Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18265 papers
The slowness of legal proceedings in the common law legal system is a widely known fact. Any tool which could help reduce the time taken for the resolution of a case is invaluable. Common legal systems place a great importance on precedents and retrieving the correct set of precedents is considerably time consuming. Hence, for any case whose proceedings are in progress, if there are suitable prior cases, then the court has to follow the same interpretations that were passed in the prior cases. This is to ensure that similar situations receive similar treatment, thus maintaining uniformity amongst the legal proceedings across all courts at all times. Hence, precedent cases are treated as important as any other written law (a statute) in this legal system. In this paper, we propose two new approaches to solve this information retrieval problem wherein the system accepts the current case document as the query and returns the relevant precedent cases as the result. The first approach is to calculate the document similarity using Wordnet, which is a lexical database that could be leveraged to quantify the semantic relatedness between two documents, using a semantic network. The second approach is the use of a Siamese Manhattan Long Short Term Memory network, which is a supervised model trained to understand the underlying similarity between two documents.
This article contends that researchers can and should be active participants in making sound archives more accessible. In fact, such advocacy can be consequential in setting up possibilities for creative research on race within radio history and sound studies. Using the example of NPR’s All Things Considered archive spanning from 1971–1983, I demonstrate how academics and archivists can work together to make possible the preservation and accessibility of sound archives. This particular collaboration offers an opportunity to take a cultural approach to understanding newsroom diversity, more specifically: the cultural constraints of linguistic norms and the emergent cultures that arise as aberrations from such norms. The article reflects on this project’s implications for other scholars who work with archives that wish to invest in sound archive preservation and use.
The Ministry of Education and Science of the Russian Federation has established a score-rating system to assess the achievement quality within academic disciplines. The word "rating" is a foreign-language term, literal translation into Russian means "assessment". The second meaning of the word "rating" is a numerical measure; it characterizes performance of a student, pupil, etc., during a certain period of training, usually by the 20-point scale, or 100-point scale. This article highlights the history of formation and implementation of a score-rating system into the curriculum. Due to development of advanced technologies, there is a need for training highly qualified personnel to solve increasingly complex problems in professional activities. In our opinion, pedagogical assessment of student's knowledge in the form of score-rating system is a key role for preparing highly qualified personnel. The authors study processes of training skilled professionals in the course of academic education in technical universities of our country. Key words: score-rating system (SRS), competence, student, progress, credit, rating, factor.
Although SGD requires shuffling the training data between epochs, currently none of the word-level language modeling systems do this. Naively shuffling all sentences in the training data would not permit the model to learn inter-sentence dependencies. Here we present a method that partially shuffles the training data between epochs. This method makes each batch random, while keeping most sentence ordering intact. It achieves new state of the art results on word-level language modeling on both the Penn Treebank and WikiText-2 datasets.
Music has been shown to influence the behavioral responses of individuals in real-world scenarios, but little research exists on the effects that music has on the in-game behaviors of video game players. A song can be rated in terms of the level of arousal, or emotional intensity, it incites, and this study explores how music of various arousal ratings can be used to influence players' choices in an interactive narrative role-playing game. We hypothesized that high-arousal music would influence players to exhibit avoidance behaviors in-game, and that low-arousal music would influence players to exhibit social behaviors. Experimentation showed that players were statistically significantly more likely to make avoidance behavior choices when high-arousal music was played. These findings are the first step into understanding how music can be used by game developers to influence player behaviors in interactive narrative games.
Music is hierarchically structured, both in how it is perceived by listeners and how it is composed. Such structure can be captured elegantly using probabilistic grammatical models similar to those used to study natural language. They address the complexity of the structure using abstract categories in a recursive formalism. Most existing grammatical models of musical structure focus on one single dimension of music--such as melody, harmony, or rhythm. While these grammar models often work well on short musical excerpts, accurate analysis of longer pieces requires taking into account the constraints from multiple domains of structure. The present paper proposes abstract product grammars--a formalism which integrates multiple dimensions of musical structure into a single grammatical model--along with efficient parsing and inference algorithms for this formalism. We use this model to study the combination of hierarchically-structured harmonic syntax and hierarchically-structured rhythmic information. The latter is modeled by a novel grammar of rhythm that is capable of expressing temporal regularities in musical phrases. It integrates grouping structure and meter. The combined model of harmony and rhythm outperforms both single-dimension models in computational experiments. All models are trained and evaluated on a treebank of hand-annotated Jazz standards.
The goal of this paper is to use all available Polish language data sets to seek the best possible performance in supervised sentiment analysis of short texts. We use text collections with labeled sentiment such as tweets, movie reviews and a sentiment treebank, in three comparison modes. In the first, we examine the performance of models trained and tested on the same text collection using standard cross-validation (in-domain). In the second we train models on all available data except the given test collection, which we use for testing (one vs rest cross-domain). In the third, we train a model on one data set and apply it to another one (one vs one cross-domain). We compare wide range of methods including machine learning on bag-of-words representation, bidirectional recurrent neural networks as well as the most recent pre-trained architectures ELMO and BERT. We formulate conclusions as to cross-domain and in-domain performance of each method. Unsurprisingly, BERT turned out to be a strong performer, especially in the cross-domain setting. What is surprising however, is solid performance of the relatively simple multinomial Naive Bayes classifier, which performed equally well as BERT on several data sets.
Several linguistic studies have shown the prevalence of various lexical and grammatical patterns in texts authored by a person of a particular gender, but models for part-of-speech tagging and dependency parsing have still not adapted to account for these differences. To address this, we annotate the Wall Street Journal part of the Penn Treebank with the gender information of the articles' authors, and build taggers and parsers trained on this data that show performance differences in text written by men and women. Further analyses reveal numerous part-of-speech tags and syntactic relations whose prediction performances benefit from the prevalence of a specific gender in the training data. The results underscore the importance of accounting for gendered differences in syntactic tasks, and outline future venues for developing more accurate taggers and parsers. We release our data to the research community.
Language models are a key technology in various tasks, such as, speech recognition and machine translation. They are usually used on texts covering various domains and as a result domain adaptation has been a long ongoing challenge in language model research. With the rising popularity of neural network based language models, many methods have been proposed in recent years. These methods can be separated into two categories: model based and feature based adaptation methods. Feature based domain adaptation has compared to model based domain adaptation the advantage that it does not require domain labels in the corpus. Most existing feature based adaptation methods are based on bias adaptation. We propose a novel feature based domain adaptation technique using hidden layer factorisation. This method is fundamentally different from existing methods because we use the domain features to calculate a linear combination of linear layers. These linear layers can capture domain specific information and information common to different domains. In the experiments, we compare our proposed method with existing adaptation methods. The compared adaptation techniques are based on two different ideas, that is, bias based adaptation and gating of hidden units. All language models in our comparison use state-of-the-art long short-term memory based recurrent neural networks. We demonstrate the effectiveness of the proposed method with perplexity results for the well-known Penn Treebank and speech recognition results for a corpus of TED talks.
In past studies, the few quantitative approaches to discourse structure were mostly confined to the presentation of the frequency of discourse relations. However, quantitative approaches should take into account both hierarchical and relational layers in the discourse structure. This study considers these factors and addresses the issue of how discourse relations and discourse units are related. It draws upon the available corpora of discourse structure (rhetorical structure theory-discourse treebank (RST-DT)) from a new perspective. Since an RST tree can be converted into a syntactic dependency tree, the data extracted from the RST-DT can be useful for calculating the discourse distance in much the same way as syntactic dependency distance is calculated. Discourse distance is also applicable to measuring the depth of the human processing of discourse. Furthermore, the data derived from the RST-DT are also easily converted into network data. This study finds that discourse structure has its discourse distance minimum and each type of RST relations has its range of discourse distance. The frequency distribution of discourse data basically follows the power law on several levels, while a network approach reveals how discourse units are arranged spatially in regular patterns. The two methods are mutually complementary in revealing the interaction between discourse relations and discourse units in a comprehensive manner, as well as in revealing how people process and comprehend discourse dynamically. Accordingly, we propose merging the two methods so as to yield a computational model for assessing discourse complexity and comprehension.
One of the central goals of Recurrent Neural Networks (RNNs) is to learn long-term dependencies in sequential data. Nevertheless, the most popular training method, Truncated Backpropagation through Time (TBPTT), categorically forbids learning dependencies beyond the truncation horizon. In contrast, the online training algorithm Real Time Recurrent Learning (RTRL) provides untruncated gradients, with the disadvantage of impractically large computational costs. Recently published approaches reduce these costs by providing noisy approximations of RTRL. We present a new approximation algorithm of RTRL, Optimal Kronecker-Sum Approximation (OK). We prove that OK is optimal for a class of approximations of RTRL, which includes all approaches published so far. Additionally, we show that OK has empirically negligible noise: Unlike previous algorithms it matches TBPTT in a real world task (character-level Penn TreeBank) and can exploit online parameter updates to outperform TBPTT in a synthetic string memorization task. Code availiable on github.
Building on recent work on capsule networks, we propose a new, general-purpose form of "routing by agreement" that activates output capsules in a layer as a function of their net benefit to use and net cost to ignore input capsules from earlier layers. To illustrate the usefulness of our routing algorithm, we present two capsule networks that apply it in different domains: vision and language. The first network achieves new state-of-the-art accuracy of 99.1% on the smallNORB visual recognition task with fewer parameters and an order of magnitude less training than previous capsule models, and we find evidence that it learns to perform a form of "reverse graphics." The second network achieves new state-of-the-art accuracies on the root sentences of the Stanford Sentiment Treebank: 58.5% on fine-grained and 95.6% on binary labels with a single-task model that routes frozen embeddings from a pretrained transformer as capsules. In both domains, we train with the same regime. Code is available at https://github.com/glassroom/heinsen_routing along with replication instructions.
<p>This dataset of Part-of-Speech (POS) tagged building codes contains 1,522 sentences from Chapters 5 and 10 of 2015 International Building Code. It adopts the original version of Penn Treebank tag set for the POS tags. It includes tagging results from 5 human annotators and 7 machine taggers. It also provides the most commonly chosen POS tag for each word by machine taggers and by human annotators. For detailed explanations of the meanings of the POS tags, please refer to <em>Building a Large Annotated Corpus of English: The Penn Treebank </em>[1]. For an explanation of the development of this dataset, please refer to the following paper [2].</p> <p>The authors would like to thank the National Science Foundation (NSF). This material is based on work supported by the NSF under Grant No. 1827733. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF.</p> <p><em>1. Marcus, Mitchell &amp; Ann Marcinkiewicz, Mary &amp; Santorini, Beatrice. (2002). Building a Large Annotated Corpus of English: The Penn Treebank. Computational Linguistics. 19. 313-330.</em></p> <p><em>2. Xue, X., and Zhang, J. (2019).&nbsp;&quot;Evaluation of Seven Part-of-Speech Taggers in Tagging Building Codes: Identifying the Best Performing Tagger and Common Sources of Errors.&quot;&nbsp;Proc.,&nbsp;ASCE&nbsp;Construction Research Congress,&nbsp;ASCE, Reston, VA, submitted.</em></p>
This paper focuses on the problem of unsupervised alignment of hierarchical\ndata such as ontologies or lexical databases. This is a problem that appears\nacross areas, from natural language processing to bioinformatics, and is\ntypically solved by appeal to outside knowledge bases and label-textual\nsimilarity. In contrast, we approach the problem from a purely geometric\nperspective: given only a vector-space representation of the items in the two\nhierarchies, we seek to infer correspondences across them. Our work derives\nfrom and interweaves hyperbolic-space representations for hierarchical data, on\none hand, and unsupervised word-alignment methods, on the other. We first\nprovide a set of negative results showing how and why Euclidean methods fail in\nthis hyperbolic setting. We then propose a novel approach based on optimal\ntransport over hyperbolic spaces, and show that it outperforms standard\nembedding alignment techniques in various experiments on cross-lingual WordNet\nalignment and ontology matching tasks.\n
Introduction: Bipolar disorder (BD) is an emotional disorder characterized by emotional instability. This study aims to investigate whether emotional disorders affect the cortical processing that regulates emotional stimuli. Methods: This study uses a passive viewing paradigm to compare event-related potentials (ERPs) of 21 subjects with bipolar disorder (BD), 25 subjects with schizophrenia (SZ), and 25 normal subjects (HC). Results: In the frontal and parietal areas (1) the late positive potential (LPP) amplitude of BD was greater than that of SZ and HC, and there was no difference in terms of peak latency. (2) In the occipital area, the amplitude and latency of P300 of SZ and BD were significantly larger than HC, but there was no difference between SZ and BD. (3) The behavioral results demonstrated that BDs’ valence rating on both positive and negative stimuli were significantly higher than those of HC and SZ when viewing face images, while BDs’ arousal scores for positive stimuli were significantly lower than HC and SZ. BDs’ arousal score for negative stimuli was significantly higher than those of HC and SZ. When viewing non-face pictures, BDs’ positive valence rating on negative stimuli and arousal scores for positive stimuli were significantly higher than those of HC and SZ. Conclusion: Bipolar patients have complete early emotional recognition, and their emotional abnormalities exist in the late evaluation of emotional stimuli, mainly occurring in the frontal and parietal areas.
The paper introduces the project of the Index Thomisticus Treebank (IT-TB). The IT-TB is a dependency-based treebank based on the corpus of the Index Thomisticus by father Roberto Busa (IT), which includes the opera omnia of Thomas Aquinas, for a total of approximately 11 million words. Currently, the IT-TB is the largest Latin treebank available, with more than 350,000 nodes in around 17,000 sentences. The annotation covers the entire books 1, 2 and 3 of Summa contra Gentiles, plus excerpts from Scriptum super Sententiis Magistri Petri Lombardi and Summa Theologiae. The paper details the multi-layer annotation style of the IT-TB and its background theoretical motivations. The conversion process to the now widely used Universal Dependencies style is described as well. Across more than a decade, the project has developed a number of linguistic resources and NLP tools for Latin connected to the IT-TB. As for the resources, the paper presents the syntaxbased subcategorization lexicon IT-VaLex and the valency lexicon Latin Vallex. As for the tools, the automatic dependency parsing process is described, highlighting the core issue of portability of NLP tools across the wide diachronic and diatopic span of Latin texts. A section is dedicated to automatic morphological analysis of Latin, introducing the analyzer Lemlat and its recent enhancement with information on derivational morphology and a new set of lexical entries covering a large Onomasticon (from Forcellini dictionary) and Medieval Latin (from Du Cange glossary).
This article focuses on the empirical experience and conclusions, resulting from the creation of language research and acquisition tools for Livonian – one of the smallest languages in Europe. A cluster was created for Livonian containing three interconnected databases, each with distinct types of data – lexical, morphological, and a corpus. The lexical database contains the lemmas and their data, the morphological database stores morphological forms, while all textual material, including the dictionary examples, is in the corpus. When indexing the corpus, every word refers to a lemma in the lexical database and its morphological information (new lemmas are added prior to indexation), ensuring consistency of the language data, and from each database the full data set of the other databases can be accessed. The function of each cluster is to extract the maximum amount of information from limited data sources. While technologies designed for languages with a large number of speakers focus on using quantitative methods and automation to extract qualitative information from a large and constantly expanding amount of linguistic data, the main function of technologies designed for small languages is to extract the same type of information from a limited and largely static data set. This article also examines a string of problems faced when working with a small amount of resources (inadequate language data, insufficient personnel, lack of rules for automating processes, etc.) and methods for resolving these problems in the case of Livonian.
Dependency grammar induction is the task of learning dependency syntax without annotated training data. Traditional graph-based models with global inference achieve state-ofthe-art results on this task but they require O(n3) run time. Transition-based models enable faster inference with O(n) time complexity, but their performance still lags behind. In this work, we propose a neural transition-based parser for dependency grammar induction, whose inference procedure utilizes rich neural features with O(n) time complexity. We train the parser with an integration of variational inference, posterior regularization and variance reduction techniques. The resulting framework outperforms previous unsupervised transition-based dependency parsers and achieves performance comparable to graph-based models, both on the English Penn Treebank and on the Universal Dependency Treebank. In an empirical comparison, we show that our approach substantially increases parsing speed over graphbased models.
This paper presents a full procedure for the development of a segmented, POS-tagged and chunk-parsed corpus of Old Tibetan. As an extremely lowresource language, Old Tibetan poses non-trivial problems in every step towards the development of a searchable treebank. We demonstrate, however, that a carefully developed, semisupervised method of optimising and extending existing tools for Classical Tibetan, as well as creating specific ones for Old Tibetan, can address these issues. We thus also present the very first Tibetan Treebank in a variety of formats to facilitate research in the fields of NLP, historical linguistics and Tibetan Studies.
This study focuses on a comprehensive analysis and manual re-annotation of the Turkish IMST-UD Treebank, which was automatically converted from the IMST Treebank (Sulubacak et al., 2016b). In accordance with the Universal Dependencies' guidelines and the necessities of Turkish grammar, the existing treebank was revised. The current study presents the revisions that were made alongside the motivations behind the major changes. Moreover, it reports the parsing results of a transition-based dependency parser and a graph-based dependency parser obtained over the previous and updated versions of the treebank. In light of these results, we have observed that the re-annotation of the Turkish IMST-UD treebank improves performance with regards to dependency parsing.
In this paper, we will present an approach in transforming Serbian language Morphological dictionaries from a DELA text format to a lexical database dubbed Leximirka. Considering the benefits of storing data within a database when compared to storing them in textual documents, we will outline some of the functionality that the database has made possible. We will also show how hand-made rules that use category labels lexical entries are marked with can be used to link lexical entries. The initial morphological dictionaries were Serbian Morphological Dictionaries. However, we will show multilingual application of Leximirka using French Morphological Dictionaries.
Head-driven phrase structure grammar (HPSG) enjoys a uniform formalism representing rich contextual syntactic and even semantic meanings. This paper makes the first attempt to formulate a simplified HPSG by integrating constituent and dependency formal representations into head-driven phrase structure. Then two parsing algorithms are respectively proposed for two converted tree representations, division span and joint span. As HPSG encodes both constituent and dependency structure information, the proposed HPSG parsers may be regarded as a sort of joint decoder for both types of structures and thus are evaluated in terms of extracted or converted constituent and dependency parsing trees. Our parser achieves new state-of-the-art performance for both parsing tasks on Penn Treebank (PTB) and Chinese Penn Treebank, verifying the effectiveness of joint learning constituent and dependency structures. In details, we report 95.84 F1 of constituent parsing and 97.00% UAS of dependency parsing on PTB.
The dorsolateral prefrontal cortex (DLPFC) plays a key role in the modulation of affective processing. However, its specific role in the regulation of neurocognitive processes underlying the interplay of affective perception and visual awareness has remained largely unclear. Using a mixed factorial design, this study investigated effects of inhibitory continuous theta-burst stimulation (cTBS) of the right DLPFC (rDLPFC) compared to an Active Control condition on behavioral (N=48) and electroencephalographic (N=38) correlates of affective processing in healthy Chinese participants. Event-related potentials (ERPs) in response to passively viewed subliminal and supraliminal negative and neutral natural scenes were recorded before and after cTBS application. We applied minimum-norm approaches to estimate the corresponding neuronal sources. On a behavioral level, we found evidence for reduced emotional interference by, and less negative and arousing ratings of negative supraliminal stimuli following rDLPFC inhibition. We found no evidence for stimulation effects on self-reported mood or the behavioral discrimination of subliminal stimuli. On a neurophysiological level, rDLPFC inhibition relatively enhanced occipito-parietal brain activity for both subliminal and supraliminal negative compared to neutral images (112-268ms; 320-380ms). The early onset and localization of these effects suggests that rDLPFC inhibition boosts automatic processes of ‘emotional attention’ independently of visual awareness. Further, our study reveals the first available evidence for a differential influence of rDLPFC inhibition on subliminal versus supraliminal neural emotion processing. Explicitly, our findings indicate that rDLPFC inhibition selectively enhances rather late (292-360ms) activity in response to supraliminal negative images. We tentatively suggest that this differential frontal activity likely reflects enhanced awareness-dependent down-regulation of negative scene processing eventually leading to facilitated disengagement from and less negative and arousing evaluations of negative supraliminal stimuli.
Frequency distribution of words, syntax and semantics in many languages abides by certain laws. However, because of the shortage of discourse corpora, few studies have examined whether the frequency of discourse relations follows some distributional patterns. Although there is some research based on the Rhetorical Structure Theory discourse treebank (RST-DT), each of these studies is limited to a single language. Otherwise to the RST-DT, the Penn Discourse Treebank (PDTB), adopting another annotation system, has had an enormous influence on the study of discourse structure and discourse annotation. Discourse corpora in other languages, such as Chinese, Hindi, Turkish, Czech and Arabic have been annotated following PDTB style. With the data from these discourse treebanks, we find that the rank-frequency of discourse relations follow the same pattern and that these languages share significant similarities in using semantic relations to organize the discourse. It is evidenced in our research that humans assume the relationship between two consecutive sentences is a causal connection or expansion link for fewer connectives used, but the relation of contrast is the most marked by connectives. This research will be of significance for understanding the homogeneity of discourse structure across languages.
Recurrent neural networks (RNNs) provide excellent performance on applications with sequential data such as speech recognition. On-chip implementation of RNNs is difficult due to the significantly large number of parameters and computations. In this work, we first present a training method for LSTM model for language modeling on Penn Treebank dataset with binary weights and multi-bit activations and then map it onto a fully parallel RRAM array architecture ("XNOR-RRAM"). An energy-efficient XNOR-RRAM array based system for LSTM RNN is implemented and benchmarked on Penn Treebank dataset. Our results show that 4-bit activation precision can provide a near-optimal perplexity of 115.3 with an estimated energy-efficiency of ~27 TOPS/W.
Det är lättare att minnas emotionellt laddad information. Även om minnet ofta försämras vid normalt åldrande, kvarstår den emotionella förstärkningseffekten. Studier antyder att unga vuxna minns bättre negativt än positivt material, men att minnespreferensen för negativt material minskar eller t.o.m. ersätts av en preferens för positivt material med åldern. I avhandlingen undersöktes hur unga och äldre friska vuxna personer minns ord med olika emotionellt värde (positiv/negativ) och emotionell intensitet. Även underliggande mekanismer i hjärnan hos äldre vuxna utforskades. Avhandlingen visade att det inte fanns skillnader mellan hur unga (21-35 år) och äldre (50-79 år) vuxna minns emotionella ord. Här sågs alltså ingen skillnad i minnespreferens för negativa eller positiva ord eller för ord med olika stark emotionell intensitet mellan åldersgrupperna. I avhandlingen identifierades samband mellan minne för positiva ord och gråsubstansvolym i hjärnans bakre delar hos friska äldre vuxna. Bättre minne för positiva ord hade samband med lägre integritet hos vitsubstansbanor i hjärnan, vilket antyder att en kompensationsmekanism kunde spela en roll. Avhandlingen visade även att språkliga och kulturella faktorer har större betydelse än ålder och kön för bedömningen av emotionella egenskaper hos ord. ------------------------------------- On helpompi muistaa tunnesisältöistä materiaalia. Vaikka muisti usein heikkenee normaalissa ikääntymisessä, tunnesisällön muistia vahvistava vaikutus pysyy. Tutkimukset osoittavat, että nuoret aikuiset muistavat kielteistä materiaalia myönteistä paremmin, mutta että kielteisen materiaalin suosiminen vähenee tai jopa vaihtuu myönteisen materiaalin suosimiseksi ikääntyessä. Väitöskirjassa tutkittiin miten nuoret ja ikääntyvät terveet aikuiset muistavat erityyppisiä tunnesisältöisiä sanoja. Selvitettiin myös tunnemuistin aivostollista perustaa ikääntyvillä aikuisilla. Väitöskirja osoitti, ettei nuorten (21-35 v.) ja ikääntyvien (50-79 v.) aikuisten välillä ollut eroja tunnesisältöisten sanojen muistamisessa. Tässä ei siis havaittu ikään liittyvää eroa kielteisten tai myönteisten sanojen tai eri tunneintensiteetillä varustettujen sanojen suosimisessa. Väitöskirjassa havaittiin yhteys myönteisten sanojen muistamisen ja aivojen takaosien harmaan aineen volyymin välillä terveillä ikääntyvillä aikuisilla. Mitä paremmin myönteisiä sanoja muistettiin, sen matalampi valkean aineen radastojen integriteetti oli normaalisti ikääntyvillä, mikä voi viitata korvaavan mekanismin vaikutukseen. Väitöskirja osoitti myös, että kielellisillä ja kulttuurillisilla tekijöillä on suurempi merkitys kuin iällä tai sukupuolella sanojen tunnesisällön arvioinnissa.
The multiple state theory of working memory suggests that representations held in working memory are separated into two states: a currently-relevant active representation and accessory memory items held for future use. While the characteristics and consequences of active versus accessory states have been the subject of several investigations, the exact neurocognitive mechanisms that move representations between the two states remain unclear. Of the two competing hypotheses, one suggests that inhibition is applied to keep a representation in an accessory state, while the other suggests that accessory representations simply receive less top-down cortical amplification than active representations, but are not subjected to inhibition. Here we capitalize on the different affective consequences for stimuli whose memory representations are subjected to inhibition (negative ratings) or active enhancement (positive ratings) to test these competing hypotheses. On each trial participants memorized four items and then were cued to focus on a single item within memory. They then completed either a visual search or an affective evaluation task. Search times were slower when a search distractor matched the colour of the active item but not when it matched the colour of the accessory item, replicating findings of a division in working memory whereby only active items guide attention. Also, accessory items were affectively devalued compared to baseline and active memory items. This finding of devaluation supports the hypothesis that inhibition is used to keep representations in an accessory state, and adds to past findings that similar mechanisms of attention and emotion govern prioritization in working memory and the prioritization of external stimuli. Meeting abstract presented at VSS 2018
Despite the significant improvement of datadriven dependency parsing systems in recent years, they still achieve a considerably lower performance in parsing spoken language data in comparison to written data. On the example of Spoken Slovenian Treebank, the first spoken data treebank using the UD annotation scheme, we investigate which speechspecific phenomena undermine parsing performance, through a series of training data and treebank modification experiments using two distinct state-of-the-art parsing systems. Our results show that utterance segmentation is the most prominent cause of low parsing performance, both in parsing raw and pre-segmented transcriptions. In addition to shorter utterances, both parsers perform better on normalized transcriptions including basic markers of prosody and excluding disfluencies, discourse markers and fillers. On the other hand, the effects of written training data addition and speech-specific dependency representations largely depend on the parsing system selected.
There has been a vast development of personal informatics devices combining sleep monitoring with alarm systems, in order to find an optimal time to awaken a sleeping person in a pleasant way. Most of these systems implement auditory feedback, which is not always pleasant and may disturb other sleepers. We present an adaptive alarm system that detects sleeping cycles and triggers alarm signal during shallow sleep, to minimize sleep inertia. Since tactile sensation is associated with positive valence, vibrotactile stimulation is investigated as a silent alarm to enhance pleasant awakening. Three modulation techniques to render the tactile stimuli for pleasant awakening are considered, namely simultaneous, continuous, and successive stimulation. Two experimental studied are conducted. Experiment 1 studied exogenous attention towards tactile stimulation in a multimodal scenario (involving visual and haptic interactions) with fully awake individuals. Results from the attention task and the subjective valence rating suggest that the vibrotactile stimulation should be based on the continuous modulation, since this not only is very perceivable but also associated with positive attention. Experiment 2 evaluated the user experience with tactile stimulation patterns during sleep. Results confirmed the findings of experiment 1. Continuous modulation was rated highest for pleasant yet arousing sleep-awake transition.
PURPOSE: To assess postmatch perceived exertion, feeling, and wellness according to the match outcome (winning, drawing, or losing) in professional soccer players. METHODS: In total, 12 outfield players were followed during 52 official matches where the outcomes (win, draw, or lose) were noted. Following each match, players completed both a 10-point Borg scale modified by Foster and an 11-point Hardy and Rejeski scale rating of perceived feeling. Rating of perceived sleep quality, stress, fatigue, and muscle soreness was collected separately on a 7-point scale the day following each match. RESULTS: Player rating of perceived exertion was higher by a very large magnitude following a loss compared with a draw or a win and higher by a small magnitude after a draw compared with a win. Players felt more pleasure after a win compared with a draw or loss and more displeasure after a loss compared with draw. The players reported a largely and moderately better perceived sleep quality, less stress, and fatigue following a win compared with a draw or a loss and a moderately bad perceived sleep quality, higher stress, and fatigue following a draw compared with a loss. In contrast, only a trivial-small change was observed in perceived muscle soreness between all outcomes. CONCLUSION: Match outcomes moderately to largely affect rating of perceived exertion, feeling, sleep quality, stress, and fatigue, whereas perceived muscle soreness remains high regardless of the match outcome. However, winning a match decreases the strain and improves both pleasure and wellness in professional soccer players.
Wikipedia has a strong norm of writing in a "neutral point of view" (NPOV). Articles that violate this norm are tagged, and editors are encouraged to make corrections. But the impact of this tagging system has not been quantitatively measured. Does NPOV tagging help articles to converge to the desired style? Do NPOV corrections encourage editors to adopt this style? We study these questions using a corpus of NPOV-tagged articles and a set of lexicons associated with biased language. An interrupted time series analysis shows that after an article is tagged for NPOV, there is a significant decrease in biased language in the article, as measured by several lexicons. However, for individual editors, NPOV corrections and talk page discussions yield no significant change in the usage of words in most of these lexicons, including Wikipedia's own list of "words to watch." This suggests that NPOV tagging and discussion does improve content, but has less success enculturating editors to the site's linguistic norms.
Nous nous intéressons à la performance du langage et aux contraintes causées par la capacité cognitive. Miller (1956) a montré que la mémoire à court-terme est limitée à 7±2 éléments, mais aujourd’hui, cette limite est actualisée à 4 selon Cowan (2001). La première question que nous nous posons est: comment mesurer la complexité syntaxique à partir des données observables considérant ce contrainte de la mémoire à court-terme? Deux autres questions que nous voudrions résoudre sont: quelles sont les relations syntaxiques qui créent vraiment de la complexité? Lorsqu’il s’agit d’une structure syntaxique plus (ou moins) complexe, quels sont les phénomènes qu’on pourrait observer dans nos données?Pour répondre à la première question, je me concentre sur des différents mesures du flux de dépendances (Kahane et al., 2017; Yan & Kahane, 2018). Notre hypothèse est que la mesure du flux à partir de treebanks annotés en syntaxe de dépendance permettraient de vérifier les hypothèses sur la mémoire à court-terme. Les deux autres questions alimentent mon actuel travail de thèse sur les configurations du flux de dépendance, ainsi que les liens entre flux et structure prosodique.
In this paper, we explore the linguistic factors that influence an author's choice of discourse connectives in the production of a coherent text. We focus on the competition between so-called primary connectives (grammaticalized and mostly one-word expressions such as therefore) and secondary connectives (not yet fully grammaticalized compositional discourse phrases such as for this reason). We attempt to describe the linguistic constraints on and preferences in connective selection. The analysis is based on manually annotated data from the Prague Discourse Treebank 2.0 (PDiT), which contains almost 50000 sentences from Czech newspaper texts. We demonstrate that discourse connectives are used in accordance with the economy principle in language, i.e. authors aim to achieve the maximal result with minimal effort. They most frequently choose short and semantically more generalized primary connectives. However, in cases where the discourse relations can be misunderstood, authors prefer more complex and specific structures.
Abstract This paper presents a treebank-based study of the effect the text form (prose vs. verse) has on the course of two grammatical changes in Medieval French: the loss of null subjects and the loss of OV word order. By means of statistical analysis, we demonstrate that naive estimates of the spread of overt subjects and VO orders give the impression that there is a significant difference between the rates of development in prose vs. verse. By contrast, estimates based on an abstract grammar competition model which distinguishes between grammar-ambiguous surface forms (overt personal subjects, null subjects in coordination contexts) and grammar-unambiguous surface forms (overt expletive subjects, null subjects in non-coordination contexts) show prose-verse parallelism, prose having an earlier change onset, in line with traditional intuitions. At a more general level, these results suggest that the product of the interaction of a particular grammar with universal pragmatic laws is constant, which can be observed if the factors responsible for variation in grammatical choices are controlled for.
In situations of real threat, showing a fear reaction makes sense, thus, increasing the chance to survive. The question is, how could anybody differentiate between a real and an apparent threat? Here, the slogan counts “better safe than sorry”, meaning that it is better to shy away once too often from nothing than once too little from a real threat. Furthermore, in a complex environment it is adaptive to generalize from one threatening situation or stimulus to another similar situation/stimulus. But, the danger hereby is to generalize in a maladaptive manner involving as it is to strong and/or fear too often “harmless” (safety) situations/stimuli, as it is known to be a criterion of anxiety disorders (AD). Fear conditioning and fear generalization paradigms are well suited to investigate fear learning processes. It is remarkable that despite increasing interest in this topic there is only little research on fear generalization. Especially, most research on human fear conditioning and its generalization has focused on adults, whereas only little is known about these processes in children, even though AD is typically developing during childhood. To address this knowledge gap, four experiments were conducted, in which a discriminative fear conditioning and generalization paradigm was used. In the first two experiments, developmental aspects of fear learning and generalization were of special interest. Therefore, in the first experiment 267 children and 285 adults were compared in the differential fear conditioning paradigm and generalization test. Skin conductance responses (SCRs) and ratings of valence and arousal were obtained to indicate fear learning. Both groups displayed robust and similar differential conditioning on subjective and physiological levels. However, children showed heightened fear generalization compared to adults as indexed by higher arousal ratings and SCRs to the generalization stimuli. Results indicate overgeneralization of conditioned fear as a developmental correlate of fear learning. The developmental change from a shallow to a steeper generalization gradient is likely related to the maturation of brain structures that modulate efficient discrimination between threatening and (ambiguous) safety cues. The question hereby is, at which developmental stage fear generalization gradients of children adapt to the gradients of adults. Following up on this question, in a second experiment, developmental changes in fear conditioning and fear generalization between children and adolescents were investigated. According to experiment 1 and previous studies in children, which showed changes in fear learning with increasing age, it was assumed that older children were better at discriminating threat and safety stimuli. Therefore, 396 healthy participants (aged 8 to 12 years) were examined with the fear conditioning and generalization paradigm. Again, ratings of valence, arousal, and SCRs were obtained. SCRs indicated differences in fear generalization with best fear discrimination in 12-year-old children suggesting that the age of 12 years seems to play an important role, since generalization gradients were similar to that of adults. These age differences were seen in boys and girls, but best discrimination was found in 12-year-old boys, indicating different development of generalization gradients according to sex. This result fits nicely with the fact that the prevalence of AD is higher in women than in men. In a third study, it was supposed that the developmental trajectory from increased trait anxiety in childhood to manifest AD could be mediated by abnormal fear conditioning and generalization processes. To this end, 394 children aged 8 to 12 years with different scores in trait anxiety were compared with each other. Results provided evidence that children with high trait anxiety showed stronger responses to threat cues and impaired safety signal learning contingent on awareness as indicated by arousal at acquisition. Furthermore, analyses revealed that children with high trait anxiety showed overall higher arousal ratings at generalization. Contrary to what was expected, high trait anxious children did not show significantly more fear generalization than children with low trait anxiety. However, high-trait-anxious (HA) participants showed a trend for a more linear gradient, whereas moderate-trait-anxious (MA) and low-trait-anxious (LA) participants showed more quadratic gradients according to arousal. Additionally, after controlling for age, sex and negative life experience, SCR to the safety stimulus predicted the trait anxiety level of children suggesting that impaired safety signal learning may be a risk factor for the development of AD. Results provide hints that frontal maturation could develop differently according to trait anxiety resulting in different stimuli discrimination. Thus, in a fourth experiment, 40 typically developing volunteers aged 10 to 18 years were screened for trait anxiety and investigated with the differential fear conditioning and generalization paradigm in the scanner. Functional magnetic resonance imaging (fMRI) were used to identify the neural mechanisms of fear learning and fear generalization investigating differences in this neural mechanism according to trait anxiety, developmental aspects and sex. At acquisition, HA participants showed reduced activation in frontal brain regions, but at generalization, HA participants showed an increase in these frontal regions with stronger linear increase in activation with similarity to CS+ in HA when compared to LA participants. This indicates that there is a hyper-regulation in adolescents to compensate the higher difficulties at generalization in form of a compensatory mechanism, which decompensates with adulthood and/or may be collapsed in manifest AD. Additionally, significant developmental effects were found: the older the subjects the stronger the hippocampus and frontal activation with resemblance to CS+, which could explain the overgeneralization of younger children. Furthermore, there were differences according to sex: males showed stronger activation with resemblance to CS+ in the hippocampus and frontal regions when compared to females fitting again nicely with the observation that prevalence rates for AD are higher for females than males. In sum, the studies suggest that investigating developmental aspects of (maladaptive) overgeneralization may lead to better understanding of the mechanisms of manifest anxiety disorders, which could result in development and provision of prevention strategies. Although, there is need for further investigations, the present work gives some first hints for such approaches.
Computational text-level discourse analysis mostly happens within Rhetorical Structure Theory (RST), whose structures have classically been presented as constituency trees, and relies on data from the RST Discourse Treebank (RST-DT); as a result, the RST discourse parsing community has largely borrowed from the syntactic constituency parsing community. The standard evaluation procedure for RST discourse parsers is thus a simplified variant of PARSEVAL, and most RST discourse parsers use techniques that originated in syntactic constituency parsing. In this article, we isolate a number of conceptual and computational problems with the constituency hypothesis. We then examine the consequences, for the implementation and evaluation of RST discourse parsers, of adopting a dependency perspective on RST structures, a view advocated so far only by a few approaches to discourse parsing. While doing that, we show the importance of the notion of headedness of RST structures. We analyze RST discourse parsing as dependency parsing by adapting to RST a recent proposal in syntactic parsing that relies on head-ordered dependency trees, a representation isomorphic to headed constituency trees. We show how to convert the original trees from the RST corpus, RST-DT, and their binarized versions used by all existing RST parsers to head-ordered dependency trees. We also propose a way to convert existing simple dependency parser output to constituent trees. This allows us to evaluate and to compare approaches from both constituent-based and dependency-based perspectives in a unified framework, using constituency and dependency metrics. We thus propose an evaluation framework to compare extant approaches easily and uniformly, something the RST parsing community has lacked up to now. We can also compare parsers’ predictions to each other across frameworks. This allows us to characterize families of parsing strategies across the different frameworks, in particular with respect to the notion of headedness. Our experiments provide evidence for the conceptual similarities between dependency parsers and shift-reduce constituency parsers, and confirm that dependency parsing constitutes a viable approach to RST discourse parsing.
English part-of-speech taggers regularly make egregious errors related to noun-verb ambiguity, despite having achieved 97%+ accuracy on the WSJ Penn Treebank since 2002. These mistakes have been difficult to quantify and make taggers less useful to downstream tasks such as translation and text-to-speech synthesis. This paper creates a new dataset of over 30,000 naturally-occurring non-trivial examples of noun-verb ambiguity. Taggers within 1% of each other when measured on the WSJ have accuracies ranging from 57% to 75% accuracy on this challenge set. Enhancing the strongest existing tagger with contextual word embeddings and targeted training data improves its accuracy to 89%, a 14% absolute (52% relative) improvement. Downstream, using just this enhanced tagger yields a 28% reduction in error over the prior best learned model for homograph disambiguation for textto-speech synthesis.
The paper analyses a corpus of metalinguistic texts (grammars, remarks) from the first grammar of French (Palsgrave 1972 [1530]) until today in order to retrace the metalinguistic perceptions and the prescriptive interventions in the domain of the French adverb. It is shown that adverbs ending in <italic>-ment</italic> and paraphrases were favored to the detriment of the old oral tradition of using short adverbs. In the second part, three methods are suggested for measuring the effect metalinguistic considerations and norms have on use. This innovative approach aims at bringing together the diachrony of grammaticography with the diachrony of use. In fact, a more objective approach is crucially needed in the domain of linguistic norm where speculation, polemic argumentation and ideology prevail.
Grammar induction is the task of learning syntactic structure without the expert-labeled treebanks (Charniak and Carroll, 1992; Klein and Manning, 2002). Recent work on latent tree learning offers a new family of approaches to this problem by inducing syntactic structure using the supervision from a downstream NLP task (Yogatama et al., 2017; Maillard et al., 2017; Choi et al., 2018). In a recent paper published at ICLR, Shen et al. (2018) introduce such a model and report near state-of-the-art results on the target task of language modeling, and the first strong latent tree learning result on constituency parsing. During the analysis of this model, we discover issues that make the original results hard to trust, including tuning and even training on what is effectively the test set. Here, we analyze the model under different configurations to understand what it learns and to identify the conditions under which it succeeds. We find that this model represents the first empirical success for neural network latent tree learning, and that neural language modeling warrants further study as a setting for grammar induction.
Sentiment analysis of short texts such as single sentence has been a research hotspot of naturallanguage processing (NLP). There still exists the challenge of effectively handling the problem with limited contextual information and semantic features. Hence, in this paper, an attention-based bidirectional LSTM neural network (AB-BiLSTM) is proposed to solve the problem. The proposed model can attend the qualitative and informative parts and learn semantic features from both directions of a sentence to perform sentiment analysis of short texts. The proposed model mainly contributes to take advantage of the attention mechanism to capture the informative parts of a sentence without any syntactic features and lexicon features. The model is conducted on the Stanford Sentiment Treebank dataset and the Movie Review Data provided by Cornell University for single sentence sentiment analysis of binary classification. The experimental results indicate that the proposal in this study has superior performances over theexisting methods, without taking the attention mechanism into account.
The contents and structure of semantic memory have been the focus of much recent research, with major advances in the development of distributional models, which use word co-occurrence information as a window into the semantics of language. In parallel, connectionist modeling has extended our knowledge of the processes engaged in semantic activation. However, these two lines of investigation have rarely been brought together. Here, we describe a processing model based on distributional semantics in which activation spreads throughout a semantic network, as dictated by the patterns of semantic similarity between words. We show that the activation profile of the network, measured at various time points, can successfully account for response times in lexical and semantic decision tasks, as well as for subjective concreteness and imageability ratings. We also show that the dynamics of the network is predictive of performance in relational semantic tasks, such as similarity/relatedness rating. Our results indicate that bringing together distributional semantic networks and spreading of activation provides a good fit to both automatic lexical processing (as indexed by lexical and semantic decisions) as well as more deliberate processing (as indexed by ratings), above and beyond what has been reported for previous models that take into account only similarity resulting from network structure.
This paper describes our approach to developing the Turkish PropBank by adopting the semantic role-labeling guidelines of the original PropBank and using the translation of the English Penn-TreeBank as a resource. We discuss the semantic annotation process of the PropBank and language-specific cases for Turkish, the tools we have developed for annotation, and quality control for multiuser annotation. In the current phase of the project, more than 9500 sentences are semantically analyzed and predicate-argument information is extracted for 1330 verbs and 1914 verb senses. Our plan is to annotate 17,000 sentences by the end of 2017.
A dependency parser generates both a syntactic structure and a shallow semantic structure of a sentence. It is a fundamental component of natural language processing (NLP) based pipelines, which are critical to facilitate research using the Electronic Health Records (EHR). However, current works mainly apply parsers developed in the general English domain to clinical text. There are no formal evaluations and comparisons of deep learning based dependency parsers in the medical domain. No state-of-the-art dependency parsing performance has been established on clinical text, either. In this study, we investigated the performance of four state-ofthe-art deep learning based dependency parsers, Stanford parser, Bist-parser, dependency_tf parser and jPTDP parser, respectively. Experiments for evaluation are conducted on two datasets: (1) The MiPACQ Treebank and (2) A Treebank of progress notes. Our results showed that the original parsers achieved lower performance in clinical text compared to general English text. After retraining on the clinical Treebank, all parsers obtained better performance. Besides, using word embeddings from Gigaword and MIMICIII yielded comparable performance. Interestingly, the transition-based parsers demonstrated stronger generalizability on different treebanks than the graph-based parsers. Overall, Bist-parser achieved the best performance on MiPACQ (88.95% UAS, 92.69% LS, 86.10% LAS). Stanford parser achieved the best performance on progress notes (84.01% UAS, 89/97% LS, 80.72% LAS).
We propose a novel approach to Vietnamese word segmentation. Our approach is based on the Single Classification Ripple Down Rules methodology (Compton and Jansen, 1990), where rules are stored in an exception structure and new rules are only added to correct segmentation errors given by existing rules. Experimental results on the benchmark Vietnamese treebank show that our approach outperforms previous state-of-the-art approaches JVnSegmenter, vnTokenizer, DongDu and UETsegmenter in terms of both accuracy and performance speed. Our code is open-source and available at: https://github.com/datquocnguyen/RDRsegmenter.