate the inequality ME model using text categorization datasets. We also propose an extension of the inequality ME model, which results in a natural integration with the Gaussian MAP estimation. Experimental results demonstrate the advantage of the inequality models and the proposed extension. | Evaluation and Extension of Maximum Entropy Models with Inequality Constraints |
d227230663 | Rumors are manufactured with no respect for accuracy, but can circulate quickly and widely by "word-of-post" through social media conversations. Conversation tree encodes important information indicative of the credibility of rumor. Existing conversation-based techniques for rumor detection either just strictly follow tree edges or treat all the posts fully-connected during feature learning. In this paper, we propose a novel detection model based on tree transformer to better utilize user interactions in the dialogue where post-level self-attention plays the key role for aggregating the intra-/inter-subtree stances. Experimental results on the TWITTER and PHEME datasets show that the proposed approach consistently improves rumor detection performance. | Debunking Rumors on Twitter with Tree Transformer |
d190014935 | ||
d239020528 | In a typical TM-MT environment, translations for segments are provided from TMs based on matching criteria while the remaining segments are translated completely by MT. However, this binary approach does not always produce desirable translations. For example, even though a contiguous portion of a sentence to be translated may exactly match a TM entry or a frequently occurring sub-segment in many TM entries, if the match for the entire sentence does not exceed some arbitrary threshold, the smaller matches will not be used, and the entire sentence will be machine translated, resulting in a less than perfect translation, even for those portions that could have matched perfectly. In this report, we describe our approach to flexibly combine the capability of MT and TMs, applying exact TM matches to sub-segments of sentences and allowing MT to handle the remaining portions of the sentences. We specifically focus on the scenario where the matched phrases, clauses, and/or sentences are quotations in the text to be translated. | Enhancing a Production TM-MT Environment Using a Quotation TM |
d6320263 | This document describes a tool which extracts term and lexicon entries from SMT phrase tables, without further reference to monolingual data. It applies filters to such tables, and builds lexicon entries from the 'good' candidates. Error rates of the tool can be as low as 7.3%, accumulated from source, target, and transfer errors. 1 | Creating Term and Lexicon Entries from Phrase Tables |
d17002200 | This paper claims that reliance on discourse focus to guide the production of rhetorically structured texts is insufficient over lengthier stretches of prose. Instead, this paper argues that at least three distinct attentional constraints are required: discourse focus [Sidner, 1979[Sidner, , 1983Grosz and Sidner, 1986], temporal focus [Webber, 1988], and a novel notion of spatial focus. The paper illustrates the operation of this tripartite theory of focus in a computational system (TEXPLAN) that plans multisentential text. | Using Discourse Focus, Temporal Focus, and Spatial to Generate Multisentential Text Focus |
d3111932 | Transition-based dependency parsing systems can utilize rich feature representations. However, in practice, features are generally limited to combinations of lexical tokens and part-of-speech tags. In this paper, we investigate richer features based on supertags, which represent lexical templates extracted from dependency structure annotated corpus. First, we develop two types of supertags that encode information about head position and dependency relations in different levels of granularity. Then, we propose a transition-based dependency parser that incorporates the predictions from a CRF-based supertagger as new features. On standard English Penn Treebank corpus, we show that our supertag features achieve parsing improvements of 1.3% in unlabeled attachment, 2.07% root attachment, and 3.94% in complete tree accuracy. | Improving Dependency Parsers with Supertags |
d13558449 | This paper presents an overview of the broad-coverage, application-independent natural language generation component of the NLP system being developed at Microsoft Research. It demonstrates how this component functions within a multilingual Machine Translation system (MSR-MT), using the languages that we are currently working on (English, Spanish, Japanese, and Chinese). Section 1 provides a system description of MSR-MT. Section 2 focuses on the generation component and its set of core rules. Section 3 describes an additional layer of generation rules with examples that address issues specific to MT. Section 4 presents evaluation results in the context of MSR-MT. Section 5 addresses generation issues outside of MT. | Generation for Multilingual MT |
d17850709 | In this paper, we describe how NLP can semi-automate the construction and analysis of knowledge in Eunomos, a legal knowledge management service which enables users to view legislation from various sources and find the right definitions and explanations of legal concepts in a given context. NLP can semi-automate some routine tasks currently performed by knowledge engineers, such as classifying norms, or linking key terms within legislation to ontological concepts. This helps overcome the resource bottleneck problem of creating specialist knowledge management systems. While accuracy is of the utmost importance in the legal domain, and the information should be verified by domain experts as a matter of course, a semi-automated approach can result in considerable efficiency gains. | NLP Challenges for Eunomos, a Tool to Build and Manage Legal Knowledge |
d201641506 | This paper describes the neural machine translation (NMT) systems of the LIUM Laboratory developed for the French ↔ German news translation task of the Fourth Conference on Machine Translation (WMT 2019). The chosen language pair is included for the first time in the WMT news translation task. We describe how the training and the evaluation data was created. We also present our participation in the French ↔ German translation directions using self-attentional Transformer networks with small and big architectures. | LIUM's Contributions to the WMT2019 News Translation Task: Data and Systems for German↔French Language Pairs |
d253734994 | Over the last years, software development in domains with high security demands transitioned from traditional methodologies to uniting modern approaches from software development and operations (DevOps). Key principles of DevOps gained more importance and are now applied to security aspects of software development, resulting in the automation of security-enhancing activities. In particular, it is common practice to use automated security testing tools that generate reports after inspecting a software artifact from multiple perspectives. However, this raises the challenge of generating duplicate security findings. To identify these duplicate findings manually, a security expert has to invest resources like time, effort, and knowledge. A partial automation of this process could reduce the analysis effort, encourage DevOps principles, and diminish the chance of human error. In this study, we investigated the potential of applying Natural Language Processing for clustering semantically similar security findings to support the identification of problem-specific duplicate findings. Towards this goal, we developed a web application for annotating and assessing security testing tool reports and published a human-annotated corpus of clustered security findings. In addition, we performed a comparison of different semantic similarity techniques for automatically grouping security findings. Finally, we assess the resulting clusters using both quantitative and qualitative evaluation methods. | Semantic Similarity-Based Clustering of Findings From Security Testing Tools |
d52139452 | This paper presents several experiments on constructing Indonesian-Korean StatisticalMachine Translation (SMT) system. A parallel corpus containing around 40,000 segments on each side has been developed for training the baseline SMT system that is built based on n-gram language model and the phrase-based translation table model. This system still has several problems, including non-translated phrases, mistranslation, incorrect phrase orders, and remaining Korean particles in the target language. To overcome these problems, some techniques are employed i.e. POS (part-of-speech) tag model, POS-based reordering rules, multiple steps translation, additional post-process, and their combinations. We then test the SMT system by randomly extracting segments from the parallel corpus. In general, the additional techniques lead to better performance in terms of BLEU score compared to the baseline systemRelated WorkParallel corpus is a valuable component needed in SMT to train models, optimize the model parameters, and test the translation quality. However, a good parallel corpus of low-resource languages such as Indonesian and Korean is hard to obtain. Therefore, we do not only use books as the source for constructing, but also subtitles and Bible. Automatic parallel corpus extraction from movie subtitles has been introduced in(Caroline et al., 2007). From this study, it was reported that 37,625 287 | Rule-based Reordering and Post-Processing for Indonesian-Korean Statistical Machine Translation |
d1237618 | Decoding algorithms for syntax based machine translation suffer from high computational complexity, a consequence of intersecting a language model with a context free grammar. Left-to-right decoding, which generates the target string in order, can improve decoding efficiency by simplifying the language model evaluation. This paper presents a novel left to right decoding algorithm for tree-to-string translation, using a bottom-up parsing strategy and dynamic future cost estimation for each partial translation. Our method outperforms previously published tree-to-string decoders, including a competing left-to-right method. | Left-to-Right Tree-to-String Decoding with Prediction |
d260063179 | This paper describes the approach that we used to take part in the multi-label multiclass emotion classification as Track 3 of the WASSA 2023 Empathy, Emotion and Personality Shared Task at ACL 2023. The overall goal of this track is to build models that can predict 8 classes (7 emotions + neutral) based on short English essays written in response to news article that talked about events perceived as harmful to people. We used OpenAI generative pretrained transformers with full-scale APIs for the emotion prediction task by finetuning a GPT-3 model and doing prompt engineering for zero-shot / few-shot learning with ChatGPT and GPT-4 models based on multiple experiments on the dev set. The most efficient method was fine-tuning a GPT-3 model which allowed us to beat our baseline characterbased XGBoost Classifier and rank 2nd among all other participants by achieving a macro F1 score of 0.65 and a micro F1 score of 0.7 on the final blind test set. | Team Bias Busters at WASSA 2023 Empathy, Emotion and Personality Shared Task: Emotion Detection with Generative Pretrained Transformers |
d59913540 | Les techniques de résumé automatique multi-documents par extraction ont récemment évolué vers des méthodes statistiques pour la sélection des phrases à extraire. Dans cet article, nous présentons un système conforme à l'« état de l'art » -CBSEAS -que nous avons développé pour les tâches Opinion (résumés d'opinions issues de blogs) et Update (résumés de dépêches et mise à jour du résumé à partir de nouvelles dépêches sur le même événement) de la campagne d'évaluation TAC 2008, et montrons l'intérêt d'analyses structurelles et linguistiques des documents à résumer. Nous présentons également notre étude sur la structure des dépêches et l'impact de son intégration à CBSEAS.Abstract. Automatic multi-document summarization techniques have recently evolved into statistical methods for selecting the sentences that will be used to generate the summary. In this paper, we present a system in accordance with « State-of-the-art » -CBSEAS -that we have developped for the « Opinion Task » (automatic summaries of opinions from blogs) and the « Update Task » (automatic summaries of newswire articles and information update) of the TAC 2008 evaluation campaign, and show the interest of structural and linguistic analysis of the documents to summarize . We also present our study on news structure and its integration to CBSEAS impact.Mots-clés : Résumé automatique, structure de documents. | Une approche mixte-statistique et structurelle -pour le résumé automatique de dépêches |
d252624518 | In this article, we present an exploratory study on perceived word sense difficulty by native and non-native speakers of French. We use a graded lexicon in conjunction with the French Wiktionary to generate tasks in bundles of four items. Annotators manually rate the difficulty of the word senses based on their usage in a sentence by selecting the easiest and the most difficult word sense out of four. Our results show that the native and non-native speakers largely agree when it comes to the difficulty of words. Further, the rankings derived from the manual annotation broadly follow the levels of the words in the graded resource, although these levels were not overtly available to annotators. Using clustering, we investigate whether there is a link between the complexity of a definition and the difficulty of the associated word sense. However, results were inconclusive. The annotated data set will be made available for research purposes. | A Dictionary-Based Study of Word Sense Difficulty |
d53235606 | Embedding and projection matrices are commonly used in neural language models (NLM) as well as in other sequence processing networks that operate on large vocabularies. We examine such matrices in fine-tuned language models and observe that a NLM learns word vectors whose norms are related to the word frequencies. We show that by initializing the weight norms with scaled log word counts, together with other techniques, lower perplexities can be obtained in early epochs of training. We also introduce a weight norm regularization loss term, whose hyperparameters are tuned via a grid search. With this method, we are able to significantly improve perplexities on two word-level language modeling tasks (without dynamic evaluation): from 54.44 to 53.16 on Penn Treebank (PTB) and from 61.45 to 60.13 on WikiText-2 (WT2). | Improving Neural Language Models with Weight Norm Initialization and Regularization |
d29217299 | In this paper, a new HMM structure is proposed to work with a limited training corpus in order to obtain improved synthetic-speech fluency. Spectral fluency is improved because this HMM structure can model the context-dependent spectral characteristics of a speech unit. In addition, instead of using a decision tree to cluster contexts, the knowledge of phoneme articulation is based to cluster contexts and reduce the enormous quantity of context combinations. To evaluate the proposed HMM structure, we construct three Mandarin speech synthesis systems each uses one different HMM structure for comparisons. In these systems, the prosodic parameters are all generated with same ANN modules studied previously * 國立臺灣科技大學資訊工程系 but the spectral coefficients are generated with different HMM adopted by its corresponding system. As to the synthesis of signal waveform, the signal model, harmonic plus noise model (HNM), studied previously is commonly adopted in the three systems. According to the results of listening tests, the speech synthesized by the system using the proposed HMM structure is indeed more fluent than the speeches synthesized by the other two systems. In addition, average spectral distances are measured between recorded sentences and synthetic sentences. The results show that the HMM structure proposed here also obtains smaller average spectral distance than the other two HMM structures. | 基於發音知識以建構頻譜 HMM 之國語語音合成方法 A Mandarin Speech Synthesis Method Using Articulation-knowledge Based Spectral HMM Structure |
d18322992 | ||
d14538687 | ||
d18292719 | We present several methods for stemming and lemmatization based on discriminative string transduction. We exploit the paradigmatic regularity of semi-structured inflection tables to identify stems in an unsupervised manner with over 85% accuracy. Experiments on English, Dutch and German show that our stemmers substantially outperform Snowball and Morfessor, and approach the accuracy of a supervised model. Furthermore, the generated stems are more consistent than those annotated by experts. Our direct lemmatization model is more accurate than Morfette and Lemming on most datasets. Finally, we test our methods on the data from the shared task on morphological reinflection. | Leveraging Inflection Tables for Stemming and Lemmatization |
d6618252 | In this paper we present a novel featureenriched approach that learns to detect the conversation focus of threaded discussions by combining NLP analysis and IR techniques. Using the graph-based algorithm HITS, we integrate different features such as lexical similarity, poster trustworthiness, and speech act analysis of human conversations with featureoriented link generation functions. It is the first quantitative study to analyze human conversation focus in the context of online discussions that takes into account heterogeneous sources of evidence. Experimental results using a threaded discussion corpus from an undergraduate class show that it achieves significant performance improvements compared with the baseline system. | Learning to Detect Conversation Focus of Threaded Discussions |
d11047443 | Shared Task 1 at SemEval-2017 deals with assessing the semantic similarity between sentences, either in the same or in different languages. In our system submission, we employ multilingual word representations, in which similar words in different languages are close to one another. Using such representations is advantageous, since the increasing amount of available parallel data allows for the application of such methods to many of the languages in the world. Hence, semantic similarity can be inferred even for languages for which no annotated data exists. Our system is trained and evaluated on all language pairs included in the shared task (English, Spanish, Arabic, and Turkish). Although development results are promising, our system does not yield high performance on the shared task test sets. | ResSim at SemEval-2017 Task 1: Multilingual Word Representations for Semantic Textual Similarity |
d52054464 | State-of-the-art systems in deep question answering proceed as follows: (1) an initial document retrieval selects relevant documents, which (2) are then processed by a neural network in order to extract the final answer. Yet the exact interplay between both components is poorly understood, especially concerning the number of candidate documents that should be retrieved. We show that choosing a static number of documents -as used in prior research -suffers from a noiseinformation trade-off and yields suboptimal results. As a remedy, we propose an adaptive document retrieval model. This learns the optimal candidate number for document retrieval, conditional on the size of the corpus and the query. We report extensive experimental results showing that our adaptive approach outperforms state-of-the-art methods on multiple benchmark datasets, as well as in the context of corpora with variable sizes. | Adaptive Document Retrieval for Deep Question Answering |
d14068705 | We examine a key task in biomedical text processing, normalization of disorder mentions. We present a multi-pass sieve approach to this task, which has the advantage of simplicity and modularity. Our approach is evaluated on two datasets, one comprising clinical reports and the other comprising biomedical abstracts, achieving state-of-the-art results. | Sieve-Based Entity Linking for the Biomedical Domain |
d6893994 | This paper discusses some improvements in recent and planned versions of the multimodal annotation tool ELAN, which are targeted at improving the usability of annotated files. Increased support for multilingual documents is provided, by allowing for multilingual vocabularies and by specifying a language per document, annotation layer (tier) or annotation. In addition, improvements in the search possibilities and the display of the results have been implemented, which are especially relevant in the interpretation of the results of complex multi-tier searches. | Improving the exploitation of linguistic annotations in ELAN |
d227230279 | Emotion recognition in conversations (ERC) has received much attention recently in the natural language processing community. Considering that the emotions of the utterances in conversations are interactive, previous works usually implicitly model the emotion interaction between utterances by modeling dialogue context, but the misleading emotion information from context often interferes with the emotion interaction. We noticed that the gold emotion labels of the context utterances can provide explicit and accurate emotion interaction, but it is impossible to input gold labels at inference time. To address this problem, we propose an iterative emotion interaction network, which uses iteratively predicted emotion labels instead of gold emotion labels to explicitly model the emotion interaction. This approach solves the above problem, and can effectively retain the performance advantages of explicit modeling. We conduct experiments on two datasets, and our approach achieves state-of-the-art performance. * Email corresponding. This work is licensed under a Creative Commons Attribution 4.0 International Licence. Licence details: http:// creativecommons.org/licenses/by/4.0/. | An Iterative Emotion Interaction Network for Emotion Recognition in Conversations |
d6908311 | We describe a number of experiments carried out to address the problem of creating summaries from multiple sources in multiple languages. A centroid-based sentence extraction system has been developed which decides the content of the summary using texts in different languages and uses sentences from English sources alone to create the final output. We describe the evaluation of the system in the recent Multilingual Summarization Evaluation MSE 2005 using the pyramids and ROUGE methods. | Multilingual Multidocument Summarization Tools and Evaluation |
d232021505 | ||
d6454720 | We propose a new unsupervised method for topic detection that automatically identifies the different facets of an event. We use pointwise Kullback-Leibler divergence along with the Jaccard coefficient to build a topic graph which represents the community structure of the different facets. The problem is formulated as a weighted set cover problem with dynamically varying weights. The algorithm is domainindependent and generates a representative set of informative and discriminative phrases that cover the entire event. We evaluate this algorithm on a large collection of blog postings about different news events and report promising results. | Detecting multiple facets of an event using graph-based unsupervised methods |
d218973694 | ||
d6097881 | We present a new version of QUEST -an open source framework for machine translation quality estimation -which brings a number of improvements: (i) it provides a Web interface and functionalities such that non-expert users, e.g. translators or lay-users of machine translations, can get quality predictions (or internal features of the framework) for translations without having to install the toolkit, obtain resources or build prediction models; (ii) it significantly improves over the previous runtime performance by keeping resources (such as language models) in memory; (iii) it provides an option for users to submit the source text only and automatically obtain translations from Bing Translator; (iv) it provides a ranking of multiple translations submitted by users for each source text according to their estimated quality. We exemplify the use of this new version through some experiments with the framework. | An efficient and user-friendly tool for machine translation quality estimation |
d40301297 | 以 以 以 以部落格 部落格 部落格 部落格語料進行 語料進行 語料進行 語料進行情緒趨勢分析 情緒趨勢分析 情緒趨勢分析 情緒趨勢分析 楊昌樺 高虹安 陳信希 國立台灣大學資訊工程學系 | |
d18796136 | Prepositional phrase (PP) attachment is a well known challenge to parsing. In this paper, we combine the insights of different works, namely: (1) treating PP attachment as a classification task with an arbitrary number of attachment candidates;(2) using auxiliary distributions to augment the data beyond the hand-annotated training set; (3) using topological fields to get information about the distribution of PP attachment throughout clauses and (4) using state-of-the-art techniques such as word embeddings and neural networks. We show that jointly using these techniques leads to substantial improvements. We also conduct a qualitative analysis to gauge where the ceiling of the task is in a realistic setup. | PP Attachment: Where do We Stand? |
d8936396 | We show that using confidence-weighted classification in transition-based parsing gives results comparable to using SVMs with faster training and parsing time. We also compare with other online learning algorithms and investigate the effect of pruning features when using confidenceweighted classification. | Transition-based parsing with Confidence-Weighted Classification |
d7534990 | Belief tracking is a promising technique for adding robustness to spoken dialog systems, but current research is fractured across different teams, techniques, and domains. This paper amplifies past informal discussions (Raux, 2011) to call for a belief tracking challenge task, based on the Spoken dialog challenge corpus(Black et al., 2011). Benefits, limitations, evaluation design issues, and next steps are presented. | A belief tracking challenge task for spoken dialog systems |
d235127372 | ||
d2473928 | We propose a method of automatically constructing an English-Chinese bilingual Fra-meNet where the English FrameNet lexical entries are linked to the appropriate Chinese word senses. This resource can be used in machine translation and cross-lingual IR systems. We coerce the English FrameNet into Chinese using a bilingual lexicon, frame context in FrameNet and taxonomy structure in HowNet. Our approach does not require any manual mapping between FrameNet and HowNet semantic roles. Evaluation results show that we achieve a promising 82% average Fmeasure for the most ambiguous lexical entries. | Automatic Construction of an English-Chinese Bilingual FrameNet |
d16387995 | Theoretically, an improvement in a language model occurs as the size of the n-grams increases from 3 to 5 or higher. As the n-gram size increases, the number of parameters and calculations, and the storage requirement increase very rapidly if we attempt to store all possible combinations of n-grams. To avoid these problems, the reduced n-grams' approach previously developed by O' Boyle and Smith [1993] can be applied. A reduced n-gram language model, called a reduced model, can efficiently store an entire corpus's phrase-history length within feasible storage limits. Another advantage of reduced n-grams is that they usually are semantically complete. In our experiments, the reduced n-gram creation method or the O' Boyle-Smith reduced n-gram algorithm was applied to a large Chinese corpus. The Chinese reduced n-gram Zipf curves are presented here and compared with previously obtained conventional Chinese n-grams. The Chinese reduced model reduced perplexity by 8.74% and the language model size by a factor of 11.49. This paper is the first attempt to model Chinese reduced n-grams, and may provide important insights for Chinese linguistic research. | Reduced N-Grams for Chinese Evaluation |
d7600459 | ADVERBS AND SEMANTIC INFERENCES | |
d9781962 | Incrementality as a way of managing the interactions between a dialogue system and its users has been shown to have concrete advantages over the traditional turn-taking frame. Incremental systems are more reactive, more human-like, offer a better user experience and allow the user to correct errors faster, hence avoiding desynchronisations. Several incremental models have been proposed, however, their core underlying architecture is different from the classical dialogue systems. As a result, they have to be implemented from scratch. In this paper, we propose a method to transform traditional dialogue systems into incremental ones. A new module, called the Scheduler is inserted between the client and the service so that from the client's point of view, the system behaves incrementally, even though the service does not. | An easy method to make dialogue systems incremental |
d16652144 | A great deal of research effort has been expended in support of natural language (NL) database querying.English and English-like query systems already exist, such as ROBOT[Ha77], TQA[Da78], LUNAREWo76] and those described by Kaplan[Ka79], Walker[Wa78] and Waltz[Wa75].Little effort has gone to NL database update [KD81, Br81, Sk80, CHSI]. We want to extend l~ interaction to include data modification (insert, delete, modify) rather than simply data extraction.The desirability and unavailability of NL database modification has been noted by Wiederhold, et al[WiSl].Database systems currently do not contain structures for explicit modelling of real world changes. | NATNRAL LANGUAGE UPDATES* |
d6895264 | This paper addresses the problem of scientific research analysis. We use the topic model Latent Dirichlet Allocation [2] and a novel classifier to classify research papers based on topic and language. Moreover, we show various insightful statistics and correlations within and across three research fields: Linguistics, Computational Linguistics, and Education. In particular, we show how topics change over time within each field, what relations and influences exist between topics within and across fields, as well as what trends can be established for some of the world's natural languages. Finally, we talk about trend prediction and topic suggestion as future extensions of this research. | Topic Modeling of Research Fields: An Interdisciplinary Perspective |
d53080615 | Sequence encoders are crucial components in many neural architectures for learning to read and comprehend. This paper presents a new compositional encoder for reading comprehension (RC). Our proposed encoder is not only aimed at being fast but also expressive. Specifically, the key novelty behind our encoder is that it explicitly models across multiple granularities using a new dilated composition mechanism. In our approach, gating functions are learned by modeling relationships and reasoning over multi-granular sequence information, enabling compositional learning that is aware of both long and short term information. We conduct experiments on three RC datasets, showing that our proposed encoder demonstrates very promising results both as a standalone encoder as well as a complementary building block. Empirical results show that simple Bi-Attentive architectures augmented with our proposed encoder not only achieves state-of-the-art / highly competitive results but is also considerably faster than other published works. | Multi-Granular Sequence Encoding via Dilated Compositional Units for Reading Comprehension |
d30668669 | This work presents a novel approach to Automatic Post-Editing (APE) and Word-Level Quality Estimation (QE) using ensembles of specialized Neural Machine Translation (NMT) systems. Word-level features that have proven effective for QE are included as input factors, expanding the representation of the original source and the machine translation hypothesis, which are used to generate an automatically post-edited hypothesis. We train a suite of NMT models that use different input representations, but share the same output space. These models are then ensembled together, and tuned for both the APE and the QE task. We thus attempt to connect the state-of-the-art approaches to APE and QE within a single framework. Our models achieve state-of-the-art results in both tasks, with the only difference in the tuning step which learns weights for each component of the ensemble. | Ensembling Factored Neural Machine Translation Models for Automatic Post-Editing and Quality Estimation |
d10031888 | This paper explores methods to alleviate the effect of lexical sparseness in the classification of verbal arguments. We show how automatically generated selectional preferences are able to generalize and perform better than lexical features in a large dataset for semantic role classification. The best results are obtained with a novel second-order distributional similarity measure, and the positive effect is specially relevant for out-of-domain data. Our findings suggest that selectional preferences have potential for improving a full system for Semantic Role Labeling. | Generalizing over Lexical Features: Selectional Preferences for Semantic Role Classification |
d3224392 | This article describes a method to build semantic representations of composite expressions in a compositional way by using WordNet relations to represent the meaning of words. The meaning of a target word is modelled as a vector in which its semantically related words are assigned weights according to both the type of the relationship and the distance to the target word. Word vectors are compositionally combined by syntactic dependencies. Each syntactic dependency triggers two complementary compositional functions: the named head function and dependent function. The experiments show that the proposed compositional method performs as the state-of-the-art for subjectverb expressions, and clearly outperforms the best system for transitive subject-verbobject constructions. | Compositional Semantics using Feature-Based Models from WordNet |
d11684721 | Speech recognition errors are inevitable in a speech dialog system. This paper presents an error handling method based on correction grammars which recognize the correction utterances which follow a recognition error. Correction grammars are dynamically created from existing grammars and a set of correction templates. We also describe a prototype dialog system which incorporates this error handling method, and provide empirical evidence that this method can improve dialog success rate and reduce the number of dialog turns required for error recovery. | Correction Grammars for Error Handling in a Speech Dialog System |
d218462190 | In this paper we describe the research that was carried out and the resources that were developed within the DISCO (Development and Integration of Speech technology into COurseware for language learning) project. This project aimed at developing an ASR-based CALL system that automatically detects pronunciation and grammar errors in Dutch L2 speaking and generates appropriate, detailed feedback on the errors detected. We briefly introduce the DISCO system and present its design, architecture and speech recognition modules. We then describe a first evaluation of the complete DISCO system and present some results. The resources generated through DISCO are subsequently described together with possible ways of efficiently generating additional resources in the future. | The DISCO ASR-based CALL system: practicing L2 oral skills and beyond |
d235258255 | ||
d184483084 | In this study we deal with the problem of identifying and categorizing offensive language in social media. Our group, BNU-HKBU UIC NLP Team2, use supervised classification along with multiple version of data generated by different ways of pre-processing the data. We then use the state-of-the-art model Bidirectional Encoder Representations from Transformers, or BERT (Devlin et al.(2018)), to capture linguistic, syntactic and semantic features. Long range dependencies between each part of a sentence can be captured by BERT's bidirectional encoder representations. Our results show 85.12% accuracy and 80.57% F1 scores in Subtask A (offensive language identification), 87.92% accuracy and 50% F1 scores in Subtask B (categorization of offense types), and 69.95% accuracy and 50.47% F1 score in Subtask C (offense target identification). Analysis of the results shows that distinguishing between targeted and untargeted offensive language is not a simple task. More work needs to be done on the unbalance data problem in Subtasks B and C. Some future work is also discussed. | BNU-HKBU UIC NLP Team 2 at SemEval-2019 Task 6: Detecting Offensive Language Using BERT model |
d219300220 | ||
d62569305 | BBN's PLUM Probabilistic Language Understanding System The PLUM System Group* | |
d8145155 | We extended the work ofLow, Ng, and Guo (2005)to create a Chinese word segmentation system based upon a maximum entropy statistical model. This system was entered into the Third International Chinese Language Processing Bakeoff and evaluated on all four corpora in their respective open tracks. Our system achieved the highest F-score for the UPUC corpus, and the second, third, and seventh highest for CKIP, CITYU, and MSRA respectively. Later testing with the gold-standard data revealed that while the additions we made to Low et al.'s system helped our results for the 2005 data with which we experimented during development, a number of them actually hurt our scores for this year's corpora. | Maximum Entropy Word Segmentation of Chinese Text |
d9787275 | The UNL project (Universal Networking Language) proposes a standard for encoding the meaning of natural language utterances as semantic hypergraphs, intended to be used as pivot in multilingual information and communication systems. Several deconverters permit to automatically translate UNL utterances into natural languages. However, a rough enconvertion from natural language texts to UNL expressions is usually done interactively with editors specially designed for the UNL project or by hand (which is very time-consuming and difficult to extrapolate to huge amounts of data). In this paper, we address the issue of using an existing incremental robust parser as main resource to enconverting French utterances into UNL expressions. | Using an incremental robust parser to automatically generate semantic UNL graphs |
d30403883 | ||
d218974528 | This paper introduces work carried out for the automatic generation of a written text in Italian starting from glosses of a fable in Italian Sign Language (LIS). The paper gives a brief overview of sign languages (SLs) and some peculiarities of SL fables such as the use of space, the strategy of Role Shift and classifiers. It also presents the annotation of the fable "The Tortoise and the Hare" -signed in LIS and made available by Alba Cooperativa Sociale -, which was annotated manually by first author for her master's thesis. The annotation was the starting point of a generation process that allowed us to automatically generate a text in Italian starting from LIS glosses. LIS sentences have been transcribed with Italian words into tables on simultaneous layers, each of which contains specific linguistic or non-linguistic pieces of information. In addition, the present work discusses problems encountered in the annotation and generation process. | Annotating a Fable in Italian Sign Language (LIS) |
d1449890 | In this paper, we present some preliminary results on Statistical Machine Translation from Bulgarian-to-English and English-to-Bulgarian. Linguistic knowledge has been added gradually as factors in the MOSES system. | Factored Models for Deep Machine Translation |
d7502134 | When individuals lose the ability to produce their own speech, due to degenerative diseases such as motor neurone disease (MND) or Parkinson's, they lose not only a functional means of communication but also a display of their individual and group identity. In order to build personalized synthetic voices, attempts have been made to capture the voice before it is lost, using a process known as voice banking. But, for some patients, the speech deterioration frequently coincides or quickly follows diagnosis. Using HMM-based speech synthesis, it is now possible to build personalized synthetic voices with minimal data recordings and even disordered speech. The power of this approach is that it is possible to use the patient's recordings to adapt existing voice models pre-trained on many speakers. When the speech has begun to deteriorate, the adapted voice model can be further modified in order to compensate for the disordered characteristics found in the patient's speech, we call this process "voice repair". In this paper we compare two methods of voice repair. The first method follows a trial and error approach and requires the expertise of a speech therapist. The second method is entirely automatic and based on some a priori statistical knowledge. A subjective evaluation shows that the automatic method achieves similar results than the manually controlled method. | A Comparison of Manual and Automatic Voice Repair for Individual with Vocal Disabilities |
d577399 | We present and partially evaluate procedures for the extraction of noun+verb collocation candidates from German text corpora, along with their morphosyntactic preferences, especially for the active vs. passive voice. We start from tokenized, tagged, lemmatized and chunked text, and we use extraction patterns formulated in the CQP corpus query language. We discuss the results of a precision evaluation, on administrative texts from the European Union: we find a considerable amount of specialized collocations, as well as general ones and complex predicates; overall the precision is considerably higher than that of a statistical extractor used as a baseline. | Tools for collocation extraction: preferences for active vs. passive |
d5665391 | Long distance reordering remains one of the greatest challenges in statistical machine translation research as the key contextual information may well be beyond the confine of translation units. In this paper, we propose Two-Neighbor Orientation (TNO) model that jointly models the orientation decisions between anchors and two neighboring multi-unit chunks which may cross phrase or rule boundaries. We explicitly model the longest span of such chunks, referred to as Maximal Orientation Span, to serve as a global parameter that constrains underlying local decisions. We integrate our proposed model into a state-of-the-art string-to-dependency translation system and demonstrate the efficacy of our proposal in a large-scale Chinese-to-English translation task. On NIST MT08 set, our most advanced model brings around +2.0 BLEU and -1.0 TER improvement. | Two-Neighbor Orientation Model with Cross-Boundary Global Contexts |
d198993887 | We present a unique dataset of student sourcebased argument essays to facilitate research on the relations between content, argumentation skills, and assessment. Two classroom writing assignments were given to college students in a STEM major, accompanied by a carefully designed rubric. The paper presents a reliability study of the rubric, showing it to be highly reliable, and initial annotation on content and argumentation annotation of the essays. | Rubric Reliability and Annotation of Content and Argument in Source-Based Argument Essays |
d246702341 | Summarisation of reviews aims at compressing opinions expressed in multiple review documents into a concise form while still covering the key opinions. Despite the advancement in summarisation models, evaluation metrics for opinionated text summaries lag behind and still rely on lexical-matching metrics such as ROUGE. In this paper, we propose using the question-answering(QA) approach to evaluate summaries of opinions in reviews. We propose to identify opinion-bearing text spans in the reference summary to generate QA pairs so as to capture salient opinions. A QA model is then employed to probe the candidate summary to evaluate information overlap between the candidate and reference summaries. We show that our metric RunQA, Review Summary Evaluation via Question Answering, correlates well with human judgments in terms of coverage and focus of information. | Evaluation of Review Summaries via Question-Answering |
d6479514 | 摘要 伴隨著網際網路快速發展與多媒體資訊的大量增長，影音的傳遞與瀏覽越來越多並且成 為我們日常生活的重要活動，這使得關於語音文件檢索(Spoken Document Retrieval, SDR)的研究成為一個有魅力的研究主題[1][2][3][4]。一般而言，SDR 的研究主要可分 成兩大研究方向：第一個研究方向為建立具強健性的索引(Robust Indexing)以表達語音 文件中詞彙和語意內涵，並且減緩語音辨識錯誤所造成的影響；第二個研究方向為發展 有效的檢索模型(Effective Retrieval Models)，基於索引所代表的詞彙和語意內涵來量化 使用者輸入的查詢(Query)和語音文件的相似程度，以協助使用者找到相關資訊，可分 | 使用查詢意向探索與類神經網路於語音文件檢索之研究 Exploring Query Intent and Neural Network modeling Techniques for Spoken Document Retrieval |
d201638186 | ||
d3837667 | This paper describes a character-based Chinese word sense induction (WSI) system for the International Chinese Language Processing Bakeoff 2010. By computing the longest common substrings between any two contexts of the ambiguous word, our system extracts collocations as features and does not depend on any extra tools, such as Chinese word segmenters. We also design a constrained clustering algorithm for this task. Experiemental results show that our system could achieve 69.88 scores of FScore on the development data set of SIGHAN Bakeoff 2010. | NEUNLPLab Chinese Word Sense Induction System for SIGHAN Bakeoff 2010 |
d15586646 | We present a general framework to incorporate prior knowledge such as heuristics or linguistic features in statistical generative word alignment models. Prior knowledge plays a role of probabilistic soft constraints between bilingual word pairs that shall be used to guide word alignment model training. We investigate knowledge that can be derived automatically from entropy principle and bilingual latent semantic analysis and show how they can be applied to improve translation performance. | Guiding Statistical Word Alignment Models With Prior Knowledge |
d256461392 | Propaganda is information or ideas that an organised group or government spreads to influence peopleś opinions, especially by not giving all the facts or secretly emphasising only one way of looking at the points. The ability to automatically detect propaganda-related language is a challenging task that researchers in the NLP community have recently started to address. This paper presents the participation of our team AraBEM in the propaganda detection shared task on Arabic tweets. Our system utilised a pre-trained BERT model to perform multi-class binary classification. It attained the best score at 0.602 micro-f1, ranking third on subtask-1, which identifies the propaganda techniques as a multilabel classification problem with a baseline of 0.079. | AraBEM at WANLP 2022 Shared Task: Propaganda Detection in Arabic Tweets |
d198231742 | In this exploratory study, we attempt to automatically induce PDTB-style relations from RST trees. We work with a German corpus of news commentary articles, annotated for RST trees and explicit PDTB-style relations and we focus on inducing the implicit relations in an automated way. Preliminary results look promising as a high-precision (but lowrecall) way of finding implicit relations where no shallow structure is annotated at all, but mapping proves more difficult in cases where EDUs and relation arguments overlap, yet do not seem to signal the same relation. | Toward Cross-theory Discourse Relation Annotation |
d18436064 | Web search can be enhanced in powerful ways if token spans in Web text are annotated with disambiguated entities from large catalogs like Freebase. Entity annotators need to be trained on sample mention snippets. Wikipedia entities and annotated pages offer high-quality labeled data for training and evaluation. Unfortunately, Wikipedia features only one-ninth the number of entities as Freebase, and these are a highly biased sample of well-connected, frequently mentioned "head" entities. To bring hope to "tail" entities, we broaden our goal to a second task: assigning types to entities in Freebase but not Wikipedia. The two tasks are synergistic: knowing the types of unfamiliar entities helps disambiguate mentions, and words in mention contexts help assign types to entities. We present TMI, a bipartite graphical model for joint type-mention inference. TMI attempts no schema integration or entity resolution, but exploits the above-mentioned synergy. In experiments involving 780,000 people in Wikipedia, 2.3 million people in Freebase, 700 million Web pages, and over 20 professional editors, TMI shows considerable annotation accuracy improvement (e.g., 70%) compared to baselines (e.g., 46%), especially for "tail" and emerging entities. We also compare with Google's recent annotations of the same corpus with Freebase entities, and report considerable improvements within the people domain. | Joint Bootstrapping of Corpus Annotations and Entity Types |
d31107411 | Social media collect and spread on the Web personal opinions, facts, fake news and all kind of information users may be interested in. Applying argument mining methods to such heterogeneous data sources is a challenging open research issue, in particular considering the peculiarities of the language used to write textual messages on social media. In addition, new issues emerge when dealing with arguments posted on such platforms, such as the need to make a distinction between personal opinions and actual facts, and to detect the source disseminating information about such facts to allow for provenance verification. In this paper, we apply supervised classification to identify arguments on Twitter, and we present two new tasks for argument mining, namely facts recognition and source identification. We study the feasibility of the approaches proposed to address these tasks on a set of tweets related to the Grexit and Brexit news topics. | Argument Mining on Twitter: Arguments, Facts and Sources |
d252619979 | Warning: This paper has contents which may be offensive, or upsetting however this cannot be avoided owing to the nature of the work. | Detecting Unintended Social Bias in Toxic Language Datasets |
d18715949 | Supervised machine learning models for automated essay scoring (AES) usually require substantial task-specific training data in order to make accurate predictions for a particular writing task. This limitation hinders their utility, and consequently their deployment in real-world settings. In this paper, we overcome this shortcoming using a constrained multi-task pairwisepreference learning approach that enables the data from multiple tasks to be combined effectively.Furthermore, contrary to some recent research, we show that high performance AES systems can be built with little or no task-specific training data. We perform a detailed study of our approach on a publicly available dataset in scenarios where we have varying amounts of task-specific training data and in scenarios where the number of tasks increases. | Constrained Multi-Task Learning for Automated Essay Scoring |
d2723528 | We present a novel evaluation method for grammatical error correction that addresses problems with previous approaches and scores systems in terms of improvement on the original text. Our method evaluates corrections at the token level using a globally optimal alignment between the source, a system hypothesis, and a reference. Unlike the M 2 Scorer, our method provides scores for both detection and correction and is sensitive to different types of edit operations. | Towards a standard evaluation method for grammatical error detection and correction |
d6932025 | Bootstrapping has a tendency, called semantic drift, to select instances unrelated to the seed instances as the iteration proceeds. We demonstrate the semantic drift of bootstrapping has the same root as the topic drift of Kleinberg's HITS, using a simplified graphbased reformulation of bootstrapping. We confirm that two graph-based algorithms, the von Neumann kernels and the regularized Laplacian, can reduce semantic drift in the task of word sense disambiguation (WSD) on Senseval-3 English Lexical Sample Task. Proposed algorithms achieve superior performance to Espresso and previous graph-based WSD methods, even though the proposed algorithms have less parameters and are easy to calibrate. | Graph-based Analysis of Semantic Drift in Espresso-like Bootstrapping Algorithms |
d6955865 | In this paper we address the problem of solving substitution ciphers using a beam search approach. We present a conceptually consistent and easy to implement method that improves the current state of the art for decipherment of substitution ciphers and is able to use high order n-gram language models. We show experiments with 1:1 substitution ciphers in which the guaranteed optimal solution for 3-gram language models has 38.6% decipherment error, while our approach achieves 4.13% decipherment error in a fraction of time by using a 6-gram language model. We also apply our approach to the famous Zodiac-408 cipher and obtain slightly better (and near to optimal) results than previously published. Unlike the previous state-of-the-art approach that uses additional word lists to evaluate possible decipherments, our approach only uses a letterbased 6-gram language model. Furthermore we use our algorithm to solve large vocabulary substitution ciphers and improve the best published decipherment error rate based on the Gigaword corpus of 7.8% to 6.0% error rate. | Beam Search for Solving Substitution Ciphers |
d227231106 | ||
d10934368 | We present new language resources for Moroccan and Sanaani Yemeni Arabic. The resources include corpora for each dialect which have been morphologically annotated, and morphological analyzers for each dialect which are derived from these corpora. These are the first sets of resources for Moroccan and Yemeni Arabic. The resources will be made available to the public. | Morphologically Annotated Corpora and Morphological Analyzers for Moroccan and Sanaani Yemeni Arabic |
d14395270 | This paper describes the LDC forced aligner which is designed to align audio and transcripts. Unlike existing forced aligners, LDC forced aligner can align partially transcribed audio files, and also audio files with large chunks of non-speech segments, such as noise, music, silence etc, by inserting optional wildcard phoneme sequences between sentence or paragraph boundaries. Based on the HTK tool kit, LDC forced aligner can align audio and transcript on sentence or word level. This paper also reports its usage on English and Mandarin Chinese data. | LDC Forced Aligner |
d232021525 | ||
d252624536 | The annotation and automatic recognition of non-fictional discourse within a text is an important, yet unresolved task in literary research. While non-fictional passages can consist of several clauses or sentences, we argue that 1) an entity-level classification of fictionality and 2) the linking of Wikidata identifiers can be used to automatically identify (non-)fictional discourse. We query Wikidata and DBpedia for relevant information about a requested entity as well as the corresponding literary text to determine the entity's fictionality status and assign a Wikidata identifier, if unequivocally possible. We evaluate our methods on an exemplary text from our diachronic literary corpus, where our methods classify 97% of persons and 62% of locations correctly as fictional or real. Furthermore, 75% of the resolved persons and 43% of the resolved locations are resolved correctly. In a quantitative experiment, we apply the entity-level fictionality tagger to our corpus and conclude that more non-fictional passages can be identified when information about real entities is available. | Levels of Non-Fictionality in Fictional Texts |
d7929514 | We treat paraphrase identification as an ordering task. We construct a corpus of 250 sets of five sentences, with each set containing a reference sentence and four paraphrase candidates, which are annotated on a scale of 1 to 5 for paraphrase proximity. We partition this corpus into 1000 pairs of sentences in which the first is the reference sentence and the second is a paraphrase candidate. We then train a DNN encoder for sentence pair inputs. It consists of parallel CNNs that feed parallel LSTM RNNs, followed by fully connected NNs, and finally a dense merging layer that produces a single output. We test it for both binary and graded predictions. The latter are generated as a by-product of training the former (the binary classifier). It reaches 70% accuracy on the binary classification task. It achieves a Pearson correlation of .59-.61 with the annotated gold standard for the gradient ranking candidate sets. | Deep Learning of Binary and Gradient Judgements for Semantic Paraphrase |
d10168800 | Both coarse-to-fine and A * parsing use simple grammars to guide search in complex ones. We compare the two approaches in a common, agenda-based framework, demonstrating the tradeoffs and relative strengths of each method. Overall, coarse-to-fine is much faster for moderate levels of search errors, but below a certain threshold A * is superior. In addition, we present the first experiments on hierarchical A * parsing, in which computation of heuristics is itself guided by meta-heuristics. Multi-level hierarchies are helpful in both approaches, but are more effective in the coarseto-fine case because of accumulated slack in A * heuristics. | Hierarchical Search for Parsing |
d2138649 | Past approaches for using reinforcement learning to derive dialog control policies have assumed that there was enough collected data to derive a reliable policy. In this paper we present a methodology for numerically constructing confidence intervals for the expected cumulative reward for a learned policy. These intervals are used to (1) better assess the reliability of the expected cumulative reward, and (2) perform a refined comparison between policies derived from different Markov Decision Processes (MDP) models. We applied this methodology to a prior experiment where the goal was to select the best features to include in the MDP statespace. Our results show that while some of the policies developed in the prior work exhibited very large confidence intervals, the policy developed from the best feature set had a much smaller confidence interval and thus showed very high reliability. | Estimating the Reliability of MDP Policies: A Confidence Interval Approach |
d391968 | This article describes the European TransRouter project which produced a prototype application designed to help users identify the best possible way to translate a particular project. As a typical knowledge society application, it is designed to facilitate easy and almost instant access to knowledge by evaluating highly complex information based on user-definable criteria. The interest shown by potential investors augurs well for the commercial exploitation and further development of this prototype. | The EU LE4 TransRouter Project |
d17173602 | Providing a comparative framework for parsers is a task that has already been tried in the | A Protocol for Evaluating Analyzers of Syntax (PEAS) |
d37606613 | This talk will describe new methods for generating Natural Language in interactive systems -methods which are similar to planning approaches, but which use statistical machine learning to develop adaptive NLG components. Employing statistical models of users, generation contexts, and of Natural Languages themselves, has several potentially beneficial features: the ability to train models on real data, the availability of precise mathematical methods for optimisation, and the capacity to adapt robustly to previously unseen situations. Rather than emulating human behaviour in generation (which can be suboptimal) these methods can even find strategies for NLG which improve upon human performance.Recently, some encouraging results have been obtained with real users of 3 different systems developed using these methods, for the tasks of Information Presentation in an automated tourist guide, Referring Expression Generation in a technical support system, and generation of Temporal Referring Expressions in an appointment scheduling system. The results show that optimised NLG significantly outperforms related prior approaches, and can also improve the global performance of dialogue systems.As well as explaining the core Reinforcement Learning and user modelling methods and concepts behind this work, I will also cover some recent work from other researchers which fits with this general perspective on NLG. Finally, I discuss some future directions for this research area, for example the issues of incremental generation and generation under uncertainty. | Talkin' bout a revolution (statistically speaking) |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.