_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d40901022 | Not everyone understands what a completely rational process this is, this maintenance of a motorcycle. <…> A motorcycle functions entirely in accordance with the laws of reason, and a study of the art of motorcycle maintenance is really a miniature study of the art of rationality itself. Robert M. Pirsig, "Zen and the Art of Motorcycle Maintenance"1 The forthcoming ISO standard on translation services [ISO/TC37/WG6] has not been referred to in this paper. 2 Person or organisation supplying translation services [EN 15038] | Zen and the Art of Quality Assurance Quality Assurance Automation in Translation: Needs, Reality and Expectations |
d18493970 | Sentiment polarity lexicons are key resources for sentiment analysis, and researchers have invested a lot of efforts in their manual creation. However, there has been a recent shift towards automatically extracted lexicons, which are orders of magnitude larger and perform much better. These lexicons are typically mined using bootstrapping, starting from very few seed words whose polarity is given, e.g., 50-60 words, and sometimes even just 5-6. Here we demonstrate that much higher-quality lexicons can be built by starting with hundreds of words and phrases as seeds, especially when they are in-domain. Thus, we combine (i) mid-sized high-quality manually crafted lexicons as seeds and (ii) bootstrapping, in order to build large-scale lexicons. | On the Impact of Seed Words on Sentiment Polarity Lexicon Induction |
d6052486 | The work analyses the head nod, a down-up movement of the head, as a polysemic social signal, that is, a signal with a number of different meanings which all share some common semantic element. Based on the analysis of 100 nods drawn from the SSPNet corpus of TV political debates, a typology of nods is presented that distinguishes Speaker's, Interlocutor's and Third Listener's nods, with their subtypes (confirmation, agreement, approval, submission and permission, greeting and thanks, backchannel giving and backchannel request, emphasis, ironic agreement, literal and rhetoric question, and others). For each nod the analysis specifies: 1. characteristic features of how it is produced, among which main direction, amplitude, velocity and number of repetitions; 2. cues in other modalities, like direction and duration of gaze; 3. conversational context in which the nod typically occurs. For the Interlocutor's or Third Listener's nod, the preceding speech act is relevant: yes/no answer or information for a nod of confirmation, expression of opinion for one of agreement, prosocial action for greetings and thanks; for the Speaker's nods, instead, their meanings are mainly distinguished by accompanying signals. | Types of Nods. The polysemy of a social signal |
d1755973 | Describing troubling events and images and reflecting on their emotional meanings are central components of most psychotherapies. The computer system described here tracks the occurrence and intensity of narration or imagery within transcribed therapy sessions and over the course of treatments; it likewise tracks the extent to which language denoting appraisal and logical thought occurs. The Discourse Attributes Analysis Program (DAAP) is a computer text analysis system that uses several dictionaries, including the Weighted Referential Activity Dictionary (WRAD), designed to detect verbal communication of emotional images and events, and the Reflection Dictionary (REF), designed to detect verbal communication denoting cognitive appraisal, as well as other dictionaries. For each dictionary and each turn of speech, DAAP uses a moving weighted average of dictionary weights, together with a fold-over procedure, to produce a smooth density function that graphically illustrates the rise and fall of each underlying psychological variable. These density functions are then used to produce several new measures, including measures of the vividness of descriptions of images or events, and a measure of the extent to which descriptions of events or images and reflection on their meaning occur separately. | A Computer Program for Tracking the Evolution of a Psychotherapy Treatment |
d16581593 | Studies have shown that Twitter can be used for health surveillance, and personal experience tweets (PETs) are an important source of information for health surveillance. To mine Twitter data requires a relatively balanced corpus and it is challenging to construct such a corpus due to the labor-intensive annotation tasks of large data sets. We developed a bootstrap method of finding PETs with the use of the machine learning-based filter. Through a few iterations, our approach can efficiently improve the balance of two class dataset with a reduced amount of annotation work. To demonstrate the usefulness of our method, a PET corpus related to effects caused by 4 dietary supplements was constructed. In 3 iterations, a corpus of 8,770 tweets was obtained from 108,528 tweets collected, and the imbalance of two classes was significantly reduced from 1:31 to 1:3. In addition, two out of three classifiers used showed improved performance over iterations. It is conceivable that our approach can be applied to various other health surveillance studies that use machine learning-based classifications of imbalanced Twitter data. | Construction of a Personal Experience Tweet Corpus for Health Surveillance |
d227230371 | ||
d8854127 | MT systems typically use parsers to help reorder constituents. However most languages do not have adequate treebank data to learn good parsers, and such training data is extremely time-consuming to annotate. Our earlier work has shown that a reordering model learnt from word-alignments using POS tags as features can improve MT performance(Visweswariah et al., 2011). In this paper, we investigate the effect of word-classing on reordering performance using this model. We show that unsupervised word clusters perform somewhat worse but still reasonably well, compared to a part-of-speech (POS) tagger built with a small amount of annotated data; while a richer tagset including case and gender-number-person further improves reordering performance by around 1.2 monolingual BLEU points. While annotating this richer tagset is more complicated than annotating the base tagset, it is much easier than annotating treebank data. | A Study of Word-Classing for MT Reordering |
d219307047 | ||
d218974018 | ||
d2466688 | In this paper, we describe a novel phrase reordering model based on predicate-argument structure. Our phrase reordering method utilizes a general predicate-argument structure analyzer to reorder source language chunks based on predicate-argument structure. We explicitly model longdistance phrase alignments by reordering arguments and predicates. The reordering approach is applied as a preprocessing step in training phase of a phrase-based statistical MT system. We report experimental results in the evaluation campaign of IWSLT 2006. | Phrase Reordering for Statistical Machine Translation Based on Predicate-Argument Structure |
d11504881 | Methodology and results of complex computational analysis of present-day standard Czech are presented. According to computer programmes various linguistic observations were achieved, concerning especially dependency syntax. | COMPUTATIONAL DATA ANALYSIS FOR SYNTAX |
d1827517 | This paper presents a possibility to extend the fonnalism of linear indexed grammars. The extension is based on the use of tuples of pushdowns instead of one pushdown to store indices during a derivation. If a restriction on the accessibility of the pushdowns is used, it can be shown that the resulting fonnalisms give rise to a hierarchy of languages that is equivalent with a hierarchy defined by Weir. For this equivalence, that was already known for a slightly different fonnalism, this paper gives a new proof. Since all languages of Weir's hierarchy are known to be mildly context sensitive, the proposed extensions of L!Gs become comparable with extensions of tree adjoining grammars and head grammars. | Extending Linear Indexed Grammars |
d52011255 | Neural network based multi-task learning has achieved great success on many NLP problems, which focuses on sharing knowledge among tasks by linking some layers to enhance the performance. However, most existing approaches suffer from the interference between tasks because they lack of selection mechanism for feature sharing. In this way, the feature spaces of tasks may be easily contaminated by useless features borrowed from others, which will confuse the models for making correct prediction. In this paper, we propose a multi-task convolutional neural network with the Leaky Unit, which has memory and forgetting mechanism to filter the feature flows between tasks. Experiments on five different datasets for text classification validate the benefits of our approach.This work is licensed under a Creative Commons Attribution 4.0 International License. License details: http:// creativecommons.org/licenses/by/4.0/ | Learning What to Share: Leaky Multi-Task Network for Text Classification |
d891205 | ON PARSING CONTROL FOR EFFICIENT TEXT ANALYSIS | |
d10401980 | Semantic role labeling (SRL) is crucial to natural language understanding as it identifies the predicate-argument structure in text with semantic labels. Unfortunately, resources required to construct SRL models are expensive to obtain and simply do not exist for most languages. In this paper, we present a two-stage method to enable the construction of SRL models for resourcepoor languages by exploiting monolingual SRL and multilingual parallel data. Experimental results show that our method outperforms existing methods. We use our method to generate Proposition Banks with high to reasonable quality for 7 languages in three language families and release these resources to the research community. | Generating High Quality Proposition Banks for Multilingual Semantic Role Labeling |
d202790046 | Automatically identifying rumours in social media and assessing their veracity is an important task with downstream applications in journalism. A significant challenge is how to keep rumour analysis tools up-to-date as new information becomes available for particular rumours that spread in a social network. This paper presents a novel open-source webbased rumour analysis tool that can continuous learn from journalists. The system features a rumour annotation service that allows journalists to easily provide feedback for a given social media post through a web-based interface. The feedback allows the system to improve an underlying state-of-the-art neural networkbased rumour classification model. The system can be easily integrated as a service into existing tools and platforms used by journalists using a REST API. | Journalist-in-the-Loop: Continuous Learning as a Service for Rumour Analysis |
d232124786 | In this work, we present a phenomenon-oriented comparative analysis of the two dominant approaches in English Resource Semantic (ERS) parsing: classic, knowledge-intensive and neural, data-intensive models. To reflect state-of-the-art neural NLP technologies, a factorizationbased parser is introduced that can produce Elementary Dependency Structures much more Volume 47, Number 1 accurately than previous data-driven parsers. We conduct a suite of tests for different linguistic phenomena to analyze the grammatical competence of different parsers, where we show that, despite comparable performance overall, knowledge-and data-intensive models produce different types of errors, in a way that can be explained by their theoretical properties. This analysis is beneficial to in-depth evaluation of several representative parsing techniques and leads to new directions for parser development. | Comparing Knowledge-Intensive and Data-Intensive Models for English Resource Semantic Parsing under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license Computational Linguistics |
d16661325 | This paper describes our participation in the shared task Named Entity Recognition in Twitter organized as part of the 2nd Workshop on Noisy User-generated Text. The shared task comprises two sub-tasks, concerning a) the detection of the boundaries of entities and b) the classification of the entities into one of 10 possible types. The proposed approach is based on Linked Open Data for extracting rich features along with standard ones which are then used by a learning to search algorithm in order to build the tagger. The submitted system scored 46.16 and 60.24 in terms of F-measure and ranked 2nd and 3rd for the classification and segmentation tasks respectively. | Learning to Search for Recognizing Named Entities in Twitter |
d16946723 | We attempt to identify citations in non-academic text such as patents. Unlike academic articles which often provide bibliographies and follow consistent citation styles, non-academic text cites scientific research in a more ad-hoc manner. We manually annotate citations in 50 patents, train a CRF classifier to find new citations, and apply a reranker to incorporate non-local information. Our best system achieves 0.83 F-score on 5-fold cross validation. | Corpus and Method for Identifying Citations in Non-Academic Text |
d252819355 | While there is much research on cross-domain text classification, most existing approaches focus on one-to-one or many-to-one domain adaptation. In this paper, we tackle the more challenging task of domain generalization, in which domain-invariant representations are learned from multiple source domains, without access to any data from the target domains, and classification decisions are then made on test documents in unseen target domains. We propose a novel framework based on supervised contrastive learning with a memory-saving queue. In this way, we explicitly encourage examples of the same class to be closer and examples of different classes to be further apart in the embedding space. We have conducted extensive experiments on two Amazon review sentiment datasets, and one rumour detection dataset. Experimental results show that our domain generalization method consistently outperforms state-of-the-art domain adaptation methods 1 . | Domain Generalization for Text Classification with Memory-Based Supervised Contrastive Learning |
d219304770 | ||
d14358598 | Deterministic dependency parsing is a robust and efficient approach to syntactic parsing of unrestricted natural language text. In this paper, we analyze its potential for incremental processing and conclude that strict incrementality is not achievable within this framework. However, we also show that it is possible to minimize the number of structures that require nonincremental processing by choosing an optimal parsing algorithm. This claim is substantiated with experimental evidence showing that the algorithm achieves incremental parsing for 68.9% of the input when tested on a random sample of Swedish text. When restricted to sentences that are accepted by the parser, the degree of incrementality increases to 87.9%. | Incrementality in Deterministic Dependency Parsing |
d213518628 | ||
d44320775 | RESUME ____________________________________________________________________________________________________________ La méthode des potentiels évoqués a permis de caractériser différentes composantes associées au traitement de la parole. Cependant il n'existe pas aujourd'hui de marqueur cortical témoignant du succès de l'accès lexical lors de la compréhension de la parole. Le but de cette étude est donc de développer un protocole expérimental et une analyse statistique des signaux électroencéphalographiques, afin d'identifier des clusters tempsfréquence dans l'activité oscillatoire corrélant avec l'intelligibilité de stimuli paroliers. Pour mettre en évidence cet effet, nous avons présenté aux sujets des mots dégradés par noise-vocoding avant et après une courte phase d'apprentissage perceptuel. Nous avons comparé les activités oscillatoires apparaissant en réponse à des stimuli évalués comme « intelligibles » et « inintelligibles » par les participants (N=12). Nous sommes ainsi parvenus à mettre à jour trois activités avec des topologies et des fréquences spécifiques liées au succès de l'accès lexical ABSTRACT _________________________________________________________________________________________________________Oscillatory cortical activity and intelligibility of degraded speechMany neurocognitive aspects associated with the processing of speech were up to now studied by the analysis of event-related potentials. However, none of these cortical responses can be considered as a direct indicator of successful lexical access during speech comprehension. The aim of the present study is to develop an experimental paradigm and a statistical analysis on electrophysiological data, in order to identify timefrequency patterns in the oscillatory cortical activity that correlate with the intelligibility of degraded speech. For this purpose we used noise-vocoded speech that is very difficult to understand without prior exposure. Noise-vocoded words were presented before and after a short period of perceptual learning, and we compared the oscillatory activity following stimuli rated as "intelligible" or "unintelligible" by participants (N=12). Results show that we were able to identify three oscillatory activities with specific topology and latency resulting from a successful lexical access. | Oscillations corticales et intelligibilité de la parole dégradée |
d258378218 | Eye-tracking data in Chinese languages present unique challenges due to the non-alphabetic and unspaced nature of the Chinese writing systems. This paper introduces the first deeply-annotated joint Mandarin-Cantonese eye-tracking dataset, from which we achieve a unified eye-tracking prediction system for both language varieties. In addition to the commonly studied first fixation duration and the total fixation duration, this dataset also includes the second fixation duration, expressing fixation patterns that are more relevant to higher-level, structural processing.A basic comparison of the features and measurements in our dataset revealed variation between Mandarin and Cantonese on fixation patterns related to word class and word position. The test of feature usefulness suggested that traditional features are less powerful in predicting the second-pass fixation, to which the linear distance to root makes a leading contribution in Mandarin. In contrast, Cantonese eye-movement behavior relies more on word position and part of speech. | Comparing and Predicting Eye-tracking Data in Mandarin and Cantonese |
d256460910 | In this paper, I present an approach using one-vs-one classification scheme with TF-IDF term weighting on character n-grams for identifying Arabic dialects used in social media. The scheme was evaluated in the context of the third Nuanced Arabic Dialect Identification (NADI 2022) shared task for identifying Arabic dialects used in Twitter messages. The approach was implemented with logistic regression loss and trained using stochastic gradient decent (SGD) algorithm. This simple method achieved a macro F1 score of 22.89% and 10.83% on TEST A and TEST B, respectively, in comparison to an approach based on AraBERT pretrained transformer model which achieved a macro F1 score of 30.01% and 14.84%, respectively. My submission based on AraBERT scored a macro F1 average of 22.42% and was ranked 10 out of the 19 teams who participated in the task. | SQU-CS @ NADI 2022: Dialectal Arabic Identification using One-vs-One Classification with TF-IDF Weights Computed on Character n-grams |
d256461028 | One of the remarkable characteristics of the drug trafficking lexicon is its elusive nature. In order to communicate information related to drugs or drug trafficking, the community uses several terms that are mostly unknown to regular people, or even to the authorities. For instance, the terms jolly green, joystick, or jive are used to refer to marijuana. The selection of such terms is not necessarily a random or senseless process, but a communicative strategy in which figurative language plays a relevant role. In this study, we describe an ongoing research to identify drug-related terms by applying machine learning techniques. To this end, a data set regarding drug trafficking in Spanish was built. This data set was used to train a word embedding model to identify terms used by the community to creatively refer to drugs and related matters. The initial findings show an interesting repository of terms created to consciously veil drug-related contents by using figurative language devices, such as metaphor or metonymy. These findings can provide preliminary evidence to be applied by law agencies in order to address actions against crime, drug transactions on the internet, illicit activities, or human trafficking. | |
d174799844 | When reading a text, it is common to become stuck on unfamiliar words and phrases, such as polysemous words with novel senses, rarely used idioms, internet slang, or emerging entities. If we humans cannot figure out the meaning of those expressions from the immediate local context, we consult dictionaries for definitions or search documents or the web to find other global context to help in interpretation. Can machines help us do this work? Which type of context is more important for machines to solve the problem? To answer these questions, we undertake a task of describing a given phrase in natural language based on its local and global contexts. To solve this task, we propose a neural description model that consists of two context encoders and a description decoder. In contrast to the existing methods for non-standard English explanation (Ni and Wang, 2017) and definition generation(Noraset et al., 2017;Gadetsky et al., 2018), our model appropriately takes important clues from both local and global contexts. Experimental results on three existing datasets (including WordNet, Oxford and Urban Dictionaries) and a dataset newly created from Wikipedia demonstrate the effectiveness of our method over previous work. | Learning to Describe Unknown Phrases with Local and Global Contexts |
d12677273 | Determining whether two terms in text have an ancestor relation (e.g. Toyota and car) or a sibling relation (e.g. Toyota and Honda) is an essential component of textual inference in NLP applications such as Question Answering, Summarization, and Recognizing Textual Entailment. Significant work has been done on developing stationary knowledge sources that could potentially support these tasks, but these resources often suffer from low coverage, noise, and are inflexible when needed to support terms that are not identical to those placed in them, making their use as general purpose background knowledge resources difficult. In this paper, rather than building a stationary hierarchical structure of terms and relations, we describe a system that, given two terms, determines the taxonomic relation between them using a machine learning-based approach that makes use of existing resources. Moreover, we develop a global constraint optimization inference process and use it to leverage an existing knowledge base also to enforce relational constraints among terms and thus improve the classifier predictions. Our experimental evaluation shows that our approach significantly outperforms other systems built upon existing well-known knowledge sources. | Constraints based Taxonomic Relation Classification |
d227231772 | ||
d220446233 | ||
d174800839 | Supervised approaches to named entity recognition (NER) are largely developed based on the assumption that the training data is fully annotated with named entity information. However, in practice, annotated data can often be imperfect with one typical issue being the training data may contain incomplete annotations. We highlight several pitfalls associated with learning under such a setup in the context of NER and identify limitations associated with existing approaches, proposing a novel yet easy-to-implement approach for recognizing named entities with incomplete data annotations. We demonstrate the effectiveness of our approach through extensive experiments. 1 | Better Modeling of Incomplete Annotations for Named Entity Recognition |
d21687180 | Since the idea of combining Lexicalized Tree Adjoining Grammars (LTAG) and frame semantics was proposed(Kallmeyer and Osswald, 2013), a set of resources of this type has been created. These grammars are composed of pairs of elementary trees and frames, where syntactic and semantic arguments are linked using unification variables. This allows to build semantic representations when parsing, by composing the frames according to the combination of the elementary trees. However, the lack of a parser using such grammars makes it complicated to check whether these resources are correct implementations of the theory or not. The development of larger resources, that is to say large-coverage grammars, is also conditioned by the existence of such a parser. In this paper, we present our solution to this problem, namely an extension of the TuLiPA parser with frame semantics. We also present the frameworks used to build the resources used by the parser: the theoretical framework, composed of LTAG and frame semantics, and the software framework, XMG2. | A Parser for LTAG and Frame Semantics |
d7078111 | Italian is a language presenting a lot of syntactical problems, sucb as a rather unrestricted word order, unbounded agreement controls, long distance structure checkings and so on. Things get worse and worse if we pass from "sentences of linguists" to real texts. In this paper we will present a system able to retrieve and signal syntactic errors in real italian texts.'~ English translations will be word by word In the last chapter we dont provide any translation, since ill formed phrases and sentences exemplifying the coverage are language specific Acr~ DE COLING-92, NANTES, 23-28 AOUT 1992 l 003 | JDIh Parsing Italian with a Robust Constraint Grammar |
d226283978 | ||
d12903863 | Recent work on natural language processing systems is aimed at more conversational, context-adaptive systems in multiple domains. An important requirement for such a system is the automatic detection of the domain and a domain consistency check of the given speech recognition hypotheses. We report a pilot study addressing these tasks, the underlying data collection and investigate the feasibility of annotating the data reliably by human annotators. | Annotating Semantic Consistency of Speech Recognition Hypotheses |
d16406435 | This paper presents MED, the main system of the LMU team for the SIGMOR-PHON 2016 Shared Task on Morphological Reinflection as well as an extended analysis of how different design choices contribute to the final performance. We model the task of morphological reinflection using neural encoder-decoder models together with an encoding of the input as a single sequence of the morphological tags of the source and target form as well as the sequence of letters of the source form. The Shared Task consists of three subtasks, three different tracks and covers 10 different languages to encourage the use of language-independent approaches. MED was the system with the overall best performance, demonstrating our method generalizes well for the low-resource setting of the SIGMORPHON 2016 Shared Task. | MED: The LMU System for the SIGMORPHON 2016 Shared Task on Morphological Reinflection |
d2621672 | ||
d18433092 | This paper describes the system implemented by Fundació Barcelona Media (FBM) for classifying the polarity of opinion expressions in tweets and SMSs, and which is supported by a UIMA pipeline for rich linguistic and sentiment annotations. FBM participated in the SEMEVAL 2013 Task 2 on polarity classification. It ranked 5th in Task A (constrained track) using an ensemble system combining ML algorithms with dictionary-based heuristics, and 7th (Task B, constrained) using an SVM classifier with features derived from the linguistic annotations and some heuristics. | FBM: Combining lexicon-based ML and heuristics for Social Media Polarities |
d11223001 | This paper considers the popular but questionable technique of 'round-trip translation' (RTT) as a means of evaluating free on-line Machine Translation systems. Two experiments are reported, both relating to common requirements of lay-users of MT on the web. In the first we see whether RTT can accurately predict the overall quality of the MT system. In the second, we ask whether RTT can predict the translatability of a given text. In both cases, we find RTT to be a poor predictor of quality, with high BLEU and F-scores for RTTs when the forward translation was poor. We discuss why this is the case, and conclude that, even if it seemed obvious that RTT was good for nothing, at least we now have some tangible evidence. | Round-Trip Translation: What Is It Good For? |
d233364900 | ||
d233364939 | Bidirectional Encoder Representations from Transformers (BERT) has gained popularity in recent years producing state-of-the-art performances across Natural Language Processing tasks. In this paper, we used AraBERT language model to binary classify pairs of verses provided by the QurSim dataset to either be semantically related or not. We have pre-processed The QurSim dataset and formed three datasets for comparisons. Also, we have used both versions of AraBERT, which are AraBERTv0.2 and AraBERTv2, to recognise which version performs the best with the given datasets. The best results was AraBERTv0.2 with 92% accuracy score using a dataset comprised of label '2' and label '-1', the latter was generated outside of QurSim dataset. | Quranic Verses Semantic Relatedness Using AraBERT |
d15779728 | Thought Thought Thought Thought De De De De se se se se, , , , first first first first person person person person indexicals indexicals indexicals indexicals and and and and Chinese Chinese Chinese Chinese reflexive reflexive reflexive reflexive ziji ziji ziji zijiAbstract Abstract Abstract AbstractIn this paper, we make a distinction between the de se and non-de se interpretations of first person indexicals and Chinese reflexive ziji. Based on the distinction, we discuss the relationship between these expressions in Chinese, and point out the problems with Wechsler's (2010) de se theory of person indexicals as well as the inappropriateness of characterizing Chinese long-distance ziji as a logophor. | |
d236999864 | ||
d250164058 | Numerical tables are widely employed to communicate or report the classification performance of machine learning (ML) models with respect to a set of evaluation metrics. For non-experts, domain knowledge is required to fully understand and interpret the information presented by numerical tables. This paper proposes a new natural language generation (NLG) task where neural models are trained to generate textual explanations, analytically describing the classification performance of ML models based on the metrics' scores reported in the tables. Presenting the generated texts along with the numerical tables will allow for a better understanding of the classification performance of ML models. We constructed a dataset comprising numerical tables paired with their corresponding textual explanations written by experts to facilitate this NLG task. Experiments on the dataset are conducted by fine-tuning pre-trained language models (T5 and BART) to generate analytical textual explanations conditioned on the information in the tables. Furthermore, we propose a neural module, Metrics Processing Unit (MPU), to improve the performance of the baselines in terms of correctly verbalising the information in the corresponding table. Evaluation and analysis conducted indicate, that exploring pre-trained models for data-to-text generation leads to better generalisation performance and can produce high-quality textual explanations. | Generating Textual Explanations for Machine Learning Models Performance: A Table-to-Text Task |
d18650536 | In this paper we tackle the problem of automatic caption generation for news images. Our approach leverages the vast resource of pictures available on the web and the fact that many of them are captioned. Inspired by recent work in summarization, we propose extractive and abstractive caption generation models. They both operate over the output of a probabilistic image annotation model that preprocesses the pictures and suggests keywords to describe their content. Experimental results show that an abstractive model defined over phrases is superior to extractive methods. | How Many Words is a Picture Worth? Automatic Caption Generation for News Images |
d33064197 | ||
d7180520 | We release a cross-lingual wikification system for all languages in Wikipedia. Given a piece of text in any supported language, the system identifies names of people, locations, organizations, and grounds these names to the corresponding English Wikipedia entries. The system is based on two components: a cross-lingual named entity recognition (NER) model and a crosslingual mention grounding model. The cross-lingual NER model is a language-independent model which can extract named entity mentions in the text of any language in Wikipedia. The extracted mentions are then grounded to the English Wikipedia using the cross-lingual mention grounding model. The only resources required to train the proposed system are the multilingual Wikipedia dump and existing training data for English NER. The system is online at | Illinois Cross-Lingual Wikifier: Grounding Entities in Many Languages to the English Wikipedia 1 Motivation |
d219303023 | ||
d8223389 | Combination of features contributes to a significant improvement in accuracy on tasks such as part-of-speech (POS) tagging and text chunking, compared with using atomic features. However, selecting combination of features on learning with large-scale and feature-rich training data requires long training time. We propose a fast boosting-based algorithm for learning rules represented by combination of features. Our algorithm constructs a set of rules by repeating the process to select several rules from a small proportion of candidate rules. The candidate rules are generated from a subset of all the features with a technique similar to beam search. Then we propose POS tagging and text chunking based on our learning algorithm. Our tagger and chunker use candidate POS tags or chunk tags of each word collected from automatically tagged data. We evaluate our methods with English POS tagging and text chunking. The experimental results show that the training time of our algorithm are about 50 times faster than Support Vector Machines with polynomial kernel on the average while maintaining stateof-the-art accuracy and faster classification speed. | A Fast Boosting-based Learner for Feature-Rich Tagging and Chunking |
d8851444 | This is an overall description of ADESSE ("Base de datos de verbos, Alternancias de Diátesis y Esquemas Sintactico-Semánticos del Español"), an online database (http://adesse.uvigo.es/) with syntactic and semantic information for all clauses in a corpus of Spanish. The manually annotated corpus has 1.5 million words, 159,000 clauses and 3,450 different verb lemmas. ADESSE is an expanded version of BDS ("Base de datos sintácticos del español actual"), which contains the grammatical features of verbs and verb-arguments in the corpus. ADESSE has added semantic features such as verb sense, verb class and semantic role of arguments to make possible a detailed syntactic and semantic corpus-based characterization of verb valency. Each verb entry in the database is described in terms of valency potential and valency realizations (diatheses). The former includes a set of semantic roles of participants in a particular event type and a classification into a conceptual hierarchy of process types. Valency realizations are described in terms of correspondences of voice, syntactic functions and categories, and semantic roles. Verbs senses are discriminated at two levels: a more abstract level linked to a valency potential, and more specific verb senses taking into account particular lexical instantiations of arguments. | ADESSE. A Database with Syntactic and Semantic Annotation of a Corpus of Spanish |
d5260223 | In morphologically rich languages such as Arabic, the abundance of word forms resulting from increased morpheme combinations is significantly greater than for languages with fewer inflected forms(Kirchhoff et al., 2006). This exacerbates the out-of-vocabulary (OOV) problem. Test set words are more likely to be unknown, limiting the effectiveness of the model. The goal of this study is to use the regularities of Arabic inflectional morphology to reduce the OOV problem in that language. We hope that success in this task will result in a decrease in word error rate in Arabic automatic speech recognition. | Arabic Language Modeling with Finite State Transducers |
d213681 | Beside the word order problem, word choice is another major obstacle for machine translation. Though phrase-based statistical machine translation (SMT) has an advantage of word choice based on local context, exploiting larger context is an interesting research topic. Recently, there have been a number of studies on integrating word sense disambiguation (WSD) into phrase-based SMT. The WSD score has been used as a feature of translation. In this paper, we will show that by bootstrapping WSD models using unlabeled data, we can bootstrap an SMT system. Our experiments on English-Vietnamese translation showed that BLEU scores have been improved significantly. | Bootstrapping Phrase-based Statistical Machine Translation via WSD Integration |
d12623074 | In this paper, we describe the SemEval-2010 shared task on "Linking Events and Their Participants in Discourse". This task is a variant of the classical semantic role labelling task. The novel aspect is that we focus on linking local semantic argument structures across sentence boundaries. Specifically, the task aims at linking locally uninstantiated roles to their coreferents in the wider discourse context (if such co-referents exist). This task is potentially beneficial for a number of NLP applications and we hope that it will not only attract researchers from the semantic role labelling community but also from co-reference resolution and information extraction. | SemEval-2010 Task 10: Linking Events and Their Participants in Discourse |
d16051864 | We take a novel approach to rapid, low-cost development of morpho-syntactically annotated resources without using parallel corpora or bilingual lexicons. The overall research question is how to exploit language resources and properties to facilitate and automate the creation of morphologically annotated corpora for new languages. This portability issue is especially relevant to minority languages, for which such resources are likely to remain unavailable in the foreseeable future. We compare the performance of our system on languages that belong to different language families (Romance vs. Slavic), as well as different language pairs within the same language family (Portuguese via Spanish vs. Catalan via Spanish). We show that across language families, the most difficult category is the category of nominals (the noun homonymy is challenging for morphological analysis and the order variation of adjectives within a sentence makes it challenging to create a realiable model), whereas different language families present different challenges with respect to their morpho-syntactic descriptions: for the Slavic languages, case is the most challenging category; for the Romance languages, gender is more challenging than case. In addition, we present an alternative evaluation metric for our system, where we measure how much human labor will be needed to convert the result of our tagging to a high precision annotated resource. | A Cross-language Approach to Rapid Creation of New Morpho-syntactically Annotated Resources |
d14503241 | In this paper, we investigate the usage of a non-canonical German passive alternation for ditransitive verbs, the recipient passive, in naturally occuring corpus data. We propose a classifier that predicts the voice of a ditransitive verb based on the contextually determined properties of its arguments. As the recipient passive is a low frequent phenomenon, we first create a special data set focussing on German ditransitive verbs which are frequently used in the recipient passive. We use a broad-coverage grammar-based parser, the German LFG parser, to automatically annotate our data set for the morpho-syntactic properties of the involved predicate arguments. We train a Maximum Entropy classifier on the automatically annotated sentences and achieve an accuracy of 98.05%, clearly outperforming the baseline that always predicts active voice (94.6%). | A Corpus-based Study of the German Recipient Passive |
d44177418 | Recent research in language and vision has developed models for predicting and disambiguating verbs from images. Here, we ask whether the predictions made by such models correspond to human intuitions about visual verbs. We show that the image regions a verb prediction model identifies as salient for a given verb correlate with the regions fixated by human observers performing a verb classification task. | An Evaluation of Image-based Verb Prediction Models against Human Eye-tracking Data |
d36100175 | Sarcasm is a pervasive phenomenon in social media, permitting the concise communication of meaning, affect and attitude. Concision requires wit to produce and wit to understand, which demands from each party knowledge of norms, context and a speaker's mindset. Insight into a speaker's psychological profile at the time of production is a valuable source of context for sarcasm detection. Using a neural architecture, we show significant gains in detection accuracy when knowledge of the speaker's mood at the time of production can be inferred. Our focus is on sarcasm detection on Twitter, and show that the mood exhibited by a speaker over tweets leading up to a new post is as useful a cue for sarcasm as the topical context of the post itself. The work opens the door to an empirical exploration not just of sarcasm in text but of the sarcastic state of mind. | Magnets for Sarcasm: Making Sarcasm Detection Timely, Contextual and Very Personal |
d12928034 | We offer a noisy channel approach for recognizing and correcting erroneous words in referring expressions. Our mechanism handles three types of errors: it removes noisy input, inserts missing prepositions, and replaces mis-heard words (at present, they are replaced by generic words). Our mechanism was evaluated on a corpus of 295 spoken referring expressions, improving interpretation performance. | A Noisy Channel Approach to Error Correction in Spoken Referring Expressions |
d9406440 | In this work, we tackle the problem of language and translation models domainadaptation without explicit bilingual indomain training data. In such a scenario, the only information about the domain can be induced from the source-language test corpus. We explore unsupervised adaptation, where the source-language test corpus is combined with the corresponding hypotheses generated by the translation system to perform adaptation. We compare unsupervised adaptation to supervised and pseudo supervised adaptation. Our results show that the choice of the adaptation (target) set is crucial for successful application of adaptation methods. Evaluation is conducted over the German-to-English WMT newswire translation task. The experiments show that the unsupervised adaptation method generates the best translation quality as well as generalizes well to unseen test sets. | Unsupervised Adaptation for Statistical Machine Translation |
d6826327 | This work describes the UoE-LMU submission for the CoNLL-SIGMORPHON 2017 Shared Task on Universal Morphological Reinflection, Subtask 1: given a lemma and target morphological tags, generate the target inflected form. We evaluate several ways to improve performance in the 1000-example setting: three methods to augment the training data with identical input-output pairs (i.e., autoencoding), a heuristic approach to identify likely pairs of inflectional variants from an unlabeled corpus, and a method for crosslingual knowledge transfer. We find that autoencoding random strings works surprisingly well, outperformed only slightly by autoencoding words from an unlabelled corpus. The random string method also works well in the 10,000-example setting despite not being tuned for it. Among 18 submissions our system takes 1st and 6th place in the 10k and 1k settings, respectively. | Training Data Augmentation for Low-Resource Morphological Inflection |
d260480536 | This paper is part of our broader investigation into the utility of discourse structure for performance analysis. In our previous work, we showed that several interaction parameters that use discourse structure predict our performance metric. Here, we take a step forward and show that these correlations are not only a surface relationship. We show that redesigning the system in light of an interpretation of a correlation has a positive impact. | Discourse Structure and Performance Analysis: Beyond the Correlation |
d23666159 | ||
d252847503 | Reflection about a learning process is beneficial to students in higher education(Bubnys, 2019). The importance of machine understanding of reflective texts grows as applications supporting students become more widespread. Nevertheless, due to the sensitive content, there is no public corpus available yet for the classification of text reflectiveness. We provide the first open-access corpus of reflective student essays in German. We collected essays from three different disciplines (Software Development, Ethics of Artificial Intelligence and Teacher Training). We annotated the corpus at sentence level with binary reflective/non-reflective labels, using an iterative annotation process with linguistic and didactic specialists, mapping the reflective components found in the data to existing schemes and complementing them. We propose and evaluate linguistic features of reflectiveness and analyse their distribution within the resulted sentences according to their labels. Our contribution constitutes the first open-access corpus to help the community towards a unified approach for reflection detection. | A German Corpus of Reflective Sentences |
d259376903 | In this paper, we present principles of constructing and resolving ambiguity in implicit discourse relations. Following these principles, we created a dataset in both English and Egyptian Arabic that controls for semantic disambiguation, enabling the investigation of prosodic features in future work. In these datasets, examples are two-part sentences with an implicit discourse relation that can be ambiguously read as either causal or concessive, paired with two different preceding context sentences forcing either the causal or the concessive reading. We also validated both datasets by humans and language models (LMs) to study whether context can help humans or LMs resolve ambiguities of implicit relations and identify the intended relation. As a result, this task posed no difficulty for humans, but proved challenging for BERT/CamelBERT and ELEC-TRA/AraELECTRA models. | Unpacking Ambiguous Structure: A Dataset for Ambiguous Implicit Discourse Relations for English and Egyptian Arabic |
d10549264 | ||
d53088130 | Les systèmes de résumé automatique de textes (SRAT) consistent à produire une représentation condensée et pertinente à partir d'un ou de plusieurs documents textuels. La majorité des SRAT sont basés sur des approches extractives. La tendance actuelle consiste à s'orienter vers les approches abstractives. Dans ce contexte, le résumé guidé défini par la campagne d'évaluation internationale TAC (Text Analysis Conference) en 2010, vise à encourager la recherche sur ce type d'approche, en se basant sur des techniques d'analyse en profondeur de textes. Dans ce papier, nous nous penchons sur le résumé automatique guidé de textes. Dans un premier temps, nous définissons les différentes caractéristiques et contraintes liées à cette tâche. Ensuite, nous dressons un état de l'art des principaux systèmes existants en mettant l'accent sur les travaux les plus récents, et en les classifiant selon les approches adoptées, les techniques utilisées, et leurs évaluations sur des corpus de références. Enfin, nous proposons les grandes étapes d'une méthode spécifique devant permettre le développement d'un nouveau type de systèmes de résumé guidé.ABSTRACTGuided Summarization : State-of-the-art and perspectives Automatic text summarization (ATS) aims to produce from one or more texts, a summary that represents the most relevant information included in the original textual sources. Most existing ATS are mainly the extraction-based systems ; however, the trend today is to make a move toward abstraction-based systems. In 2010, the Text Analysis Conference (TAC) campaign defined the guided summarization task as a recent type of ATS, aims to encourage a deeper linguistic analysis of the source documents instead of relying only on classical extractive approaches. In this work, we provide an introduction to guided summarization task, by defining the different characteristics and constraints related to this task, and by reviewing the details of different guided summarization systems developed so far. We also classify these systems according to the adopted approaches, techniques used, and evaluations on reference corpus. Finally, we propose the main steps of a specific method that will allow the development of a new type of guided summary systems. MOTS-CLÉS : Résumé automatique de textes, résumé guidé, approche extractive, approche abstractive, traitement automatique de la langue naturelle, extraction d'information. | Résumé automatique guidé de textes : État de l'art et perspectives |
d61207035 | MACHINE TRANSLATION AND ARTIFICIAL INTELLIGENCE | |
d19035620 | LR RECURSIVE TRANSITION NETWORKS FOR EARLEY AND TOMITA PARSING | |
d53589436 | Computational Language Documentation attempts to make the most recent research in speech and language technologies available to linguists working on language preservation and documentation. In this paper, we pursue two main goals along these lines. The first is to improve upon a strong baseline for the unsupervised word discovery task on two very lowresource Bantu languages, taking advantage of the expertise of linguists on these particular languages. The second consists in exploring the Adaptor Grammar framework as a decision and prediction tool for linguists studying a new language. We experiment 162 grammar configurations for each language and show that using Adaptor Grammars for word segmentation enables us to test hypotheses about a language. Specializing a generic grammar with language specific knowledge leads to great improvements for the word discovery task, ultimately achieving a leap of about 30% token F-score from the results of a strong baseline. | Adaptor Grammars for the Linguist: Word Segmentation Experiments for Very Low-Resource Languages |
d53082447 | Categorizing a patient's intentions during clinical interactions in general and within motivational interviewing specifically may improve decision making in clinical treatments. Within this paper, we propose a method that models the temporal flow of a conversation and the transition between topics by using domain adaptation on a clinical dialogue corpus comprising Motivational Interviewing (MI) sessions. We deploy Bi-LSTM and topic models jointly to learn theme shifts across different time stages within these hour-long MI sessions to assess the patient's intent to change their habits or to sustain them respectively. Our experiments show promising results and improvements after considering temporality in the classification task over our baseline. This result confirms and extends related literature that has manually identified that certain phases within MI sessions are more predictive of patient outcomes than others. . 2013.API design for machine learning software: experiences from the scikit-learn project. In ECML PKDD Workshop: Languages for Data Mining and Machine Learning, pages 108-122. Kate B Carey, James M Henson, Michael P Carey, and Stephen A Maisto. 2009. Computer versus inperson intervention for students violating campus alcohol policy. Journal of consulting and clinical psychology, 77(1):74. | Modeling Temporality of Human Intentions by Domain Adaptation |
d12339022 | This paper presents the model that can be used to filter the texts which the user is interested in from a large scale of source texts in Chinese or in English. Each text which the user is interested in can be represented as a vector in the vector space of classifiable sememes. The text to be sifted is represented as a vector too. The relevance of the text to the user can be measured by using the cosine angle between the text and its k nearest neighbor in the vector space. Experiments have been done and their results show that this scheme yields good results . | Cross-Lingual Text Filtering Based On Text Concepts And kNN |
d41531604 | In this work we present the open source hunvec framework for sequential tagging, built upon Theano and Pylearn2. The underlying statistical model, which connects linear CRF-s with neural networks, was used by Collobert and co-workers, and several other researchers. For demonstrating the flexibility of our tool, we describe a set of experiments on part-of-speech and named-entityrecognition tasks, using English and Hungarian datasets, where we modify both model and training parameters, and illustrate the usage of custom features. Model parameters we experiment with affect the vectorial word representations used by the model; we apply different word vector initializations, defined by Word2vec and GloVe embeddings and enrich the representation of words by vectors assigned trigram features. We extend training methods by using their regularized (l2 and dropout) version. When testing our framework on a Hungarian named entity corpus, we find that its performance reaches the best published results on this dataset, with no need for language-specific feature engineering. Our code is available at http://github.com/zseder/hunvec | The Hunvec Framework For NN-CRF-based Sequential Tagging |
d69289238 | We present a new method for unsupervised learning of multilingual symbol (e.g. character) embeddings, without any parallel data or prior knowledge about correspondences between languages. It is able to exploit similarities across languages between the distributions over symbols' contexts of use within their language, even in the absence of any symbols in common to the two languages. In experiments with an artificially corrupted text corpus, we show that the method can retrieve character correspondences obscured by noise. We then present encouraging results of applying the method to real linguistic data, including for low-resourced languages. The learned representations open the possibility of fully unsupervised comparative studies of text or speech corpora in low-resourced languages with no prior knowledge regarding their symbol sets. | Unsupervised Learning of Cross-Lingual Symbol Embeddings Without Parallel Data |
d62483340 | Book Review Language as a Cognitive Process | |
d9018481 | Several grammars have been proposed for modeling RNA pseudoknotted structure. In this paper, we focus on multiple context-free grammars (MCFGs), which are natural extension of context-free grammars and can represent pseudoknots, and extend a specific subclass of MCFGs to a probabilistic model called SMCFG. We present a polynomial time parsing algorithm for finding the most probable derivation tree and a probability parameter estimation algorithm. Furthermore, we show some experimental results of pseudoknot prediction using SMCFG algorithm. * ⇒ f [(ab, cd)] = (aabb, ccdd) by the first rule. The probability of the latter derivation is 0.3 · 0.7 = 0.21. The language generated by an SMCFG G is defined as L(G) = {w ∈ T * | S * ⇒ | Stochastic Multiple Context-Free Grammar for RNA Pseudoknot Modeling |
d13391708 | This paper evaluates the benefit of deleting fillers (e.g. you know, like) early in parsing conversational speech. Readability studies have shown that disfluencies (fillers and speech repairs) may be deleted from transcripts without compromising meaning(Jones et al., 2003), and deleting repairs prior to parsing has been shown to improve its accuracy(Charniak and Johnson, 2001). We explore whether this strategy of early deletion is also beneficial with regard to fillers. Reported experiments measure the effect of early deletion under in-domain and out-of-domain parser training conditions using a state-of-the-art parser(Charniak, 2000). While early deletion is found to yield only modest benefit for in-domain parsing, significant improvement is achieved for out-of-domain adaptation. This suggests a potentially broader role for disfluency modeling in adapting text-based tools for processing conversational speech. | Early Deletion of Fillers In Processing Conversational Speech |
d13635396 | Reordering is a major challenge for machine translation between distant languages. Recent work has shown that evaluation metrics that explicitly account for target language word order correlate better with human judgments of translation quality. Here we present a simple framework for evaluating word order independently of lexical choice by comparing the system's reordering of a source sentence to reference reordering data generated from manually word-aligned translations. When used to evaluate a system that performs reordering as a preprocessing step our framework allows the parser and reordering rules to be evaluated extremely quickly without time-consuming endto-end machine translation experiments. A novelty of our approach is that the translations used to generate the reordering reference data are generated in an alignment-oriented fashion. We show that how the alignments are generated can significantly effect the robustness of the evaluation. We also outline some ways in which this framework has allowed our group to analyze reordering errors for English to Japanese machine translation. | A Lightweight Evaluation Framework for Machine Translation Reordering |
d220047813 | In this paper, we show that neural machine translation (NMT) systems trained on large back-translated data overfit some of the characteristics of machine-translated texts. Such NMT systems better translate humanproduced translations, i.e., translationese, but may largely worsen the translation quality of original texts. Our analysis reveals that adding a simple tag to back-translations prevents this quality degradation and improves on average the overall translation quality by helping the NMT system to distinguish back-translated data from original parallel data during training. We also show that, in contrast to high-resource configurations, NMT systems trained in lowresource settings are much less vulnerable to overfit back-translations. We conclude that the back-translations in the training data should always be tagged especially when the origin of the text to be translated is unknown. | Tagged Back-translation Revisited: Why Does It Really Work? |
d249455893 | Named entity recognition (NER) in a realworld setting remains challenging and is impacted by factors like text genre, corpus quality, and data availability. NER models trained on CoNLL do not transfer well to other domains, even within the same language. This is especially the case for multi-lingual models when applied to low-resource languages, and is mainly due to missing entity information.We propose an approach that with limited effort and data, addresses the NER knowledge gap across languages and domains. Our novel approach uses a token-level gating layer to augment pre-trained multilingual transformers with gazetteers containing named entities (NE) from a target language or domain. This approach provides the flexibility to jointly integrate both textual and gazetteer information dynamically: entity knowledge from gazetteers is used only when a token's textual representation is insufficient for the NER task.Evaluation on several languages and domains demonstrates: (i) a high mismatch of reported NER performance on CoNLL vs. domain specific datasets, (ii) gazetteers significantly improve NER performance across languages and domains, and (iii) gazetteers can be flexibly incorporated to guide knowledge transfer. On cross-lingual transfer we achieve an improvement over the baseline with F1=+17.6%, and with F1=+21.3% for cross-domain transfer. | Dynamic Gazetteer Integration in Multilingual Models for Cross-Lingual and Cross-Domain Named Entity Recognition |
d226283626 | The CoNLL-2003 corpus for Englishlanguage named entity recognition (NER) is one of the most influential corpora for NER model research. A large number of publications, including many landmark works, have used this corpus as a source of ground truth for NER tasks. In this paper, we examine this corpus and identify over 1300 incorrect labels (out of 35089 in the corpus). In particular, the number of incorrect labels in the test fold is comparable to the number of errors that state-of-the-art models make when running inference over this corpus.We describe the process by which we identified these incorrect labels, using novel variants of techniques from semi-supervised learning. We also summarize the types of errors that we found, and we revisit several recent results in NER in light of the corrected data. Finally, we show experimentally that our corrections to the corpus have a positive impact on three state-ofthe-art models. * The last four authors have contributed equally. | Identifying Incorrect Labels in the CoNLL-2003 Corpus |
d67239762 | A new method is presented for simplifying the logical expressions used to represent utterance meaning in a natural language system. | A Terminological Simplification Transformation for Natural Language Question-Answering Systems |
d9107928 | Textual records of business-oriented conversations between customers and agents need to be analyzed properly to acquire useful business insights that improve productivity. For such an analysis, it is critical to identify appropriate textual segments and expressions to focus on, especially when the textual data consists of complete transcripts, which are often lengthy and redundant. In this paper, we propose a method to identify important segments from the conversations by looking for changes in the accuracy of a categorizer designed to separate different business outcomes. We extract effective expressions from the important segments to define various viewpoints. In text mining a viewpoint defines the important associations between key entities and it is crucial that the correct viewpoints are identified. We show the effectiveness of the method by using real datasets from a car rental service center. | Automatic Identification of Important Segments and Expressions for Mining of Business-Oriented Conversations at Contact Centers |
d233365325 | Data in general encodes human biases by default; being aware of this is a good start, and the research around how to handle it is ongoing. The term 'bias' is extensively used in various contexts in NLP systems. In our research the focus is specific to biases such as gender, racism, religion, demographic and other intersectional views on biases that prevail in text processing systems responsible for systematically discriminating specific population, which is not ethical in NLP. These biases exacerbate the lack of equality, diversity and inclusion of specific population while utilizing the NLP applications. The tools and technology at the intermediate level utilize biased data, and transfer or amplify this bias to the downstream applications. However, it is not enough to be colourblind, gender-neutral alone when designing a unbiased technologyinstead, we should take a conscious effort by designing a unified framework to measure and benchmark the bias. In this paper, we recommend six measures and one augment measure based on the observations of the bias in data, annotations, text representations and debiasing techniques. | An Overview of Fairness in Data -Illuminating the Bias in Data Pipeline |
d7210871 | A desirable quality of a coreference resolution system is the ability to handle transitivity constraints, such that even if it places high likelihood on a particular mention being coreferent with each of two other mentions, it will also consider the likelihood of those two mentions being coreferent when making a final assignment. This is exactly the kind of constraint that integer linear programming (ILP) is ideal for, but, surprisingly, previous work applying ILP to coreference resolution has not encoded this type of constraint. We train a coreference classifier over pairs of mentions, and show how to encode this type of constraint on top of the probabilities output from our pairwise classifier to extract the most probable legal entity assignments. We present results on two commonly used datasets which show that enforcement of transitive closure consistently improves performance, including improvements of up to 3.6% using the b 3 scorer, and up to 16.5% using cluster f-measure. | Enforcing Transitivity in Coreference Resolution |
d51870345 | Due to the presence of both Twitterspecific conventions and non-standard and dialectal language, Twitter presents a significant parsing challenge to current dependency parsing tools. We broaden English dependency parsing to handle social media English, particularly social media African-American English (AAE), by developing and annotating a new dataset of 500 tweets, 250 of which are in AAE, within the Universal Dependencies 2.0 framework. We describe our standards for handling Twitter-and AAE-specific features and evaluate a variety of crossdomain strategies for improving parsing with no, or very little, in-domain labeled data, including a new data synthesis approach. We analyze these methods' impact on performance disparities between AAE and Mainstream American English tweets, and assess parsing accuracy for specific AAE lexical and syntactic features. Our annotated data and a parsing model are available at: | |
d5385823 | Resea/ch based on a treebank is active for many natural language applications. However, the work to build a large scale treebank is laborious and tedious. This paper proposes a probabilistic chunker to help the development of a partially bracketed corpus. The chunker partitions the part-of-speech sequence into segments called chunks. Rather than using a treebank as our training corpus, a corpus which is tagged with part-of-speech information only is used. The experimental results show the probabilistic chunker has more than 92% correct rate in outside test. The well-formed partially bracketed corpus is a milestone in the development of a treebank. Besides, the simple but effective chunker can also be applied to many natural language applications. | Development of a Partially Bracketed Corpus with Part-of-Speech Information Only |
d10984754 | Recently, speech recognition performance has been drastically improved by statistical methods and huge speech databases. Now performance improvement under such realistic environments as noisy conditions is being focused on. Since October 2001, we from the working group of the Information Processing Society in Japan have been working on evaluation methodologies and frameworks for Japanese noisy speech recognition. We have released frameworks including databases and evaluation tools called CENSREC-1 (Corpus and Environment for Noisy Speech RECognition 1; formerly AURORA-2J), CENSREC-2 (in-car connected digits recognition), CENSREC-3 (in-car isolated word recognition), and CENSREC-1-C (voice activity detection under noisy conditions). In this paper, we newly introduce a collection of databases and evaluation tools named CENSREC-4, which is an evaluation framework for distant-talking speech under hands-free conditions. Distant-talking speech recognition is crucial for a hands-free speech interface. Therefore, we measured room impulse responses to investigate reverberant speech recognition. The results of evaluation experiments proved that CENSREC-4 is an effective database suitable for evaluating the new dereverberation method because the traditional dereverberation process had difficulty sufficiently improving the recognition performance. The framework was released in March 2008, and many studies are being conducted with it in Japan. | Evaluation Framework for Distant-talking Speech Recognition under Reverberant Environments -Newest Part of the CENSREC Series |
d2947828 | This paper, the 5 th in a series of biennial progress reports, reviews the activities of the Linguistic Data Consortium with particular emphasis on general trends in the language resource landscape and on changes that distinguish the two years since LDC's last report at LREC from the preceding 8 years. After providing a perspective on the current landscape of language resources, the paper goes on to describe our vision of the role of LDC within the research communities it serves before sketching briefly specific publications and resources creations projects that have been the focus our attention since the last report. | 5 Years of Language Resource Creation and Sharing: A Progress Report on LDC Activities |
d28847083 | Nos recherches ont d6but6 il y a bientOt quatre ans. Ce projet fut d'abord int6gr6 aux activit6s g6n6rales de la Facult6 des Lettres de l'Universit6 de Montr6al; maintenant, il d6pend directement du vice-recteur ~ la recherche de la mSme universit6.Le P.l. avance de gauche ~ droite sur la chaTne d'entr6e, ajoutant un symhole ~ la lois (s6par6 par des blancs de part et d'autre) au segment d6j~ interpr6t6. Pour chaque segment, il 6tablit un r6seau de noeuds 6tiquet6s, que nous pouvons repr6senter comme un demi treillis [el.fig. lJ. | I -HISTORIQUE DU PROJET |
d232021585 | ||
d778011 | To date, there are no WSD systems for Arabic. In this paper we present and evaluate a novel unsupervised approach, SALAAM, which exploits translational correspondences between words in a parallel Arabic English corpus to annotate Arabic text using an English WordNet taxonomy. We illustrate that our approach is highly accurate in ¢ ¡ ¤ £ ¦ ¥ § © of the evaluated data items based on Arabic native judgement ratings and annotations. Moreover, the obtained results are competitive with state-of-the-art unsupervised English WSD systems when evaluated on English data.SALAAM favors the last meaning definition for congregation. | An Unsupervised Approach for Bootstrapping Arabic Sense Tagging |
d13891858 | In this paper we present a novel approach to multi-word terminology extraction combining a well-known automatic term recognition approach, the C-NC value method, with a contrastive ranking technique, aimed at refining obtained results either by filtering noise due to common words or by discerning between semantically different types of terms within heterogeneous terminologies. The proposed methodology has been tested in two case studies carried out in the History of Art and Legal domains with promising results. | A Contrastive Approach to Multi-word Term Extraction from Domain Corpora |
d21720027 | In this paper, we introduce the first version of ForFun, Prague Database of Forms and Functions, as an invaluable resource for profound linguistic research, particularly in describing syntactic functions and their formal realizations. ForFun is built with the use of already existing richly syntactically annotated corpora, collectively called Prague Dependency Treebanks. ForFun brings this complex annotation of Czech sentences closer to researchers. We demonstrate that ForFun 1.0 provides valuable and rich material allowing to elaborate various syntactic issues in depth. We believe that nowadays when corpus linguistics differs from traditional linguistics in its insistence on a systematic study of authentic examples of language in use, our database will contribute to the comprehensive syntactic description. | ForFun 1.0: Prague Database of Forms and Functions An Invaluable Resource for Linguistic Research |
d14135824 | To solve the unknown morpheme problem in Japanese morphological analysis, we previously proposed a novel framework of online unknown morpheme acquisition and its implementation. This framework poses a previously unexplored problem, online unknown morpheme detection. Online unknown morpheme detection is a task of finding morphemes in each sentence that are not listed in a given lexicon. Unlike in English, it is a non-trivial task because Japanese does not delimit words by white space. We first present a baseline method that simply uses the output of the morphological analyzer. We then show that it fails to detect some unknown morphemes because they are over-segmented into shorter registered morphemes. To cope with this problem, we present a simple solution, the use of orthographic variation of Japanese. Under the assumption that orthographic variants behave similarly, each over-segmentation candidate is checked against its counterparts. Experiments show that the proposed method improves the recall of detection and contributes to improving unknown morpheme acquisition. | Online Japanese Unknown Morpheme Detection using Orthographic Variation |
d248780370 | Warning: this paper contains examples that may be offensive or upsetting.The social impact of natural language processing and its applications has received increasing attention. In this position paper, we focus on the problem of safety for end-to-end conversational AI. We survey the problem landscape therein, introducing a taxonomy of three observed phenomena: the INSTIGATOR, YEA-SAYER, and IMPOSTOR effects. We then empirically assess the extent to which current tools can measure these effects and current systems display them. We release these tools as part of a "first aid kit" (SAFETYKIT) to quickly assess apparent safety concerns. Our results show that, while current tools are able to provide an estimate of the relative safety of systems in various settings, they still have several shortcomings. We suggest several future directions and discuss ethical considerations. | SAFETYKIT: First Aid for Measuring Safety in Open-domain Conversational Systems |
d11143530 | Discourse markers are complex discontinuous linguistic expressions which are used to explicitly signal the discourse structure of a text. This paper describes efforts to improve an automatic tagging system which identifies and classifies discourse markers in Chinese texts by applying machine learning (ML) to the disambiguation of discourse markers, as an integral part of automatic text summarization via rhetorical structure. Encouraging results are reported. | Enhancement of a Chinese Discourse Marker Tagger with C4.5 |
d19607997 | Edinburgh: Edinburgh University Press (Edinburgh Information Technology Series 7, edited by S. Michaelson and M. Steedman), 1990, x + 276 pp. Hardbound, ISBN 0-7486-0162-7 (distributed in North America by Columbia University Press), $60.00The art of automatic speech recognition has advanced remarkably in the past decade. With the advances in accuracy and scope, there has come, for the time being, a strong convergence on a class of statistical methods based on a structure called a hidden Markov model (HMM). HMM-based systems dominate speech recognition, and success in the speech domain has spawned many attempts to extend HMM methods to related patternrecognition fields such as document recognition and handwriting recognition.The book under review answers a clear need. It introduces most of the theory and techniques needed to build a complete HMM-based speech recognition system. Huang, Akiri, and Jack use the first 90 pages to cover general methods of pattern recognition, speech-signal processing, and statistical language modeling. The next 90 pages cover signal quantization and the theory of HMMs for speech recognition. The last 60 pages cover practical issues with examples provided and explained.The style of the book is very dry but clear; it reads like an abstract of an engineering text. The majority of the pages are thick with equations, but the authors seem to use only as much mathematics as will be needed to actually implement the algorithms. There are few derivations and no proofs. The authors present the basic algorithms and the best algorithms (in their judgment), but offer neither historical perspective nor critical review of current research. The book gives algorithm examples written in a high-level English pseudo-code. These examples are concrete and will be helpful for a novice. The tone of the book is decidedly practical, although it is too concise. The book would need substantial expansion to work as a graduate-level text.In sum, for the reader-cum-system-builder who is ready to implement an HMM speech recognition system from the bottom up, this is the best available book that I have seen. Huang, Akiri, and Jack have distilled and presented the essential techniques of the trade.--Jared Bernstein, SRI International Hardbound, ISBN 0-7923-9214-0, $62.50, £43.25, Dfl 145.00Michael Mauldin has attempted the difficult task of building a full-fledged information retrieval system in the traditional design, but one with a language-understanding flavor. His purpose is to demonstrate that use of semantic knowledge will improve our capability to retrieve information. And he demonstrated his achievement by testing his system, FERRET, against a Boolean retrieval prototype of contemporary design. Unfortunately, his effort is marred by the commonly recognized flaws of the IR experimental methodology. However, given the realities of human interaction, he had little choice but to employ that methodology to convince his target audience. For the computational linguist, the work is of interest because it demonstrates the practicality of recognized techniques. There is little new here; however, in putting the pieces together, the author has provided us with convincing evidence that a high-level parse can be extremely useful for a number of activities requiring analysis of text. Mauldin includes a catalog of suggested uses in his last chapter.FERRET employs an adapted version of DeJong's (1979) FRUMP to parse articles from UseNet concerning astronomy. The parse is really a skimming of the text. Complex bits are passed over. Then "sketchy scripts" and case frames are compiled from the resulting conceptual dependencies (CD)(Schank et al. 1975). It was the author's purpose to demonstrate that a wide range of textual domains could be successfully treated in such a way, and the resulting representa-217 | Briefly Noted Hidden Markov Models for Speech Recognition Conceptual Information Retrieval: A Case Study in Adaptive Partial Parsing |
d252624658 | This paper discusses work in progress on the digitization of a sketch map of the Taz River basin -a region that is lacking highly detailed open-source cartography data. The original sketch is retrieved from the archive of Selkup materials gathered by Angelina Ivanovna Kuzmina in the 1960s and 1970s. The data quality and challenges that come with it are evaluated and a task-specific workflow is designed. The process of the turning a series of hand-drawn images with non-structured geographical and linguistic data into an interactive, geographically precise digital map is described both from linguistic and technical perspectives. Furthermore, the map objects in focus are differentiated based on the geographical type of the object and the etymology of the name. This provides an insight into the peculiarities of the linguistic history of the region and contributes to the cartography of the Uralic languages. | How to Digitize Completely: Interactive Geovizualization of a Sketch Map from the Kuzmina Archive |
d236486181 | Abusive language is a growing phenomenon on social media platforms. Its effects can reach beyond the online context, contributing to mental or emotional stress on users. Automatic tools for detecting abuse can alleviate the issue. In practice, developing automated methods to detect abusive language relies on good quality data. However, there is currently a lack of standards for creating datasets in the field. These standards include definitions of what is considered abusive language, annotation guidelines and reporting on the process. This paper introduces an annotation framework inspired by legal concepts to define abusive language in the context of online harassment. The framework uses a 7point Likert scale for labelling instead of class labels. We also present ALYT -a dataset of Abusive Language on YouTube. ALYT includes YouTube comments in English extracted from videos on different controversial topics and labelled by Law students. The comments were sampled from the actual collected data, without artificial methods for increasing the abusive content. The paper describes the annotation process thoroughly, including all its guidelines and training steps. | Abusive Language on Social Media Through the Legal Looking Glass |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.