Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
16504 papers
Images and language convey meaning that depend on the viewpoints and contextual background of those perceiving them. Taxonomies help order meaning such that information granules, image elements and words, make sense in relation to one another and to their mutual global context. In linguistics, hypernyms cover semantically broader context then their subordinate hyponym. In images, superordinate spatial-taxons (object groups or foreground) cover more abstract regions then their child subordinate spatial-taxons (objects or salient object parts). In this paper I use fuzzy granularization and fuzzy perceptualization as proposed by Zadeh 2002 to explore image annotation by using Zadeh's Restriction-centered Theory of Truth and Meaning as proposed in 2013. The approach uses human annotated image data, search engine queries and data collected from WordNet (A Lexical Database for English maintained by Princeton University). I discuss implications for Shannon, Integrated, and Zadeh Information Theory.
This article explores whether and how network visualization can benefit philological and historical-linguistic study. This is illustrated with a corpus-based investigation of scribes' language use in a lemmatized and morphologically annotated corpus of documentary Latin (Late Latin Charter Treebank, LLCT2). We extract four continuous linguistic variables from LLCT2 and utilize a gradient colour palette in Gephi to visualize the variable values as node attributes in a trimodal network which consists of the documents, writers, and writing locations underlying the same corpus. We call this network the "LLCT2 network". The geographical coordinates of the location nodes form an approximate map, which allows for drawing geographical conclusions. The linguistic variables are examined both separately and as a sum variable, and the visualizations presented as static images and as interactive Sigma.js visualizations. The variables represent different domains of language competence of scribes who learnt written Latin practically as a second-language. The results show that the network visualization of linguistic features helps in observing patterns which support linguistic-philological argumentation and which risk passing unnoticed with traditional methods. However, the approach is subject to the same limitations as all visualization techniques: the human eye can only perceive a certain, relatively small amount of information at a time.
Word vectors are at the core of many natural language processing tasks. Recently, there has been interest in post-processing word vectors to enrich their semantic information. In this paper, we introduce a novel word vector post-processing technique based on matrix conceptors (Jaeger2014), a family of regularized identity maps. More concretely, we propose to use conceptors to suppress those latent features of word vectors having high variances. The proposed method is purely unsupervised: it does not rely on any corpus or external linguistic database. We evaluate the post-processed word vectors on a battery of intrinsic lexical evaluation tasks, showing that the proposed method consistently outperforms existing state-of-the-art alternatives. We also show that post-processed word vectors can be used for the downstream natural language processing task of dialogue state tracking, yielding improved results in different dialogue domains.
Language regulation, as Hynninen explains it in this volume, ‘is a concept that is intended to depict the kind of language use that may lead to formation of (new) norms’ (p. 20). This concept draws attention to the usage of language in micro-level communication and its impact on norms development. Hynninen’s study explores the phenomenon of language regulation in spoken academic settings using English as a Lingua Franca (henceforth ELF), looking at it from both interactional and ideological perspectives. An important distinction in this context was made by Bartsch (1987), who distinguishes between grammatically-conforming language use and acceptable language use. The former is measured against established linguistic norms, whereas the latter is driven by the interactive norm to accomplish mutual understanding. This means that interaction becomes a site for negotiating norms, in the sense that if deviation from what is grammatically correct occurs repeatedly over time and is accepted in talk, new norms may be formed. In this book Hynninen constructs language norms as acceptable linguistic conduct, focusing particularly on the negotiation of acceptable language in interaction. Nevertheless, she does not restrict the exploration of language norms only to interactional behaviours of regulatory negotiations, as research indicates that the conformity to a norm is influenced by expectations of language use. In addition, we also know that speakers’ beliefs about what is correct usage of language is not necessarily reflected in their actual linguistic behaviour. This calls for a binary approach in exploring language norms. In this light, Hynninen conceptualises language norms as a combination of communicative linguistic behaviours (the interactive dimension of the study) and speakers’ beliefs about and expectations of others’ behaviours (the ideological element studied in this research).
This paper describes a collection of modules for Italian language processing based on CoreNLP and Universal Dependencies (UD). The software will be freely available for download under the GNU General Public License (GNU GPL). Given the flexibility of the framework, it is easily adaptable to new languages provided with an UD Treebank.
The so-called epoch of the Internet, which has replaced the epoch of book, change dramatically the Russian literary language. That is to say, the language of media obtained the leading position amongst other language variations. The language of media is becoming a model language of the society and reflecting the current language practice. The media language represents the city linguistic norms and is focused on modernization (literary linguistic norms, in particular). This article is to set the hypotheses that there are two main form of the national modern literary language, existing in the modern information society. They are: 1) the elite literary language; 2) the mass literary language (and its implementation in the media as an option, which we mean “media language”). In mass-media communication, focused on mass media mind, literary standards are on the periphery of media content. Standard standards are superseded by norms of media communication, which draw their resources from the mass language of the whole society and themselves form this mass language. The present hypothesis is based on three fundamental principles: 1) the change of social status of modern media; 2) the definition of mass media language as common national language; 3) the concept of mediatization, which we develop in correlation to mass media language. This article concludes that traditional language norms are challenged by new technologies and are not relevant to speaking practice of mass media.
There are many mechanisms to sense arousal. Most of them are either intrusive, prone to bias, costly, require skills to set-up or do not provide additional context to the user's measure of arousal. We present arousal detection through the analysis of pupillary response from eye trackers. Using eye-trackers, the user's focal attention can be detected with high fidelity during user interaction in an unobtrusive manner. To evaluate this, we displayed twelve images of varying arousal levels rated by the International Affective Picture System (IAPS) to 41 participants while they reported their arousal levels. We found a moderate correlation between the self-reported arousal and the algorithm's arousal rating, r(47)=0.46, p<.01. The results show that eye trackers can serve as a multi-sensory device for measuring arousal, and relate the level of arousal to the user's focal attention. We anticipate that in the future, high fidelity web cameras can be used to detect arousal in relation to user attention, to improve usability, UX and understand visual behaviour.
Language Modeling (LM) is a subtask in Natural Language Processing (NLP), and the goal of LM is to build a statistical language model that can learn and estimate a probability distribution of natural language over sentences of terms. Recently, many recurrent neural network based LM, a type of deep neural network for dealing with sequential data, have been proposed and achieved remarkable results. However, they only rely upon the analysis on the words occurred in the sentences even though every sentence contains various useful morphological information, such as Part-of-Speech (POS) tag that is necessary for constituting a sentence and can be used for an analysis as a feature. Although morphological information can be useful for LM, using that information as the input data to neural network based LM is not straightforward because adding features between words as a one-dimensional array can cause the vanishing gradient problem by increasing the time steps of recurrent neural network. In order to solve this problem, in this paper, we propose a CNN-LSTM based language model that deals with textual data regarding a multi-dimensional data with respect to the input of the network. To train this multi-dimensional input to Long-Short Term Memory (LSTM), we use a convolutional neural network (CNN) with a 1×1 filter for dimensionality reduction of input data to avoid the vanishing gradient problem by decreasing the time step between input words. In addition, our approach that uses multi-dimension data reduced by CNN can be used as a plugin with many customized LSTM based LM. On the Penn Treebank corpus, our model has shown improvement of the perplexity with not only vanilla LSTM but customized LSTM models.
Abstract The aim of the article is to discuss the legal language transformations from a diachronic perspective taking into account the following factors: (i) spatial and temporal, (ii) linguistic norm changes, (iii) political, (iv) social (customs), and (v) globalization as well as (vi) EU-induced. Spatial and temporal factors include legal relations influenced by climate and the cycles of nature. Linguistic factors include spelling reforms and grammatical changes each language undergoes, for example, as a result of usage. As far as the law is concerned, normative changes can be observed when laws are amended. Other factors such as customs, usage, etc. cannot be neglected when discussing the language of the law. Analogously political correctness and usage can be observed in gender sensitive language and the introduction of such terms as chairperson instead of chairman. Social factors should not be overlooked. As a result of social changes, numerous terms have been introduced to legal lexicons in many countries starting with same-sex unions or same-sex-marriages. The so-called political correctness enforces some language changes and leads to the introduction of new terms and at the same time the abandonment of others. Consequently, some terms cease to be used and consequently become archaic. The aim of the article is to focus on diachronic changes in legal languages and present the communication problems resulting from them from intra- and inter-lingual perspectives.
We describe the first automatic approach for merging coreference annotations obtained from multiple annotators into a single gold standard. This merging is subject to certain linguistic hard constraints and optimisation criteria that prefer solutions with minimal divergence from annotators. The representation involves an equivalence relation over a large number of elements. We use Answer Set Programming to describe two representations of the problem and four objective functions suitable for different data-sets. We provide two structurally different real-world benchmark data-sets based on the METU-Sabanci Turkish Treebank and we report our experiences in using the Gringo, Clasp and Wasp tools for computing optimal adjudication results on these data-sets.
Recently, deep learning methods have achieved good results in dependency parsing for many natural languages. In this paper, we investigate the use of bidirectional long short-term memory network models for both transition-based and graph-based dependency parsing for the Vietnamese language. We also reportour contribution in building a Vietnamese dependency treebank whose tagset conforms to the Universal Dependency schema. Various experiments demonstrate the efficiency of this method, which achieves the best parsing accuracy in comparison to other existing approaches on the same corpus, with unlabeled attachment score of 84.45% or labeled attachment scoreof 78.56%.
Despite increased use of behavioral analogues to identify casual mechanisms of self-injurious behavior (e.g., suicide attempts; non-suicidal self-injury), little is known about the impact on participants. The current study examined the impact of a specific behavior analogue, Self-Aggressive Paradigm (SAP), on participant affect. Community participants (n = 507) reported several affective ratings before and after completing SAP task procedures. Following the SAP, participants reported reductions in nervousness and fear and increases in calmness and anger (d =.21). Participants with a current anxiety disorder reported greater increases in happiness; those with a suicide attempt history reported greater increases in sadness. Findings demonstrate the SAP has no adverse mood effects, supporting its use in experimental research.
Abstract This study uses treebanking to investigate how spoken language infiltrated legal Latin in early medieval Italy. The documents used are always formulaic, but they also always contain a ‘free’ part where the case in question is described in free prose. This paper uses this difference to measure how ten linguistic features, representative of the evolution that took place between Classical and Late Latin, are distributed between the formulaic and free parts. Some variants are attested equally often in both parts of the documents, while perceptually or conceptually salient variants appear to be preserved in their conservative form mainly in the formulaic parts.
Unsupervised learning of syntactic structure is typically performed using generative models with discrete latent variables and multinomial parameters. In most cases, these models have not leveraged continuous word representations. In this work, we propose a novel generative model that jointly learns discrete syntactic structure and continuous word representations in an unsupervised fashion by cascading an invertible neural network with a structured generative prior. We show that the invertibility condition allows for efficient exact inference and marginal likelihood computation in our model so long as the prior is well-behaved. In experiments we instantiate our approach with both Markov and tree-structured priors, evaluating on two tasks: part-of-speech (POS) induction, and unsupervised dependency parsing without gold POS annotation. On the Penn Treebank, our Markov-structured model surpasses state-of-the-art results on POS induction. Similarly, we find that our tree-structured model achieves state-of-the-art performance on unsupervised dependency parsing for the difficult training condition where neither gold POS annotation nor punctuation-based constraints are available.
Conventional lexical-clustering algorithms treat text fragments as a mixed collection of words, with a semantic similarity between them calculated based on the term of how many the particular word occurs within the compared fragments. Whereas this technique is appropriate for clustering large-sized textual collections, it operates poorly when clustering small-sized texts such as sentences. This is due to compared sentences that may be linguistically similar despite having no words in common. This chapter presents a new version of the original k-means method for sentence-level text clustering that is relay on the idea of use of the related synonyms in order to construct the rich semantic vectors. These vectors represent a sentence using linguistic information resulting from a lexical database founded to determine the actual sense to a word, based on the context in which it occurs. Therefore, while traditional k-means method application is relay on calculating the distance between patterns, the new proposed version operates by calculating the semantic similarity between sentences. This allows it to capture a higher degree of semantic or linguistic information existing within the clustered sentences. Experimental results illustrate that the proposed version of clustering algorithm performs favorably against other well-known clustering algorithms on several standard datasets.
Language models, being at the heart of many NLP problems, are always of great interest to researchers. Neural language models come with the advantage of distributed representations and long range contexts. With its particular dynamics that allow the cycling of information within the network, `Recurrent neural network' (RNN) becomes an ideal paradigm for neural language modeling. Long Short-Term Memory (LSTM) architecture solves the inadequacies of the standard RNN in modeling long-range contexts. In spite of a plethora of RNN variants, possibility to add multiple memory cells in LSTM nodes was seldom explored. Here we propose a multi-cell node architecture for LSTMs and study its applicability for neural language modeling. The proposed multi-cell LSTM language models outperform the state-of-the-art results on well-known Penn Treebank (PTB) setup.
In natural language processing, the syntactic analysis process (such as constituency parsing) is required to understand word context in the sentence. We propose a modification on using binarization technique alternative and feature multiplication factors for shift-reduce constituency parser using beam search approach and structured learning algorithm. Our modification in binarization technique is inspired from assorted tagging schemes in NER, while the feature multiplication factors is used to scale up our scoring system for beam search algorithm. For evaluation, we mainly used the new INACL Treebank (consisting 11,356 and 4,457 instances for training and test set), resulted 50.3% in f <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> -score. Our parser also compared with previous work by using the same training and test set for IDN-Treebank, resulted 74.0% in f <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> -score.
Embedding and projection matrices are commonly used in neural language models (NLM) as well as in other sequence processing networks that operate on large vocabularies. We examine such matrices in fine-tuned language models and observe that a NLM learns word vectors whose norms are related to the word frequencies. We show that by initializing the weight norms with scaled log word counts, together with other techniques, lower perplexities can be obtained in early epochs of training. We also introduce a weight norm regularization loss term, whose hyperparameters are tuned via a grid search. With this method, we are able to significantly improve perplexities on two word-level language modeling tasks (without dynamic evaluation): from 54.44 to 53.16 on Penn Treebank (PTB) and from 61.45 to 60.13 on WikiText-2 (WT2).
DGT is an intense user of specialised IT applications. Computer assisted translation tools and linguistic databases are an integral part of DGT's differentiated resource mix (in-house translators, free-lance translators, tools). They contribute to the efficient delivery of high quality services to DGT's customers. More specifically the IT systems developed, used and maintained by DGT serve the following main purposes:
An outstanding question in empirical aesthetics concerns whether negative emotions (e.g., fear, disgust) can improve aesthetic judgments of liking. Although negative emotions are sometimes linked with enjoyment in music or visual design/art, emotion priming studies have shown conflicting results, reporting both more negative and more positive assessments. These divergences may be driven by key differences in priming procedures. Specifically, past studies' use of either emotional faces or emotional scenes as primes as well as differing negative emotion content (fear, disgust) may involve differing processes leading to opposing effects, particularly in aesthetic judgments. To differentiate among these, we presented emotion primes (20 ms) consisting of either emotional faces or scenes, further subdivided in disgusting, fearful, neutral, or positive emotional content and tested how liking, valence, and arousal ratings of abstract patterns were affected. Additionally, facial electromyography (fEMG) over M. frontalis (indicator of fear), M. levator labii (disgust), and M. zygomaticus (positive) muscles was recorded, to see whether primes would elicit prime-emotion congruent changes. However, fEMG activations indicated no prime congruent changes. Critically, primes influenced ratings in an emotion congruent manner in both faces and emotional scenes. Stimuli were rated as more liked and positively valenced after positive primes and less liked/more negatively valenced after fear or disgust primes. The similarity of priming effects in both prime types in absence of congruent fEMG changes may suggest that priming exerts its influence via a cognitive rather than a more immediate emotional route. Overall-at least in emotional priming-negative emotions seem to be incompatible with higher liking. (PsycINFO Database Record (c) 2019 APA, all rights reserved).
OBJECTIVE: Previous research has consistently shown that the ability to recognize emotions from facial expressions is impaired in Huntington's disease (HD). The aim of this study was to examine whether people with the gene expansion for HD visually scan the most emotionally informative features of human faces less than unaffected individuals, and whether altered visual scanning predicts emotion recognition in HD beyond general disease-related decline. METHOD: We recorded eye movements of 25 participants either in the late premanifest or early stage of HD and 25 age-matched healthy control participants during a face-viewing task. The task involved the viewing of pictures depicting human faces with angry, disgusted, fearful, happy, and neutral expressions, and evaluating each face on a valence rating scale. For data analysis, we defined 2 regions of interest (ROIs) on each picture, including an eye-ROI and a nose/mouth-ROI. Emotion recognition abilities were measured using an established emotion-recognition task and general disease-related decline was measured using the UHDRS motor score. RESULTS: Compared to the control participants, the HD participants spent less time looking at the ROIs relative to the total time spent looking at the pictures (partial η2 = 0.10), and made fewer fixations on the ROIs (partial η2 = 0.16). Furthermore, visual scanning of the eye-ROI, but not the nose/mouth-ROI, predicted emotion recognition performance in the HD group, over and beyond general disease-related decline. CONCLUSION: The emotion recognition deficit in HD may partly be explained by general disease-related decline in cognition and motor functioning and partly by a social-emotional deficit, which is reflected in reduced eye-viewing. (PsycINFO Database Record
Using the accurate relative pronoun (RP) in a formal writing task presents challenges for writers since they seem to be influenced by forms used in the popular oral variety of French which are far from the linguistic norm (Blanche-Benveniste, 2010). Studies describing the teaching of the relative clause (RC) in the secondary classroom have highlighted the problems encountered by students not only with handling this grammatical object, but also with using their grammatical knowledge in revising their text (Dolz & Schneuwly, 2009). However, to our knowledge, no study has yet been conducted to conceive and test an intervention for teaching RCs in French L1 classes. Based on theoretical and empirical work converging toward the fostering of sustained verbal interactions throughout grammatical and revision instruction, a series of lessons was implemented with 52 grade nine students enrolled in a French course (Montreal, Canada). Pretest and posttest texts were analysed in terms of RC frequency, usage and accuracy. While no difference was found in the general frequency of RCs, results show a significant increase in the use of complex RPs. Students, especially the weaker ones, also make significantly fewer mistakes overall on RPs and also on complex RPs. These results could indicate that certain structures associated with complexity and formal register are used more frequently and more accurately during written production after our intervention. Our results contribute to the ongoing discussion on the complementarity between direct grammar instruction and writing and revision instruction and their positive impact on students' syntactic constructions in texts.
En este artículo presentamos el primer detector de la Unidad Central (CU) de resúmenes científicos en castellano basado en técnicas de aprendizaje automático. Para ello, nos hemos basado en la anotación del Spanish RST Treebank anotado bajo la Teoría de la Estructura Retórica o Rhetorical Structure Theory (RST). El método empleado para detectar la unidad central es el modelo de bolsa de palabras utilizando clasificadores como Naive Bayes y SVM. Finalmente, evaluamos el rendimiento de los clasificadores y hemos creado el detector de CUs usando el mejor clasificador.
This paper addresses and targets morpheme segmentation of Kannada words using supervised classification. We have used manually annotated Kannada treebank corpus, which is recently developed by us. Kannada bears resemblance to other Dravidian languages in morphological structure. It is an agglutinative language, hence its words have complex morphological form with each word comprising of a root and an optional set of suffixes. These suffixes carry additional meaning, apart from the root word in a context. This paper discusses the extraction of morphemes of a word by using Support Vector Machines for Classification. Additional features representing the properties of the Kannada words were extracted and the different letters were classified into labels that result in the morphological segmentation of the word. Various methods for evaluation were considered and an accuracy of 85.97% was achieved.
An increasing amount of research tackles the challenge of text generation from abstract ontological or semantic structures, which are in their very nature potentially large connected graphs. These graphs must be "packaged" into sentence-wise subgraphs. We interpret the problem of sentence packaging as a community detection problem with post optimization. Experiments on the texts of the Verb-Net/FrameNet structure annotated-Penn Treebank, which have been converted into graphs by a coreference merge using Stanford CoreNLP, show a high F 1 -score of 0.738.
Abstract The P3a observed after novel events is an event-related potential comprising an early fronto-central phase and a late fronto-parietal phase. It has classically been considered to reflect the attention processing of distracting stimuli. However, novel sounds can lead to behavioral facilitation as much as behavioral distraction. This illustrates the duality of the orienting response which includes both an attentional and an arousal component. Using a paradigm with visual or auditory targets to detect and irrelevant unexpected distracting sounds to ignore, we showed that the facilitation effect by distracting sounds is independent of the target modality and endures more than 1500 ms. These results confirm that the behavioral facilitation observed after distracting sounds is related to an increase in unspecific phasic arousal on top of the attentional capture. Moreover, the amplitude of the early phase of the P3a to distracting sounds positively correlated with subjective arousal ratings, contrary to other event-related potentials. We propose that the fronto-central early phase of the P3a would index the arousing properties of distracting sounds and would be linked to the arousal component of the orienting response. Finally, we discuss the relevance of the P3a as a marker of distraction.
Dependency grammar induction is the task of learning dependency syntax without annotated training data. Traditional graph-based models with global inference achieve state-of-the-art results on this task but they require $O(n^3)$ 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 graph-based models.
We introduce a class of convolutional neural networks (CNNs) that utilize recurrent neural networks (RNNs) as convolution filters. A convolution filter is typically implemented as a linear affine transformation followed by a non-linear function, which fails to account for language compositionality. As a result, it limits the use of high-order filters that are often warranted for natural language processing tasks. In this work, we model convolution filters with RNNs that naturally capture compositionality and long-term dependencies in language. We show that simple CNN architectures equipped with recurrent neural filters (RNFs) achieve results that are on par with the best published ones on the Stanford Sentiment Treebank and two answer sentence selection datasets.
This paper proposes the model for searching similar collocations in English texts in order to determine semantically connected text fragments for social network data streams analysis. The logical-linguistic model uses semantic and grammatical features of words to obtain a sequence of semantically related to each other text fragments from different actors of a social network. In order to implement the model, we leverage Universal Dependencies parser and Natural Language Toolkit with the lexical database WordNet. Based on the Blog Authorship Corpus, the experiment achieves over 0.92 precision.
Web queries with question intent manifest a complex syntactic structure and the processing of this structure is important for their interpretation. However, their algorithms rely on resources typically not available outside of big web corporates. We propose a new BiLSTM query parser that: (1) Explicitly accounts for the unique grammar of web queries; and (2) Utilizes named entity (NE) information from a BiLSTM NE tagger, that can be jointly trained with the parser. In order to train our model we annotate the query treebank of When trained on 2500 annotated queries our parser achieves UAS of 83.5% and segmentation F1score of 84.5, substantially outperforming existing state-of-the-art parsers. 1
Sentiment analysis involves classifying text into positive, negative and neutral classes according to the emotions expressed in the text. Extensive study has been carried out in performing sentiment analysis using the traditional ‘bag of words’ approach which involves feature selection, where the input is given to classifiers such as Naive Bayes and SVMs. A relatively new approach to sentiment analysis involves using a deep learning model. In this approach, a recently discovered technique called word embedding is used, following which the input is fed into a deep neural network architecture. As sentiment analysis using deep learning is a relatively unexplored domain, we plan to perform in-depth analysis into this field and implement a state of the art model which will achieve optimal accuracy. The proposed methodology will use a hybrid architecture, which consists of CNNs (Convolutional Neural Networks) and RNNs (Recurrent Neural Networks), to implement the deep learning model on the SAR14 and Stanford Sentiment Treebank data sets.
This paper describes the system of our team Phoenix for participating CoNLL 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies. Given the annotated gold standard data in CoNLL-U format, we train the tokenizer, tagger and parser separately for each treebank based on an open source pipeline tool UDPipe. Our system reads the plain texts for input, performs the preprocessing steps (tokenization, lemmas, morphology) and finally outputs the syntactic dependencies. For the low-resource languages with no training data, we use cross-lingual techniques to build models with some close languages instead. In the official evaluation, our system achieves the macro-averaged scores of 65.61%, 52.26%, 55.71% for LAS, MLAS and BLEX respectively.
This chapter reports on the comprehensive reform of an early childhood teacher education program aiming to substantively address the improvement of educational circumstances of culturally and linguistically diverse children. This project engages sociocultural theory, research, and praxis and expands a funds of knowledge perspective. This reform calls for new contexts for action, new forms of family and community engagement, and new circulation of literacies across households, neighborhoods, and U.S. schools. Overarching design goals for the program are focused on challenging the following: (a) narrow conceptions of language literacy and stories; (b) bounded or contained views of learning contexts; (c) hegemonic cultural and linguistic norms, matters regarding U.S. language and educational policies and power; (d) issues of voice and silencing; (e) quantitative and static views of “resources”; and (f) the limited attention to the agency, identities, and strategic actions of diverse students and their families as they traverse home, community, and institutional contexts.
In this paper, we focus on parsing rare and non-trivial constructions, in particular ellipsis. We report on several experiments in enrichment of training data for this specific construction, evaluated on five languages: Czech, English, Finnish, Russian and Slovak. These data enrichment methods draw upon self-training and tri-training, combined with a stratified sampling method mimicking the structural complexity of the original treebank. In addition, using these same methods, we also demonstrate small improvements over the CoNLL-17 parsing shared task winning system for four of the five languages, not only restricted to the elliptical constructions.
With the recent growth in attention to transgender people’s experiences, language has become a focused site of both anxiety and contestation. This chapter focuses on three challenges for trans-inclusive language: how to select gendered labels and pronouns, how to make language more gender-neutral when gender isn’t relevant, and how to talk about gender more precisely when it is relevant, such as in discussions of identity, social inequality, physiology, or socialization. In each case, concrete strategies are presented that reflect the way transgender people themselves reshape language to work in more inclusive and affirming ways. These include more direct negotiation regarding appropriate identifying terms, the selection of gender-neutral or gender-inclusive forms, and avoiding the conflation of different aspects of gender. Although transphobia will not be eliminated through linguistic reform alone, changes to linguistic norms that delegitimize trans identities and normalize cis (non-trans) identities are a critical part of trans liberation.
Despite all the impressive advances of recurrent neural networks, sequential data is still in need of better modelling. Truncated backpropagation through time (TBPTT), the learning algorithm most widely used in practice, suffers from the truncation bias, which drastically limits its ability to learn long-term dependencies.The Real-Time Recurrent Learning algorithm (RTRL) addresses this issue, but its high computational requirements make it infeasible in practice. The Unbiased Online Recurrent Optimization algorithm (UORO) approximates RTRL with a smaller runtime and memory cost, but with the disadvantage of obtaining noisy gradients that also limit its practical applicability. In this paper we propose the Kronecker Factored RTRL (KF-RTRL) algorithm that uses a Kronecker product decomposition to approximate the gradients for a large class of RNNs. We show that KF-RTRL is an unbiased and memory efficient online learning algorithm. Our theoretical analysis shows that, under reasonable assumptions, the noise introduced by our algorithm is not only stable over time but also asymptotically much smaller than the one of the UORO algorithm. We also confirm these theoretical results experimentally. Further, we show empirically that the KF-RTRL algorithm captures long-term dependencies and almost matches the performance of TBPTT on real world tasks by training Recurrent Highway Networks on a synthetic string memorization task and on the Penn TreeBank task, respectively. These results indicate that RTRL based approaches might be a promising future alternative to TBPTT.
We propose a novel way to create categorized discourse lexicons for multiple languages. We combine information from the Penn Discourse Treebank with statistical machine translation techniques on the Europarl corpus. Using gender profiling as an application, we evaluate our approach by comparing it with an approach using features from a knowledge-based lexicon and with an Rhetorical structure theory (RST) discourse parser. Our experiments are performed on corpora for three languages (English, Dutch, and German) in two genres (news and blogs). We include a feature analysis in which we look for (in)consistencies of discourse features related to male and female authors between the different experimental settings.
Answering questions formulated in natural language is a long standing quest in Artificial Intelligence. However, even formulating the problem in precise terms has proven to be too challenging, which lead many researchers to focus on Multiple-Choice Question Answering problems. One particularly interesting type of the latter problem is solving standardized tests such as university entrance exams. The Exame Nacional do Ensino Médio (ENEM) is a High School level exam widely used by Brazilian universities as entrance exam, and the world's second biggest university entrance examination in number of registered candidates. In this work we tackle the problem of answering purely textual multiple-choice questions from the ENEM. We build on a previous solution that formulated the problem as a text information retrieval problem. In particular, we investigate how to enhance these methods by text augmentation using Word Embedding and WordNet, a structured lexical database where words are connected according to some relations like synonymy and hypernymy. We also investigate how to boost performance by building ensembles of weakly correlated solvers. Our approaches obtain accuracies ranging from 26% to 29.3%, outperforming the previous approach.
BACKGROUND: Neuroplastic underpinnings of meditation-induced changes in affective processing are largely unclear. METHODS: We included healthy older participants in an active-controlled experiment. They were involved a meditation training or a control relaxation training of eight weeks. Associations between behavioral and neural morphometric changes induced by the training were examined. RESULTS: The meditation group demonstrated a change in valence perception indexed by more neutral valence ratings of positive and negative affective images. These behavioral changes were associated with synchronous structural enlargements in a prefrontal network involving the ventromedial prefrontal cortex and the inferior frontal sulcus. In addition, these neuroplastic effects were modulated by the enlargement in the inferior frontal junction. In contrast, these prefrontal enlargements were absent in the active control group, which completed a relaxation training. Supported by a path analysis, we propose a model that describes how meditation may induce a series of prefrontal neuroplastic changes related to valence perception. These brain areas showing meditation-induced structural enlargements are reduced in older people with affective dysregulations. CONCLUSION: We demonstrated that a prefrontal network was enlarged after eight weeks of meditation training. Our findings yield translational insights in the endeavor to promote healthy aging by means of meditation.
The paper presents an extension of the Italian Universal Dependencies Treebank with an “enhanced” representation level (e-IUDT), aimed at simplifying the information extraction process. The modules developed to semi-automatically build e-IUDT were delexicalized to perform cross-language enhancements: preliminary experiments in this direction led to promising results.
This paper chooses two news-genre dependency treebanks, one in Chinese and one in English and examines the synergetics or the interrelations among dependency tree widths, heights and sentence lengths in the framework of dependency grammar. When sentences grow longer, the dependency trees grow both taller and higher. The growths of heights and widths are competing with each other, resulting in a minimization of dependency distance. The product of the widest layer and its width corresponds to the sentence length. These correlations are integrated into a tentative synergetic syntactic model based on the framework of dependency grammar.
The quantity of music content is rapidly increasing and automated affective tagging of music video clips can enable the development of intelligent retrieval, music recommendation, automatic playlist generators, and music browsing interfaces tuned to the users' current desires, preferences, or affective states. To achieve this goal, the field of affective computing has emerged, in particular the development of so-called affective brain-computer interfaces, which measure the user's affective state directly from measured brain waves using non-invasive tools, such as electroencephalography (EEG). Typically, conventional features extracted from the EEG signal have been used, such as frequency subband powers and/or inter-hemispheric power asymmetry indices. More recently, the coupling between EEG and peripheral physiological signals, such as the galvanic skin response (GSR), have also been proposed. Here, we show the importance of EEG amplitude modulations and propose several new features that measure the amplitude-amplitude cross-frequency coupling per EEG electrode, as well as linear and non-linear connections between multiple electrode pairs. When tested on a publicly available dataset of music video clips tagged with subjective affective ratings, support vector classifiers trained on the proposed features were shown to outperform those trained on conventional benchmark EEG features by as much as 6, 20, 8, and 7% for arousal, valence, dominance and liking, respectively. Moreover, fusion of the proposed features with EEG-GSR coupling features showed to be particularly useful for arousal (feature-level fusion) and liking (decision-level fusion) prediction. Together, these findings show the importance of the proposed features to characterize human affective states during music clip watching.
In Vietnamese dependency parsing, several methods have been proposed. Dependency parser which uses deep neural network model has been reported that achieved state-of-the-art results. In this paper, we proposed a new method which applies LSTM easy-first dependency parsing with pre-trained word embeddings and character-level word embeddings. Our method achieves an accuracy of 80.91% of unlabeled attachment score and 72.98% of labeled attachment score on the Vietnamese Dependency Treebank (VnDT).
Sarcasm is a sophisticated form of sentiment expression where speaker express their opinions opposite of what they mean. Sarcasm detection and Emotion detection from social net-working sites has been a great field of study. With the growth of e-services such as e-commerce, e-tourism and e-business, the companies are very keen on exploiting emotion and sarcasm analysis for their marketing strategies in order to evaluate the public attitudes towards their brand. Thus efficient emotion and sarcasm modeling system can be a good solution to the above problem. This work aims at developing a system that groups posts based on emotions, sentiment and find sarcastic posts, if present. The proposed system is to develop a prototype that help to come to an inference about the emotions of the posts namely anger, surprise, happy, fear, sorrow, trust, anticipation and disgust with three sentic levels in each. This helps in better understanding of the posts when compared to the approaches which senses the polarity of the posts and gives just their sentiments i.e., positive, negative or neutral. The posts handling these emotions might be sarcastic too. The Sentiment & emotion identification module identifies the sentiment or emotion of the post by evaluating score of each word in the comment which is used by different sarcasm detection methods to detect sarcasm. The emotion identification module uses the lexical databases WordNet, SentiWordNet to find the right sentiment scores for the words with respect to each emotion. It also uses Sarcasm detection algorithms like Emoticon sarcasm detection, Hybrid sarcasm detection, Hashtag Processing, Interjection Word Start (IWT).