Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
16504 papers
We tackle implicit discourse relation classification, a task of automatically determining semantic relationships between arguments. The attention-worthy words in arguments are crucial clues for classifying the discourse relations. Attention mechanisms have been proven effective in highlighting the attention-worthy words during encoding. However, our survey shows that some inessential words are unintentionally misjudged as the attention-worthy words and, therefore, assigned heavier attention weights than should be. We propose a penalty-based loss re-estimation method to regulate the attention learning process, integrating penalty coefficients into the computation of loss by means of overstability of attention weight distributions. We conduct experiments on the Penn Discourse TreeBank (PDTB) corpus. The test results show that our loss re-estimation method leads to substantial improvements for a variety of attention mechanisms, and it obtains highly competitive performance compared to the state-of-the-art methods.
Gender can be considered an embodied social concept encompassing biological as well as cultural components. In this paper, we explored whether the concept of gender varies as a function of different cultural and linguistic norms by comparing communities that vary in their social treatment of gender-related issues and linguistic encoding of gender. In Study 1, Italian, Dutch, and English speaking participants completed a free-listing task which showed Italians and Dutch were the most distinct in their conceptualization of gender: Italian participants focused more on sociocultural features (e.g., discrimination, politics, power), whereas Dutch participants focused more on the corporeal sphere (e.g., hormones, breasts, genitals). Study 2 replicated this finding focusing on Italian and Dutch and using a typicality rating task: sociocultural and abstract features were considered as more typical of “gender” by Italian than Dutch participants. Study 3 addressed Italian and Dutch participants’ explicit beliefs about gender with a questionnaire measuring essentialism and constructivism, and consolidated results from Study 1 and 2 showing that Dutch participants endorsed more essentialist beliefs about gender compared to Italian participants. Our results provide evidence that gender is conceptualized differently by diverse groups more in line with sociocultural constructivist accounts and is adapted to specific cultural and linguistic environments.
The purpose of this study was to experimentally investigate the relationship between positive affect elicitation (using a short video clip) prior to exercise and affect during acute aerobic exercise. A counterbalanced, within-subject experimental design was used. We conducted three related experiments. In Experiment 1, 30 adults aged 18–40 years participated in a positive affect-elicitation condition (“affective priming”) and a control condition. Participation involved watching a five-minute video clip, as well as walking on a treadmill at a (self-selected) brisk pace for ten minutes. We compared affective ratings at baseline and intra-exercise for both conditions using a 2 (condition; priming versus no priming) × 2 (time; pre- versus mid-exercise) repeated measures ANOVA. In the follow-up experiments, we re-examined the relationship between affective priming and intra-exercise affect, addressing some limitations noted with Experiment 1. In Experiment 2, we compared the affect-elicitation properties of self-selected and imposed video clips. In Experiment 3, we re-investigated the potential affective benefits of priming, while including a neutral (neither positive nor negative) video during the control condition to attenuate potential demand characteristics, and a positive video-only condition to investigate possible carryover effects. Self-selected and imposed film clips showed similar affect-elicitation properties. Comparing the priming and control conditions, there were notable differences in the mean intra-exercise affective valence ratings (p = 0.07 Experiment 1, p = 0.01 Experiment 3). The mean affective activation ratings were not significantly different (p = 0.07 Experiment 1, p = 0.86 Experiment 3). Priming the affective state prior to exercise may be beneficial for enhancing intra-exercise affect.
Pain is evolutionarily hardwired to signal potential danger and threat. It has been proposed that altered pain-related associative learning processes, i.e., emotional or fear conditioning, might contribute to the development and maintenance of chronic pain. Pain in or near the face plays a special role in pain perception and processing, especially with regard to increased pain-related fear and unpleasantness. However, differences in pain-related learning mechanisms between the face and other body parts have not yet been investigated. Here, we examined body-site specific differences in associative emotional conditioning using electrical stimuli applied to the face and the hand. Acquisition, extinction, and reinstatement of cue-pain associations were assessed in a 2-day emotional conditioning paradigm using a within-subject design. Data of 34 healthy subjects revealed higher fear of face pain as compared to hand pain. During acquisition, face pain (as compared to hand pain) led to a steeper increase in pain-related negative emotions in response to conditioned stimuli (CS) as assessed using valence ratings. While no significant differences between both conditions were observed during the extinction phase, a reinstatement effect for face but not for hand pain was revealed on the descriptive level and contingency awareness was higher for face pain compared to hand pain. Our results indicate a stronger propensity to acquire cue-pain-associations for face compared to hand pain, which might also be reinstated more easily. These differences in learning and resultant pain-related emotions might play an important role in the chronification and high prevalence of chronic facial pain and stress the evolutionary significance of pain in the head and face.
Many educational institutions in North America have declared a commitment to enhancing the diversity of their students, and to providing a learning environment free of discrimination. This diversity unequivocally includes sexual orientation as well as gender identity and expression, and we teachers are expected to play a role in fulfilling this commitment. Nevertheless, there are few opportunities for us to learn about the diversity of gender and sexuality in pre-service training or in-service professional development for Japanese-language education. This paper addresses issues that may create challenges for LGBTQ learners of Japanese, paying special attention to heteronormativity in Japanese language teaching materials and linguistic norms and ideology regarding gendered expression in Japanese, and suggests ways teachers might deal with these issues in order to create an inclusive learning environment for all students regardless of their gender and sexuality.
(NAPS; Marchewka et al., 2014). We collected ratings of perceived valence and arousal for both material groups and recorded eye movements of 42 participants during reading and picture viewing. Linear mixed-effects models were performed to analyze effects of valence (i.e., valence category, valence rating) and stimulus domain (i.e., textual, pictorial) on ratings of perceived valence and arousal, eye movements in reading, and eye movements in picture viewing. Results supported the success of our experimental manipulation: emotionally positive stimuli (i.e., vignettes, pictures) were perceived more positively and less arousing than emotionally negative ones. The cross-domain comparison indicated that vignettes are able to induce stronger valence effects than their pictorial counterparts, no differences between vignettes and pictures regarding effects on perceived arousal were found. Analyses of eye movements in reading replicated results from experiments using isolated words and sentences: perceived positive text valence attracted shorter reading times than perceived negative valence at both the supralexical and lexical level. In line with previous findings, no emotion effects on eye movements in picture viewing were found. This is the first eye tracking study reporting superior valence effects for vignettes compared to pictures and valence-specific effects on eye movements in reading at the supralexical level.
Emotion research typically searches for consistency and specificity in physiological activity across instances of an emotion category, such as anger or fear, yet studies to date have observed more variation than expected. In the present study, we adopt an alternative approach, searching inductively for structure within variation, both within and across participants. Following a novel, physiologically-triggered experience sampling procedure, participants’ self-reports and peripheral physiological activity were recorded when substantial changes in cardiac activity occurred in the absence of movement. Unsupervised clustering analyses revealed variability in the number and nature of patterns of physiological activity that recurred within individuals, as well as in the affect ratings and emotion labels associated with each pattern. There were also broad patterns that recurred across individuals. These findings support a constructionist account of emotion which, drawing on Darwin, proposes that emotion categories are populations of variable instances tied to situation-specific needs.
Traditional approaches to set goals in second language (L2) vocabulary acquisition relied either on word lists that were obtained from large L1 corpora or on collective knowledge and experience of L2 experts, teachers, and examiners. Both approaches are known to offer some advantages, but also to have some limitations. In this paper, we try to combine both sources of information, namely the official reference level description for French language and the FLElex lexical database. Our aim is to train a statistical model on the French RLD that would be able to turn the distributional information from FLElex into one of the six levels of the Common European Framework of Reference for languages (CEFR). We show that such approach yields a gain of 29\\% in accuracy compared to the method currently used in the CEFRLex project. Besides, our experiments also offer deeper insights into the advantages and shortcomings of the two traditional sources of information (frequency vs. expert knowledge).
We propose a novel constituency parsing model that casts the parsing problem into a series of pointing tasks. Specifically, our model estimates the likelihood of a span being a legitimate tree constituent via the pointing score corresponding to the boundary words of the span. Our parsing model supports efficient top-down decoding and our learning objective is able to enforce structural consistency without resorting to the expensive CKY inference. The experiments on the standard English Penn Treebank parsing task show that our method achieves 92.78 F1 without using pre-trained models, which is higher than all the existing methods with similar time complexity. Using pre-trained BERT, our model achieves 95.48 F1, which is competitive with the state-of-theart while being faster. Our approach also establishes new state-of-the-art in Basque and Swedish in the SPMRL shared tasks on multilingual constituency parsing.
The importance of affect processing to human behavior has long driven researchers to pursue its measurement. In this study, we compared the relative fidelity of measurements of neural activation and physiology (i.e., heart rate change) in detecting affective valence induction across a broad continuum of conveyed affective valence. We combined intra-subject neural activation based multivariate predictions of affective valence with measures of heart rate (HR) deceleration to predict predefined normative affect rating scores for stimuli drawn from the International Affective Picture System (IAPS) in a population (n = 50) of healthy adults. In sum, we found that patterns of neural activation and HR deceleration significantly, and uniquely, explain the variance in normative valent scores associated with IAPS stimuli; however, we also found that patterns of neural activation explain a significantly greater proportion of that variance. These traits persisted across a range of stimulus sets, differing by the polar-extremity of their positively and negatively valent subsets, which represent the positively and negatively valent polar-extremity of stimulus sets reported in the literature. Overall, these findings support the acquisition of heart rate deceleration concurrently with fMRI to provide convergent validation of induced affect processing in the dimension of affective valence.
For sequence models with large word-level vocabularies, a majority of network parameters lie in the input and output layers. In this work, we describe a new method, DeFINE, for learning deep word-level representations efficiently. Our architecture uses a hierarchical structure with novel skip-connections which allows for the use of low dimensional input and output layers, reducing total parameters and training time while delivering similar or better performance versus existing methods. DeFINE can be incorporated easily in new or existing sequence models. Compared to state-of-the-art methods including adaptive input representations, this technique results in a 6% to 20% drop in perplexity. On WikiText-103, DeFINE reduces total parameters of Transformer-XL by half with minimal impact on performance. On the Penn Treebank, DeFINE improves AWD-LSTM by 4 points with a 17% reduction in parameters, achieving comparable performance to state-of-the-art methods with fewer parameters. For machine translation, DeFINE improves a Transformer model by 2% while simultaneously reducing total parameters by 26%
This paper discusses the theoretical bases as well as the pragmatic implementation of the lemmatization of the Late Latin Charter Treebanks (LLCT). LLCT is a set of three dependency treebanks (LLCT1, LLCT2, LLCT3) of Early Medieval Latin documentary texts (charters) written in Italy between AD 714 and 1000 (c. 594,000 tokens). The original model for the lemmatization of LLCT was the Latin Dependency Treebank (LDT), which is mainly Classical standard Latin and based on the entries of Lewis and Short’s Latin Dictionary. Since LLCT reflects later linguistic developments of Latin and contains a plethora of non-standard proper names, particular attention is paid to how non-standard lexemes are lemmatized systematically to make the lemmatization maximally usable. The theoretical underpinnings to manage the lemmatization boil down to two principles: the evolutionary principle and the parsimony principle.
In this paper, we first open on important issues regarding the Penn Korean Universal Treebank (PKT-UD) and address these issues by revising the entire corpus manually with the aim of producing cleaner UD annotations that are more faithful to Korean grammar. For compatibility to the rest of UD corpora, we follow the UDv2 guidelines, and extensively revise the part-of-speech tags and the dependency relations to reflect morphological features and flexible word-order aspects in Korean. The original and the revised versions of PKT-UD are experimented with transformer-based parsing models using biaffine attention. The parsing model trained on the revised corpus shows a significant improvement of 3.0% in labeled attachment score over the model trained on the previous corpus. Our error analysis demonstrates that this revision allows the parsing model to learn relations more robustly, reducing several critical errors that used to be made by the previous model.
Parsing sentences into syntax trees can benefit downstream applications in NLP. Transition-based parsers build trees by executing actions in a state transition system. They are computationally efficient, and can leverage machine learning to predict actions based on partial trees. However, existing transition-based parsers are predominantly based on the shift-reduce transition system, which does not align with how humans are known to parse sentences. Psycholinguistic research suggests that human parsing is strongly incremental: humans grow a single parse tree by adding exactly one token at each step. In this paper, we propose a novel transition system called attach-juxtapose. It is strongly incremental; it represents a partial sentence using a single tree; each action adds exactly one token into the partial tree. Based on our transition system, we develop a strongly incremental parser. At each step, it encodes the partial tree using a graph neural network and predicts an action. We evaluate our parser on Penn Treebank (PTB) and Chinese Treebank (CTB). On PTB, it outperforms existing parsers trained with only constituency trees; and it performs on par with state-of-the-art parsers that use dependency trees as additional training data. On CTB, our parser establishes a new state of the art. Code is available at this https URL.
Abstract Two research traditions explain the way we deal with emotional situations: emotional intelligence (EI) and emotion regulation (ER). EI refers to the individual differences in the knowledge, identification, and regulation of emotions. ER describes processes in which emotions are experienced, expressed, and altered. Our study examined the EI-ER link and their moderating role on affective responses. We used self-report questionnaires and a cognitive reappraisal (CR) task, in which subjective affective responses were registered. We found that higher levels of ER difficulties correlated with lower EI. Gender had an overall impact on affective changes, indicating a more unpleasant and more arousing affective state for women compared with men. Regarding the moderating role of EI and ER difficulties, the ability to utilize emotions (Utilization) decreased the valence into a more unpleasant direction, similar to the effect of the inability to identify and differentiate emotions (Clarity). A weak control over emotions (Impulse), however, increased the valence into a more pleasant direction. The lack of attention to emotional signals (Awareness) marginally decreased the initial intensity (i.e., lower level of arousal). We demonstrated that EI and ER have distinctive routes and a different influence on the affective outcome defined by valence and arousal ratings: (1) EI has an impact through the utilization of emotions mainly on the valence dimension; and (2) individual differences in ER have a moderating effect on both valence and arousal dimensions. This study provided evidence on how individual differences contribute to a successful ER process when using a CR strategy.
Exaggerated reactivity to drug-cues and emotional dysregulations represent key symptoms of early stages of substance use disorders. The diagnostic criteria for (Internet) gaming disorder strongly resemble symptoms for substance-related addictions. However, previous cross-sections studies revealed inconsistent results with respect to neural cue reactivity and emotional dysregulations in these populations. To this end, the present fMRI study applied a combined cross-sectional and longitudinal design in regular online gamers (n = 37) and gaming-naïve controls (n = 67). To separate gaming-associated changes from predisposing factors, gaming-naive subjects were randomly assigned to 6 weeks of daily Internet gaming or a non-gaming condition. At baseline and after the training, subjects underwent an fMRI paradigm presenting gaming-related cues and non-gaming-related emotional stimuli. Cross-sectional comparisons revealed gaming-cue specific enhanced valence attribution and neural reactivity in a parietal network, including the posterior cingulate in regular gamers as compared to gaming naïve-controls. Longitudinal analysis revealed that 6 weeks of gaming elevated valence ratings as well as neural cue-reactivity in a similar parietal network, specifically the posterior cingulate in previously gaming-naïve controls. Together, the longitudinal design did not reveal supporting evidence for altered emotional processing of non-gaming associated stimuli in regular gamers whereas convergent evidence for increased emotional and neural reactivity to gaming-associated stimuli was observed. Findings suggest that exaggerated neural reactivity in posterior parietal regions engaged in default mode and automated information processing already occur during early stages of regular gaming and probably promote continued engagement in gaming behavior.
Social animals show reduced physiological responses to aversive events if a conspecific is physically present. Although humans are innately social, it is unclear whether the mere physical presence of another person is sufficient to reduce human autonomic responses to aversive events. In our study, participants experienced aversive and neutral sounds alone (alone treatment) or with an unknown person that was physically present without providing active support. The present person was a member of the participants' ethnical group (ingroup treatment) or a different ethnical group (outgroup treatment), inspired by studies that have found an impact of similarity on social modulation effects. We measured skin conductance responses (SCRs) and collected subjective similarity and affect ratings. The mere presence of an ingroup or outgroup person significantly reduced SCRs to the aversive sounds compared with the alone condition, in particular in participants with high situational anxiety. Moreover, the effect was stronger if participants perceived the ingroup or outgroup person as dissimilar to themselves. Our results indicate that the mere presence of another person was sufficient to diminish autonomic responses to aversive events in humans, and thus verify the translational validity of basic social modulation effects across different species.
The process of identifying the meaning of a polysemous word correctly from a given context is known as the Word Sense Disambiguation (WSD) in natural language processing (NLP). Adapted Lesk algorithm based system is proposed which makes use of knowledge based approach. This work utilizes WordNet as the knowledge source (lexical database). The proposed system has three units – Input query, Pre-Processing and WSD classifier. Task of input query is to take the inputs sentence (which is an unstructured query) from the user and render it to the pre-processing unit. Pre-processing unit will convert the received unstructured query into a structured query by adding some features such as Part of Speech (POS) tagging, grammatical identification (Subject, Verb, and Object) and this structured query is transferred to the WSD classifier. WSD classifier uniquely identifies the sense of the polysemous word using the context information of the query and the lexical database.
OBJECTIVE: We investigated the relations between psychopathic traits and fear enjoyment. METHOD: In Study 1, 140 undergraduate participants (62 men, 78 women) watched the footage of video game play meant to induce either excitement or fear, rating each on positive/negative adjectives. In Study 2, 150 undergraduate participants (94 women, 56 men) rated valence (positive/negative) of 20 sets of morphed surprise/fear photos. RESULTS: In Study 1, participants with higher levels of psychopathy rated the fear video as less negative and more positive. In Study 2, valence ratings became more negative as fear information increased (fear-laden faces were rated more negatively than surprise-laden faces). As well, there were significant interactions between psychopathy and morph level in predicting valence with psychopathic traits being associated with giving higher positivity ratings to fear-laden faces. CONCLUSIONS: The results of these two studies suggest that people with psychopathic traits have a more positive interpretation of the experience of fear, which could extend to evaluations of others' experiences of fear.
Dual language immersion (DLI) programs have emerged in the U.S. as effective ways to bring together language minority and language majority speakers in school settings with the goal of bilingualism and bi-literacy for all. However, the proliferation of these programs has raised concerns regarding issues of inequity and dissimilar power dynamics in these spaces (Cervantes-Soon, 2014 Cervantes-Soon, C. G. 2014. “A Critical Look at Dual Language Immersion in the New Latin@ Diaspora.” Bilingual Research Journal 37 (1): 64–82.[Taylor & Francis Online], [Google Scholar], “A Critical Look at Dual Language Immersion in the New Latin@ Diaspora.” Bilingual Research Journal 37 (1): 64–82; Flores, 2016, Do Black Lives matter in Bilingual Education [Web log post]. Accessed May 1, 2017. https://educationallinguist.wordpress.com/2016/09/11/do-black-lives-matter-in-bilingual-education/; Valdes, 1997, “Dual language immersion programs: A cautionary note concerning the education of language-minority students.” Harvard Educational Review 67: 391–430, 2018, “Analyzing the curricularization of language in two-way immersion education: Restating two cautionary notes.” Bilingual Research Journal). With this in mind, this study aims to shed light on the intricate social processes at work in DLI contexts. In particular, this paper examines first, how notions of language use, race, and ethnicity are socially constructed and intersect in DLI settings; and second, it explores how these ideas are discerned and re-shaped by young children into their own social and linguistic norms. Employing qualitative research methods, this year-long ethnographic case study uses the intersectional lens of raciolinguistics (Alim, Rickford & Ball, 2016 Alim, H. S., J. R. Rickford, and A. F. Ball, eds. 2016. Raciolinguistics: how Language Shapes our Ideas About Race. New York, NY: Oxford University Press.[Crossref], [Google Scholar], Raciolinguistics: how language shapes our ideas about race. New York, NY: Oxford University Press; Rosa & Flores, 2017, “Unsettling race and language: Toward a raciolinguistic perspective.” Language in Society 46 (5): 621–647), to examine the intricate cross-cutting dynamics at play in bilingual spaces. The exploration of these ideas helps to illuminate the ways in which language practices and interactions are shaped by social constructions from a very early age. Furthermore, it contributes to understandings of social perceptions and relations in multilingual/multicultural/multiethnic contemporary school settings.
This paper describes our system to predict enhanced dependencies for Universal Dependencies (UD) treebanks, which ranked 2 nd in the Shared Task on Enhanced Dependency Parsing with an average ELAS of 82.60%. Our system uses a hybrid two-step approach. First, we use a graph-based parser to extract a basic syntactic dependency tree. Then, we use a set of linguistic rules which generate the enhanced dependencies for the syntactic tree. The application of these rules is optimized using a classifier which predicts their suitability in the given context. A key advantage of this approach is its language independence, as rules rely solely on dependency trees and UPOS tags which are shared across all languages.
Memory-based learning can be characterized as a lazy learning method in machine learning terminology because it delays the processing of input by storing the input until needed. Linguistic structure parsing, which has been in a performance improvement bottleneck since the latest series of works was presented, determines the syntactic or semantic structure of a sentence. In this article, we construct a memory component and use it to augment a linguistic structure parser which allows the parser to directly extract patterns from the known training treebank to form memory. The experimental results show that existing state-of-the-art parsers reach new heights of performance on the main benchmarks for dependency parsing and semantic role labeling with this memory network.
Contribution of emotional valence and arousal to attentional processing over time is not fully understood. We employed a rapid serial visual paradigm (RSVP) in three experiments to investigate the role of valence and arousal. In three experiments, participants had to identify the expression of the two targets (experiment 1 - happy and angry; experiment 2 - angry and surprise; experiment 3 - happy and surprise) presented among neutral upright face distracters. In the first and third experiments, the two targets differed both in valence and arousal ratings. In experiment 2, the surprise and angry expressions differed in terms of valence but were matched for arousal. There was a happy expression advantage (lesser attentional blink) when the first target was anger (experiment 1) or surprise (experiment 3) and a surprise expression advantage when the first target was anger (experiment 2). There was a backward blink with reduced detection of the first target primarily by the relatively more positive valence second target. These results indicate that the benefit for happy and surprise expressions in comparison to angry expression identification is probably due to valence (more positive) and not arousal. Our results demonstrate a novel dynamic interplay of emotional information on temporal attention.
In this paper, we introduce the resources that we developed for Turkish\ndependency parsing, which include a novel manually annotated treebank (BOUN\nTreebank), along with the guidelines we adopted, and a new annotation tool\n(BoAT). The manual annotation process we employed was shaped and implemented by\na team of four linguists and five Natural Language Processing (NLP)\nspecialists. Decisions regarding the annotation of the BOUN Treebank were made\nin line with the Universal Dependencies (UD) framework as well as our recent\nefforts for unifying the Turkish UD treebanks through manual re-annotation. To\nthe best of our knowledge, BOUN Treebank is the largest Turkish treebank. It\ncontains a total of 9,761 sentences from various topics including biographical\ntexts, national newspapers, instructional texts, popular culture articles, and\nessays. In addition, we report the parsing results of a state-of-the-art\ndependency parser obtained over the BOUN Treebank as well as two other\ntreebanks in Turkish. Our results demonstrate that the unification of the\nTurkish annotation scheme and the introduction of a more comprehensive treebank\nlead to improved performance with regard to dependency parsing.\n
Affective responses to music have been shown to be influenced by the psychoacoustic features of the acoustic signal, learned associations between musical features and emotions, and familiarity with a musical system through exposure. The present article reports two experiments investigating whether short-term exposure has an effect on valence and consonance ratings of unfamiliar musical chords from the Bohlen-Pierce system, which are not based on a traditional Western musical scale. In a pre- and post-test design, exposure to positive, negative and neutral chord types was manipulated to test for an effect of exposure on liking. In this paradigm, short-term (“mere”) exposure to unfamiliar chords produced an increase only in valence ratings for negative chords. In neither experiment did it produce an increase in valence or pleasantness ratings for other chord types. Contrast effects for some chord types were found in both experiments, suggesting that a chord’s affect (i.e., affective response to the chord) might be emphasised when the chord is preceded by a stimulus with a contrasting affect. The results confirmed those of a previous study showing that psychoacoustic features play an important role in the perception of music. The findings are discussed in light of their psychological and musical implications.
In this paper, we propose MCNN-ReMGU model based on multi-window convolution and residual-connected minimal gated unit (MGU) network for the natural language word prediction. First, the convolution kernels with different sizes are used to extract the local feature information of different graininess between the word sequences. Then, the extracted features are fed to the residual-connected MGU network. Finally, the prediction results are output by the SoftMax layer. Through the residual-connection processing of MGU network in the model, not only the problems of vanishing gradient and network degradation are effectively solved, but also the long-term dependence between word sequences is effectively extracted to predict the next word accurately. Meanwhile, the introduction of the convolution kernel in a convolutional neural network (CNN) enables the feature information between word sequences to be extracted more fully. The experimental results on the Penn Treebank and WikiText-2 datasets show that the proposed method has certain advantages in the word prediction task.
HDT-UD, the largest German UD treebank by a large margin, as well as the German-LIT treebank, currently do not analyze preposition-determiner contractions such as zum (= zu dem, “to the”) as multi-word tokens, which is inconsistent both with UD guidelines as well as other German UD corpora (GSD and PUD). In this paper, we show that harmonizing corpora with regard to this highly frequent phenomenon using a lookup-table based approach leads to a considerable increase in automatic parsing performance.
Music style is tightly connected with listeners’ emotional processes and neural activities. However, it remains unclear how the brain works when different music styles are processed emotionally. The current study analyzed the neural activation associated with five music styles during emotion-evoking. Twenty non-musicians participated in the functional magnetic resonance imaging (fMRI) scanning and the emotional ratings of pleasure and arousal evoked by pop, rock, jazz, folk, and classical music. Results showed that classical music was associated with the highest pleasure rating and deactivation of the corpus callosum. Rock music was associated with the highest arousal rating and deactivation of the cingulate gyrus. Pop music activated the bilateral supplementary motor areas (SMA) and the superior temporal gyrus (STG) with moderate pleasure and arousal. As the first fMRI experiment investigating the relationship between the music style and emotion, it provides neural correlates of different music styles during emotion-evoking.
To overcome the lack of NLP resources for the low-resource languages, we can utilize tools that are already available for other highresource languages and then modify the output to conform to the target language. In this study, we proposed an approach to convert an Indonesian constituency treebank to a dependency treebank by utilizing an English NLP tool (Stanford CoreNLP) to create the initial dependency treebank. Some annotations in this initial treebank did not conform to Indonesian grammar, especially noun phrases' head-directionality. Noun phrases in English usually have head-final direction, while in Indonesian is the opposite, head-initial. We proposed a variant of tree rotations algorithm named headSwap for dependency trees. We used this algorithm to convert the head-directionality for noun phrases that were initially labeled as a compound. Moreover, we also proposed a set of rules to rename the dependency relation labels to conform to the recent guidelines. To evaluate our proposed method, we created a gold standard of 2,846 tokens that were annotated manually. Experiment results showed that our proposed method improved the Unlabeled Attachment Score (UAS) with a margin of 32.5% from 61.6 to 94.1% and the Labeled Attachment Score (LAS) with a margin of 41% from 44.1 to 85.1%. Finally, we created a new Indonesian dependency treebank that converted automatically using our proposed method that consists of 25,416 tokens. The dependency parser model built using this treebank has UAS of 75.90% and LAS of 70.38%.
The present study investigates the relationship between two features of dependencies, namely, dependency distances and dependency frequencies. The study is based on the analysis of a parallel dependency treebank that includes 10 Indo-European languages. Two corresponding random dependency treebanks are generated as baselines for comparison. After computing the values of dependency distances and their frequencies in these treebanks, for each lan-guage, we fit four functions, namely quadratic, exponent, logarithm, and power-law func-tions, to its original and random datasets. The preliminary result shows that there is a rela-tion between the two dependency features for all 10 Indo-European languages. The relation can be further formalized as a power-law function which can distinguish the observed data from randomly generated datasets.
Background: Alcohol priming can modulate the value of rewards, as observed through the effects of acute alcohol administration on cue reactivity. However, little is known about the psychophysiological mechanisms driving these effects. Here we examine how alcohol-induced changes in bodily states shape the development of implicit attentional biases and explicit cue reactivity. Aims: To characterize the interoceptive correlates of alcohol priming effects on alcohol attentional biases and cue reactivityMethods: In a two-session double blind alcohol administration procedure, participants (n=31) were given a 0.4g/kg dose of alcohol or a placebo drink. Cardiovascular responses were measured before and after alcohol administration to observe the effects of alcohol on viscero-afferent reactivity, as indexed through changes in heart-rate variability (HRV) at or near 0.1Hz (0.1Hz HRV). Next, participants completed a modified Flanker task to examine implicit alcohol attentional biases and provided subjective valence and arousal ratings of alcohol cues to examine explicit cue reactivity. Results: We found that changes in 0.1Hz HRV after alcohol administration positively correlated with attentional biases, and negatively correlated with alcohol valence ratings; breath alcohol content was a null predictor. Conclusion: This is novel evidence that suggests alcohol-induced changes in bodily states may mediate the occurrence of alcohol priming effects, and highlights the potentially generative role of interoceptive mechanisms in alcohol-related behaviors. The differential patterns revealed by implicit biases and explicit response tendencies is considered within the context of the dissociation between wanting and liking.
We propose a method for unsupervised parsing based on the linguistic notion of a constituency test. One type of constituency test involves modifying the sentence via some transformation (e.g. replacing the span with a pronoun) and then judging the result (e.g. checking if it is grammatical). Motivated by this idea, we design an unsupervised parser by specifying a set of transformations and using an unsupervised neural acceptability model to make grammaticality decisions. To produce a tree given a sentence, we score each span by aggregating its constituency test judgments, and we choose the binary tree with the highest total score. While this approach already achieves performance in the range of current methods, we further improve accuracy by fine-tuning the grammaticality model through a refinement procedure, where we alternate between improving the estimated trees and improving the grammaticality model. The refined model achieves 62.8 F1 on the Penn Treebank test set, an absolute improvement of 7.6 points over the previous best published result.
The deep inside-outside recursive autoencoder (DIORA; Drozdov et al. 2019a) is a selfsupervised neural model that learns to induce syntactic tree structures for input sentences without access to labeled training data. In this paper, we discover that while DIORA exhaustively encodes all possible binary trees of a sentence with a soft dynamic program, its vector averaging approach is locally greedy and cannot recover from errors when computing the highest scoring parse tree in bottom-up chart parsing. To fix this issue, we introduce S-DIORA, an improved variant of DIORA that encodes a single tree rather than a softlyweighted mixture of trees by employing a hard argmax operation and a beam at each cell in the chart. Our experiments show that through fine-tuning a pre-trained DIORA with our new algorithm, we improve the state of the art in unsupervised constituency parsing on the English WSJ Penn Treebank by 2.2 6% F1, depending on the data used for fine-tuning.
We report the results of our system on the Metaphor Detection Shared Task at the Second Workshop on Figurative Language Processing 2020. Our model is an ensemble, utilising contextualised and static distributional semantic representations, along with word-type concreteness ratings. Using these features, it predicts word metaphoricity with a deep multilayer perceptron. We are able to best the stateof-the-art from the 2018 Shared Task by an average of 8.0% F 1, and finish fourth in both subtasks in which we participate.
Implicit discourse relation recognition is a challenging task due to the lack of connectives as strong linguistic clues. Previous methods primarily encode two arguments separately or extract the specific interaction patterns for the task, which have not fully exploited the annotated relation signal. Therefore, we propose a novel TransS-driven joint learning architecture to address the issues. Specifically, based on the multi-level encoder, we 1) translate discourse relations in low-dimensional embedding space (called TransS), which could mine the latent geometric structure information of argumentrelation instances; 2) further exploit the semantic features of arguments to assist discourse understanding; 3) jointly learn 1) and 2) to mutually reinforce each other to obtain the better argument representations, so as to improve the performance of the task. Extensive experimental results on the Penn Discourse TreeBank (PDTB) show that our model achieves competitive results against several state-of-the-art systems.
The large communication cost for exchanging gradients between different nodes significantly limits the scalability of distributed training for large-scale learning models. Motivated by this observation, there has been significant recent interest in techniques that reduce the communication cost of distributed Stochastic Gradient Descent (SGD), with gradient sparsification techniques such as top-k and random-k shown to be particularly effective. The same observation has also motivated a separate line of work in distributed statistical estimation theory focusing on the impact of communication constraints on the estimation efficiency of different statistical models. The primary goal of this paper is to connect these two research lines and demonstrate how statistical estimation models and their analysis can lead to new insights in the design of communication-efficient training techniques. We propose a simple statistical estimation model for the stochastic gradients which captures the sparsity and skewness of their distribution. The statistically optimal communication scheme arising from the analysis of this model leads to a new sparsification technique for SGD, which concatenates random-k and top-k, considered separately in the prior literature. We show through extensive experiments on both image and language domains with CIFAR-10, ImageNet, and Penn Treebank datasets that the concatenated application of these two sparsification methods consistently and significantly outperforms either method applied alone.
Text discourse parsing plays an important role in understanding information flow and argumentative structure in natural language. Previous research under the Rhetorical Structure Theory (RST) has mostly focused on inducing and evaluating models from the English treebank. However, the parsing tasks for other languages such as German, Dutch, and Portuguese are still challenging due to the shortage of annotated data. In this work, we investigate two approaches to establish a neural, cross-lingual discourse parser via: (1) utilizing multilingual vector representations; and
Both syntactic and semantic structures are key linguistic contextual clues, in which parsing the latter has been well shown beneficial from parsing the former. However, few works ever made an attempt to let semantic parsing help syntactic parsing. As linguistic representation formalisms, both syntax and semantics may be represented in either span (constituent/phrase) or dependency, on both of which joint learning was also seldom explored. In this paper, we propose a novel joint model of syntactic and semantic parsing on both span and dependency representations, which incorporates syntactic information effectively in the encoder of neural network and benefits from two representation formalisms in a uniform way. The experiments show that semantics and syntax can benefit each other by optimizing joint objectives. Our single model achieves new state-of-the-art or competitive results on both span and dependency semantic parsing on Propbank benchmarks and both dependency and constituent syntactic parsing on Penn Treebank.
This paper presents theoretical and methodological questions related to the creation of a Linguistic Database, made up of samples from the Cazumbá Iracema Extractive Reserve, located in the state of Acre, and discusses the main challenges found and contributions to the teaching and learning process of Portuguese. The methodology for collecting and organizing this database is based on the theoretical assumptions of sociolinguistic patterns, the empirical foundations of the Theory of Linguistic Variation and Change, and the methodology for collecting and manipulating data in sociolinguistics. The implementation of the proposal involves the use of software that can be used in education. The results show contributions of this sample use for the creation of teaching proposals, focusing on the language in use, identification of the sociocultural factors that influence the emergence and permanence of linguistic variation and researches in the scope of natural languages.