Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18265 papers
Dialectology and Dutch syntax from the perspective of linguistic (norm) change Jeroen van Craenenbroeck demonstrates in a highly convincing way that both synchronous descriptions and theoretical approaches of the syntax of Standard Dutch can be optimized by including analyses of dialect variation. Besides a minor reservation with respect to van Craenenbroeck’s interpretation of the conjugation of conjunctions, this response adds a complementary perspective to van Craenenbroeck’s overall argumentation by arguing that the inclusion of geolinguistic research on dialect variation is also indispensable for the diachronic study of Dutch syntax.
The paper presents a short introduction to several electronic resources for Ukrainian language, namely, two treebanks: the Gold standard (ab. 130 thousand tokens), manually annotated in the Universal Dependencies flavour (https://universaldependencies.org/), which comprises the training data for a machine-trained syntactic parser, and a big (near 3 billion tokens),
Many long short-term memory (LSTM) applications need fast yet compact models. Neural network compression approaches, such as the grow-and-prune paradigm, have proved to be promising for cutting down network complexity by skipping insignificant weights. However, current compression strategies are mostly hardware-agnostic and network complexity reduction does not always translate into execution efficiency. In this work, we propose a hardware-guided symbiotic training methodology for compact, accurate, yet execution-efficient inference models. It is based on our observation that hardware may introduce substantial non-monotonic behavior, which we call the latency hysteresis effect, when evaluating network size vs. inference latency. This observation raises question about the mainstream smaller-dimension-is-better compression strategy, which often leads to a sub-optimal model architecture. By leveraging the hardware-impacted hysteresis effect and sparsity, we are able to achieve the symbiosis of model compactness and accuracy with execution efficiency, thus reducing LSTM latency while increasing its accuracy. We have evaluated our algorithms on language modeling and speech recognition applications. Relative to the traditional stacked LSTM architecture obtained for the Penn Treebank dataset, we reduce the number of parameters by 18.0x (30.5x) and measured run-time latency by up to 2.4x (5.2x) on Nvidia GPUs (Intel Xeon CPUs) without any accuracy degradation. For the DeepSpeech2 architecture obtained for the AN4 dataset, we reduce the number of parameters by 7.0x (19.4x), word error rate from 12.9% to 9.9% (10.4%), and measured run-time latency by up to 1.7x (2.4x) on Nvidia GPUs (Intel Xeon CPUs). Thus, our method yields compact, accurate, yet execution-efficient inference models.
The slowness of legal proceedings in the common law legal system is a widely known fact. Any tool which could help reduce the time taken for the resolution of a case is invaluable. Common legal systems place a great importance on precedents and retrieving the correct set of precedents is considerably time consuming. Hence, for any case whose proceedings are in progress, if there are suitable prior cases, then the court has to follow the same interpretations that were passed in the prior cases. This is to ensure that similar situations receive similar treatment, thus maintaining uniformity amongst the legal proceedings across all courts at all times. Hence, precedent cases are treated as important as any other written law (a statute) in this legal system. In this paper, we propose two new approaches to solve this information retrieval problem wherein the system accepts the current case document as the query and returns the relevant precedent cases as the result. The first approach is to calculate the document similarity using Wordnet, which is a lexical database that could be leveraged to quantify the semantic relatedness between two documents, using a semantic network. The second approach is the use of a Siamese Manhattan Long Short Term Memory network, which is a supervised model trained to understand the underlying similarity between two documents.
This article contends that researchers can and should be active participants in making sound archives more accessible. In fact, such advocacy can be consequential in setting up possibilities for creative research on race within radio history and sound studies. Using the example of NPR’s All Things Considered archive spanning from 1971–1983, I demonstrate how academics and archivists can work together to make possible the preservation and accessibility of sound archives. This particular collaboration offers an opportunity to take a cultural approach to understanding newsroom diversity, more specifically: the cultural constraints of linguistic norms and the emergent cultures that arise as aberrations from such norms. The article reflects on this project’s implications for other scholars who work with archives that wish to invest in sound archive preservation and use.
The Ministry of Education and Science of the Russian Federation has established a score-rating system to assess the achievement quality within academic disciplines. The word "rating" is a foreign-language term, literal translation into Russian means "assessment". The second meaning of the word "rating" is a numerical measure; it characterizes performance of a student, pupil, etc., during a certain period of training, usually by the 20-point scale, or 100-point scale. This article highlights the history of formation and implementation of a score-rating system into the curriculum. Due to development of advanced technologies, there is a need for training highly qualified personnel to solve increasingly complex problems in professional activities. In our opinion, pedagogical assessment of student's knowledge in the form of score-rating system is a key role for preparing highly qualified personnel. The authors study processes of training skilled professionals in the course of academic education in technical universities of our country. Key words: score-rating system (SRS), competence, student, progress, credit, rating, factor.
Although SGD requires shuffling the training data between epochs, currently none of the word-level language modeling systems do this. Naively shuffling all sentences in the training data would not permit the model to learn inter-sentence dependencies. Here we present a method that partially shuffles the training data between epochs. This method makes each batch random, while keeping most sentence ordering intact. It achieves new state of the art results on word-level language modeling on both the Penn Treebank and WikiText-2 datasets.
Music has been shown to influence the behavioral responses of individuals in real-world scenarios, but little research exists on the effects that music has on the in-game behaviors of video game players. A song can be rated in terms of the level of arousal, or emotional intensity, it incites, and this study explores how music of various arousal ratings can be used to influence players' choices in an interactive narrative role-playing game. We hypothesized that high-arousal music would influence players to exhibit avoidance behaviors in-game, and that low-arousal music would influence players to exhibit social behaviors. Experimentation showed that players were statistically significantly more likely to make avoidance behavior choices when high-arousal music was played. These findings are the first step into understanding how music can be used by game developers to influence player behaviors in interactive narrative games.
Music is hierarchically structured, both in how it is perceived by listeners and how it is composed. Such structure can be captured elegantly using probabilistic grammatical models similar to those used to study natural language. They address the complexity of the structure using abstract categories in a recursive formalism. Most existing grammatical models of musical structure focus on one single dimension of music--such as melody, harmony, or rhythm. While these grammar models often work well on short musical excerpts, accurate analysis of longer pieces requires taking into account the constraints from multiple domains of structure. The present paper proposes abstract product grammars--a formalism which integrates multiple dimensions of musical structure into a single grammatical model--along with efficient parsing and inference algorithms for this formalism. We use this model to study the combination of hierarchically-structured harmonic syntax and hierarchically-structured rhythmic information. The latter is modeled by a novel grammar of rhythm that is capable of expressing temporal regularities in musical phrases. It integrates grouping structure and meter. The combined model of harmony and rhythm outperforms both single-dimension models in computational experiments. All models are trained and evaluated on a treebank of hand-annotated Jazz standards.
The goal of this paper is to use all available Polish language data sets to seek the best possible performance in supervised sentiment analysis of short texts. We use text collections with labeled sentiment such as tweets, movie reviews and a sentiment treebank, in three comparison modes. In the first, we examine the performance of models trained and tested on the same text collection using standard cross-validation (in-domain). In the second we train models on all available data except the given test collection, which we use for testing (one vs rest cross-domain). In the third, we train a model on one data set and apply it to another one (one vs one cross-domain). We compare wide range of methods including machine learning on bag-of-words representation, bidirectional recurrent neural networks as well as the most recent pre-trained architectures ELMO and BERT. We formulate conclusions as to cross-domain and in-domain performance of each method. Unsurprisingly, BERT turned out to be a strong performer, especially in the cross-domain setting. What is surprising however, is solid performance of the relatively simple multinomial Naive Bayes classifier, which performed equally well as BERT on several data sets.
Dependency grammar induction is the task of learning dependency syntax without annotated training data. Traditional graph-based models with global inference achieve state-ofthe-art results on this task but they require O(n3) run time. Transition-based models enable faster inference with O(n) time complexity, but their performance still lags behind. In this work, we propose a neural transition-based parser for dependency grammar induction, whose inference procedure utilizes rich neural features with O(n) time complexity. We train the parser with an integration of variational inference, posterior regularization and variance reduction techniques. The resulting framework outperforms previous unsupervised transition-based dependency parsers and achieves performance comparable to graph-based models, both on the English Penn Treebank and on the Universal Dependency Treebank. In an empirical comparison, we show that our approach substantially increases parsing speed over graphbased models.
Highly regularized LSTMs achieve impressive results on several benchmark datasets in language modeling. We propose a new regularization method based on decoding the last token in the context using the predicted distribution of the next token. This biases the model towards retaining more contextual information, in turn improving its ability to predict the next token. With negligible overhead in the number of parameters and training time, our Past Decode Regularization (PDR) method improves perplexity on the Penn Treebank dataset by up to 1.8 points and by up to 2.3 points on the WikiText-2 dataset, over strong regularized baselines using a single softmax. With a mixture-of-softmax model, we show gains of up to 1.0 perplexity points on these datasets. In addition, our method achieves 1.169 bits-per-character on the Penn Treebank Character dataset for character level language modeling. Each of these results constitute improvements over models without PDR in their respective settings.
We present a novel semantic framework for modeling linguistic expressions of generalization— generic, habitual, and episodic statements—as combinations of simple, real-valued referential properties of predicates and their arguments. We use this framework to construct a dataset covering the entirety of the Universal Dependencies English Web Treebank. We use this dataset to probe the efficacy of type-level and token-level information—including hand-engineered features and static (GloVe) and contextual (ELMo) word embeddings—for predicting expressions of generalization.
This paper presents a novel crowd-sourced resource for multimodal discourse: our resource characterizes inferences in image-text contexts in the domain of cooking recipes in the form of coherence relations. Like previous corpora annotating discourse structure between text arguments, such as the Penn Discourse Treebank, our new corpus aids in establishing a better understanding of natural communication and common-sense reasoning, while our findings have implications for a wide range of applications, such as understanding and generation of multimodal documents.
Drawing on linguistic ethnographic data analysis, this article aims to expand the Rampton’s concept of ‘language crossing’ through integrating the notion of ‘linguistic racism’ experienced by Mongolian background immigrant women in Australia. These women encounter linguistic homogeneity, discrimination, and alienation in varied ways in their daily institutional and non-institutional settings based on how they speak English or their usage of heritage languages. As a result, they establish everyday linguistic resistance strategies to combat linguistic racism, which further add two new dimensions to the concept of language crossing – ‘crossing as a resistance strategy’ and ‘crossing as a passing strategy’. Adopting these crossing strategies allow these women to use their preferred forms of communication to resist dominant linguistic norms and standards in the dominant culture. These strategies further make it possible for these speakers to pass as the native speakers of that dominant language. Finally, the paper argues that it is almost impossible to understand ‘language crossing’ as a discrete understanding isolated from the concept of ‘linguistic racism’. It is better to examine these concepts together, as they seem to complement each other in terms of investigating the everyday linguistic practices, sociolinguistic realities and struggles that these immigrant women encounter.
Purpose Verbs with low concreteness are frequent in discourse samples but rarely targeted in aphasia treatments for verbs. These verbs are an important part of functional communication, and recent studies have called for more research regarding aphasia and treatment stimuli with low concreteness. The aim of this study was to pilot the use of verbs with low concreteness in a novel sentence production intervention with persons with aphasia. Method The study took the form of a single-case experimental design with multiple baselines across behaviors and across participants. Three persons with chronic nonfluent aphasia and apraxia of speech participated in the study. Each participant received treatment designed to increase the semantic networks of verbs with high frequency and low concreteness. Sentence production was closely examined over the course of treatment for treated and untreated verbs of varying concreteness levels. Additional measures of language and cognitive functioning were also taken before and after treatment. Results Results indicated improved sentence production with target verbs attributable to the treatment in the 1st phase of 2 phases for 2 of the 3 participants. The increases corresponded with the application of treatment, despite the difference in number of baseline sessions for the participants. Where there were treatment effects, there was also considerable generalization to untreated sets of items during the 1st treatment phase. Word retrieval also improved for 2 participants. Conclusions The results suggest that the novel treatment may improve sentence production and word retrieval in persons with aphasia, even when using target verbs with low concreteness ratings. Future research is warranted into the use of low concreteness verbs. Supplemental Material https://doi.org/10.23641/asha.10870958.
Abstract This chapter poses the question of whether humans might be essentially normative animals, i.e. whether traditionally prominent specificities of the human life form—our linguistic, social, and moral “natures”—might ground in a basic susceptibility, or proclivity to the deontic regulation of thought and behaviour: the “normative animal thesis.” The chapter lays out the issues at stake in attempting to answer this question. It divides into two main parts. The first begins by clarifying the—norm-related—concept of normativity at issue, distinguishing it from the—reason-related—conceptualisation current in meta-ethics and theories of rationality. It then discusses the primary candidates for generic features of norms, before dividing the normative animal thesis into various sub-claims. The second part presents the key questions at issue in the discussion of social, moral, and linguistic norms, comparing ways of conceiving them and marking the significance of such conceptualisations for the normative animal thesis.
Classifying patients' affect is a pivotal part of the mental status examination. However, this common practice is often widely inconsistent between raters. Recent advances in the field of Facial Action Recognition (FAR) have enabled the development of tools that can act to identify facial expressions from videos. In this study, we aimed to explore the potential of using machine learning techniques on FAR features extracted from videotaped semi-structured psychiatric interviews of 25 male schizophrenia inpatients (mean age 41.2 years, STD = 11.4). Five senior psychiatrists rated patients' affect based on the videos. Then, a novel computer vision algorithm and a machine learning method were used to predict affect classification based on each psychiatrist affect rating. The algorithm is shown to have a significant predictive power for each of the human raters. We also found that the eyes facial area contributed the most to the psychiatrists' evaluation of the patients' affect. This study serves as a proof-of-concept for the potential of using the machine learning FAR system as a clinician-supporting tool, in an attempt to improve the consistency and reliability of mental status examination.
70% of patients with schizophrenia suffer from auditory verbal hallucinations (AVH) which are frequently described as distressing and disabling. The content of AVH, in relation to internal thought, has never been linguistically tested in a self-monitoring study. The aim of this preliminary study was to establish if there was a significant difference between AVH and inner thoughts on the key linguistic parameters of valence (pleasantness), dominance (control) and arousal (intensity of emotion produced). Six volunteers with a diagnosis of schizophrenia from voice hearing support groups produced real-time, detailed diaries of AVH and inner thoughts using randomised/fixed timers. Analysis of content was completed using an established linguistic database. AVH were significantly more unpleasant and controlling but not more emotionally arousing than inner thoughts. Psychoeducation around the experience of hallucination in schizophrenia should include information that the voices will be significantly more unpleasant and controlling than their own thoughts but not more emotionally arousing. CBT might therefore include the use of compassion focussed techniques to help with the unpleasantness of AVH and schema level techniques to improve coping with the dominance of AVH.
PURPOSE A culturally appropriate, patient-centered measure of the quality of dying and death is needed to advance palliative care in Africa. We therefore evaluated the Quality of Dying and Death Questionnaire (QODD) in a Kenyan hospice sample and compared item ratings with those from a Canadian advanced-cancer sample. METHODS Caregivers of deceased patients from three Kenyan hospices completed the QODD. Their QODD item ratings were compared with those from 602 caregivers of deceased patients with advanced cancer in Ontario, Canada, and were correlated with overall quality of dying and death ratings. RESULTS Compared with the Ontario sample, outcomes in the Kenyan sample (N = 127; mean age, 48.21 years; standard deviation, 13.57 years) were worse on 14 QODD concerns and on overall quality of dying and death ( P values ≤.001) but better on five concerns, including interpersonal and religious/spiritual concerns ( P values ≤.005). Overall quality of dying was associated with better patient experiences with Symptoms and Personal Care, interpersonal, and religious/spiritual concerns ( P values <.01). Preparation for Death, Treatment Preferences, and Moment of Death items showed the most omitted ratings. CONCLUSION The quality of dying and death in Kenya is worse than in a setting with greater PC access, except in interpersonal and religious/spiritual domains. Cultural differences in perceptions of a good death and the acceptability of death-related discussions may affect ratings on the QODD. This measure requires revision and validation for use in African settings, but evidence from such patient-centered assessment tools can advance palliative care in this region.
In recent years, there has been increasing demand for automatic architecture search in deep learning. Numerous approaches have been proposed and led to state-of-the-art results in various applications, including image classification and language modeling. In this paper, we propose a novel way of architecture search by means of weighted networks (WeNet), which consist of a number of networks, with each assigned a weight. These weights are updated with back-propagation to reflect the importance of different networks. Such weighted networks bear similarity to mixture of experts. We conduct experiments on Penn Treebank and WikiText-2. We show that the proposed WeNet can find recurrent architectures which result in state-of-the-art performance.
Using a sample of 115 first-generation married immigrant Pakistani residing in Toronto, Canada, this study assessed the moderating effect of couples’ communication patterns (warm and moderate hostility) and gender role beliefs (transcendence and gender-linked beliefs) on the link between acculturative stress and satisfaction with marriage. Participants completed a demographic information sheet, the Multidimensional Acculturative Stress Scale (MASS), the Social Roles Questionnaire (SRQ), the Behavioral Affect Rating Scale (BARS) and the Relationship Assessment Scale (RAS). Regression analyses indicated that gender-linked beliefs and hostile styles of communication played a moderating role between acculturative stress and marital satisfaction, decreasing marital satisfaction, while gender-transcendent beliefs and a warm style of communication supported the relationship and enhanced the marital satisfaction. Clinicians and researchers need to address the communication styles and beliefs of Pakistani immigrants within their marital relationship to get a fuller picture of it.
This thesis presents several studies in neural dependency parsing for typologically diverse languages, using treebanks from Universal Dependencies (UD). The focus is on informing models with linguistic knowledge. We first extend a parser to work well on typologically diverse languages, including morphologically complex languages and languages whose treebanks have a high ratio of non-projective sentences, a notorious difficulty in dependency parsing. We propose a general methodology where we sample a representative subset of UD treebanks for parser development and evaluation. Our parser uses recurrent neural networks which construct information sequentially, and we study the incorporation of a recursive neural network layer in our parser. This follows the intuition that language is hierarchical. This layer turns out to be superfluous in our parser and we study its interaction with other parts of the network. We subsequently study transitivity and agreement information learned by our parser for auxiliary verb constructions (AVCs). We suggest that a parser should learn similar information about AVCs as it learns for finite main verbs. This is motivated by work in theoretical dependency grammar. Our parser learns different information about these two if we do not augment it with a recursive layer, but similar information if we do, indicating that there may be benefits from using that layer and we may not yet have found the best way to incorporate it in our parser. We finally investigate polyglot parsing. Training one model for multiple related languages leads to substantial improvements in parsing accuracy over a monolingual baseline. We also study different parameter sharing strategies for related and unrelated languages. Sharing parameters that partially abstract away from word order appears to be beneficial in both cases but sharing parameters that represent words and characters is more beneficial for related than unrelated languages.
Four experiments examined the claims that people can intuitively assess the logical validity of arguments, and that qualitatively different reasoning processes drive intuitive and explicit validity assessments. In each study participants evaluated arguments varying in validity and believability using either deductive criteria (logic task) or via an intuitive, affective response (liking task). Experiment 1 found that people are sensitive to argument validity on both tasks, with valid arguments receiving higher liking as well as higher deductive ratings than invalid arguments. However, the claim that this effect is driven by logical intuitions was challenged by the finding that sensitivity to validity in both liking and logic tasks was affected in similar ways by manipulations of concurrent memory load (Experiments 1 and 2) and variations in individual working memory capacity (Experiments 3 and 4). In both tasks better discrimination between valid and invalid arguments was found when more working memory resources were available. Formal signal detection models of reasoning were tested against the experimental data using signed difference analysis (Stephens, Dunn, & Hayes, 2018b). A single-process reasoning model which assumes that argument evaluation in both logic and liking tasks involves a single latent dimension for assessing argument strength but different response criteria for each task, was found to be consistent with the data from each experiment (as were some dual-process models). The experimental and modeling results confirm that people are sensitive to argument validity in both explicit logic and affect rating tasks, but that these results can be explained by a single underlying reasoning process. (PsycInfo Database Record (c) 2020 APA, all rights reserved).
Abstract We describe systematic changes that have been made to the Czech morphological dictionary related to annotating new data within the project of Prague Dependency Treebank (PDT). We bring new solutions to several complicated morphological features that occur in Czech texts. We introduced two new parts of speech, namely foreign word and segment. We adopted new principles for morphological analysis of global and inflectional variants, homonymous lemmas, abbreviations and aggregates. The changes were initiated by the need of consistency between the data and the dictionary and of the dictionary itself.
Many advances in Natural Language Processing have been based upon more expressive models for how inputs interact with the context in which they occur. Recurrent networks, which have enjoyed a modicum of success, still lack the generalization and systematicity ultimately required for modelling language. In this work, we propose an extension to the venerable Long Short-Term Memory in the form of mutual gating of the current input and the previous output. This mechanism affords the modelling of a richer space of interactions between inputs and their context. Equivalently, our model can be viewed as making the transition function given by the LSTM context-dependent. Experiments demonstrate markedly improved generalization on language modelling in the range of 3-4 perplexity points on Penn Treebank and Wikitext-2, and 0.01-0.05 bpc on four character-based datasets. We establish a new state of the art on all datasets with the exception of Enwik8, where we close a large gap between the LSTM and Transformer models.
Many adaptive gradient methods have been successfully applied to train deep neural networks, such as Adagrad, Adadelta, RMSprop and Adam. These methods perform local optimization with an element-wise scaling learning rate based on past gradients. Although these methods can achieve an advantageous training loss, some researchers have pointed out that their generalization capability tends to be poor as compared to stochastic gradient descent (SGD) in many applications. These methods obtain a rapid initial training process but fail to converge to an optimal solution due to the unstable and extreme learning rates. In this paper, we investigate the adaptive gradient methods and get the insights on various factors that may lead to poor performance of Adam. To overcome that, we propose a bounded scheduling algorithm for Adam, which can not only improve the generalization capability but also ensure the convergence. To validate our claims, we carry out a series of experiments on the image classification and the language modeling tasks on several standard benchmarks such as ResNet, DenseNet, SENet and LSTM on typical data sets such as CIFAR-10, CIFAR-100 and Penn Treebank. Experimental results show that our method can eliminate the generalization gap between Adam and SGD, meanwhile maintaining a relative high convergence rate during training.
To create emotionally expressive robots, designers of human-robot interaction routinely translate emotion theories into instruments through which we estimate, quantify and analyze human emotional responses to robot behaviour. Pragmatically, we often use straightforward models such as Russell's circumplex, treating emotion as a single point in a two-dimensional space. However, this simple metaphor and its consequent representations omit many aspects of real emotional experience, can lead to erroneous data and may undermine computational models that rely on them. Problems with emotion representations currently prevalent in human-robot interaction fall into three categories: (1)Representations are static and singular, whereas real emotions can be dynamic, multi-valued, uncertain or conflicting. (2)The framing of an interaction is unspecified (i.e., in an affective rating task: which part of an interaction involving multiple parties and perspectives the participant is meant to consider). (3) Participant responses captured with instruments and methods that are not well-understood by experimenters nor participants produce data that is hard to interpret. We propose alternative emotion representations to account for dynamic emotions inherent in interactive contexts; scrutinize framing ambiguities in study tasks and argue for mixed-methods approaches to achieve shared understanding of emotion representations between participants and researchers.
Organizing multivariate data spaces by their dimensions or attributes can be a rather difficult task. Most of the work in this area focuses on the statistical aspects such as correlation clustering, dimension reduction, and the like. These methods typically produce hierarchies in which the leaf nodes are labeled by the attribute names while the inner nodes are often represented by just a statistical measure and criterion, such as a threshold. This makes them difficult to understand for mainstream users. Taxonomies in science, biology, engineering, etc. on the other hand, are easy to comprehend since they provide meaningful labels at the inner nodes as well. Labeling inner nodes of taxonomies automatically requires the identification of hypernyms. Our proposed framework, called Taxonomizer, takes a visual analytics approach to meet this challenge. It appeals to the wisdom of humans to liaise with state of the art data analytics, neural word embeddings, and lexical databases. It consists of a set of visual tools that starts out with an automatically computed hierarchy where the leaf nodes are the original data attributes, and it then allows users to sculpt high-quality taxonomies for any multivariate dataset.
Many advances in Natural Language Processing have been based upon more expressive models for how inputs interact with the context in which they occur. Recurrent networks, which have enjoyed a modicum of success, still lack the generalization and systematicity ultimately required for modelling language. In this work, we propose an extension to the venerable Long Short-Term Memory in the form of mutual gating of the current input and the previous output. This mechanism affords the modelling of a richer space of interactions between inputs and their context. Equivalently, our model can be viewed as making the transition function given by the LSTM context-dependent. Experiments demonstrate markedly improved generalization on language modelling in the range of 3-4 perplexity points on Penn Treebank and Wikitext-2, and 0.01-0.05 bpc on four character-based datasets. We establish a new state of the art on all datasets with the exception of Enwik8, where we close a large gap between the LSTM and Transformer models.
The Penn Treebank (PTB) represents syntactic structures as graphs due to nonlocal dependencies. This paper proposes a method that approximates PTB graph-structured representations by trees. By our approximation method, we can reduce nonlocal dependency identification and constituency parsing into single treebased parsing. An experimental result demonstrates that our approximation method with an off-the-shelf tree-based constituency parser significantly outperforms the previous methods in nonlocal dependency identification.
The paper studies the effect of emotional states modulated by auditory stimuli on the cognitive control on decision making. Based on other previous neuroimaging studies, functional near-infrared spectroscopy provided reliable neuroimaging measurement in analyzing emotional states by studying the changes of hemodynamic response in prefrontal cortex (PFC). This experiment involved 16 nursing students. During the experiment, participants were given one minute to complete five nursing practice questions with five sequential repetitions in the presence of neutral and negative emotional auditory stimuli in two separated sessions under fNIRS measurement. The sound stimuli was selected from the International Affective Digitized Sound (IADS) System. The neutral auditory stimuli had neutral valence and medium arousal rating whereas negative auditory stimuli had negative valence and high arousal rating. The data collected was preprocessed by using wavelet transform to decompose the data into different frequency intervals. By selecting the frequency interval of interest, we analyzed the data based on functional connectivity within prefrontal cortex regions. We computed the regional wavelet coherence values between affective and neutral tasks. From the behavioral analysis, we found that subjects had significantly higher accuracy in affective task compared to neutral task. Based on the analysis, we found that left prefrontal cortex produced significantly lower wavelet coherence value but the highest coherence-accuracy correlation in affective task than in neutral task.
Characterizing the distribution of crossing dependencies in natural language dependency trees is a crucial task for building parsers and understanding the formal properties of human language. A number of formal restrictions on crossing dependencies have been proposed, including bounds on gap degree, edge degree, and end-point crossings. Here we ask whether the empirical distribution of crossing dependencies in dependency treebanks offers evidence for these formal restrictions as true, independent constraints on dependency trees, or whether the distribution can be explained using other, more generic constraints affecting dependency trees. Specifically, we explore the null hypothesis that crossing dependencies are formally unrestricted, but occur at a low rate. We implement the null hypothesis using random trees where crossing dependencies occur at the same rate as in natural language trees, but without any formal restrictions. We find that this baseline generally does not reproduce the same distribution of gap degree, edge degree, endpoint-crossing, and heads' depth difference as real trees, suggesting that these formal constraints are a consequence of factors beyond the rate of crossing dependencies alone.
Mirror-sensory synaesthetes mirror the pain or touch that they observe in other people on their own bodies. This type of synaesthesia has been associated with enhanced empathy. We investigated whether the enhanced empathy of people with mirror-sensory synesthesia influences the experience of situations involving touch or pain and whether it affects their prosocial decision making. Mirror-sensory synaesthetes ( N = 18, all female), verified with a touch-interference paradigm, were compared with a similar number of age-matched control individuals (all female). Participants viewed arousing images depicting pain or touch; we recorded subjective valence and arousal ratings, and physiological responses, hypothesizing more extreme reactions in synaesthetes. The subjective impact of positive and negative images was stronger in synaesthetes than in control participants; the stronger the reported synaesthesia, the more extreme the picture ratings. However, there was no evidence for differential physiological or hormonal responses to arousing pictures. Prosocial decision making was assessed with an economic game assessing altruism, in which participants had to divide money between themselves and a second player. Mirror-sensory synaesthetes donated more money than non-synaesthetes, showing enhanced prosocial behaviour, and also scored higher on the Interpersonal Reactivity Index as a measure of empathy. Our study demonstrates the subjective impact of mirror-sensory synaesthesia and its stimulating influence on prosocial behaviour. This article is part of the discussion meeting issue ‘Bridging senses: novel insights from synaesthesia’.
Automatic emotion regulation (AER) is an important type of emotion regulation in our daily life. Most of the previous studies concerning AER are done in the conscious level. Little is known about the AER under the subliminal level. The present study was to investigate the AER at the different perceptual levels (i.e., explicitly and implicitly) simultaneously, and the associated neural differences using functional magnetic resonance imaging. Priming paradigm was adopted in which the inhibition or neutral words were used as primes and the negative picutres were used as targets. In the experiment, the duration time of priming words was manipulated at 33 or 50 ms in the implicit level and 3000 ms in the explicit level. The participants were required to make emotional valence rating of the negative pictures while undergoing functional magnetic resonance imaging scanning. The results showed that the participants experienced less negative emotion in inhibition words priming condition contrary to neutral words priming condition. Significant differences were also found in the left ventrolateral prefrontal cortex and left dorsolateral prefrontal cortex at the implicit and explicit AER. The findings of this study demonstrate that inhibition words can automatically and effectively reduce an individual's negative emotion experience, and left ventrolateral prefrontal cortex and left dorsolateral prefrontal cortex have been both implicated in self-control during AER.
Alterations in fear learning/generalization are considered to be relevant mechanisms engendering the development of anxiety disorders being the most prevalent mental disorders. Although anxiety disorders almost exclusively have their first onset in childhood and adolescence, etiological research focuses on adult individuals. In this study, we evaluated findings of a recent meta-analysis of genome-wide association studies in adult anxiety disorders with significant associations of four single nucleotide polymorphisms (SNPs) in a large cohort of 347 healthy children (8-12 years) characterized for dimensional anxiety. We investigated the modulation of anxiety parameters by these SNPs in a discriminative fear conditioning and generalization paradigm in the to-date largest sample of children. Results extended findings of the meta-analysis showing a genomic locus on 2p21 to modulate anxious personality traits and arousal ratings. These SNPs might, thus, serve as susceptibility markers for a shared risk across pathological anxiety, presumably mediated by alterations in arousal.
This paper describes a novel approach for the task of end-to-end argument labeling in shallow discourse parsing. Our method describes a decomposition of the overall labeling task into subtasks and a general distance-based aggregation procedure. For learning these subtasks, we train a recurrent neural network and gradually replace existing components of our baseline by our model. The model is trained and evaluated on the Penn Discourse Treebank 2 corpus. While it is not as good as knowledge-intensive approaches, it clearly outperforms other models that are also trained without additional linguistic features.
This Sentiment analysis is mainly found in the user's social platform for a hot event or product point of view and attitude. Most existing sentiment analysis approaches heavily rely on a large amount of labeled data that usually involve time-consuming and error-prone manual annotations. In order to avoid the dependence on the manual annotation dictionary and reduce the human intervention in the machine learning process, In this paper, Based on the researches on sentiment analysis and deep learning, we propose a hybrid framework AM-Bi-LSTM that combines Attention Mechanism and Bi-directional Long-Short-Term Memory (Bi-LSTM) neural networks for sentence classification. We demonstrate the effectiveness and efficiency of our approach on a representative Stanford Sentiment Treebank (SST) dataset. For SST-1 and SST-2, Compared with the currently published state-of-the-art methods Conv-RNN, the accuracy of AM-Bi-LSTM is improved by 2.787% and 1.946% respectively.
In this article, we tackle the issue of the limited quantity of manually\nsense annotated corpora for the task of word sense disambiguation, by\nexploiting the semantic relationships between senses such as synonymy,\nhypernymy and hyponymy, in order to compress the sense vocabulary of Princeton\nWordNet, and thus reduce the number of different sense tags that must be\nobserved to disambiguate all words of the lexical database. We propose two\ndifferent methods that greatly reduces the size of neural WSD models, with the\nbenefit of improving their coverage without additional training data, and\nwithout impacting their precision. In addition to our method, we present a WSD\nsystem which relies on pre-trained BERT word vectors in order to achieve\nresults that significantly outperform the state of the art on all WSD\nevaluation tasks.\n
This present pilot study investigates the relationship between dependency distance and frequency based on the analysis of an English dependency treebank. The preliminary result shows that there is a non-linear relation between dependency distance and frequency. This relation between them can be further formalized as a power law function which can be used to predict the distribution of dependency distance in a treebank.
Perceived self-efficacy refers to a subject's expectation about the outcomes his/her behavior will have in a challenging situation. Low self-efficacy has been implicated in the origins and maintenance of phobic behavior. Correlational studies suggest an association between perceived self-efficacy and learning. The experimental manipulation of perceived self-efficacy offers an interesting approach to examine the impact of self-efficacy beliefs on cognitive and emotional functions. Recently, a positive effect of an experimentally induced increased self-efficacy on associative learning has been demonstrated. Changes in associative learning constitute a central hallmark of pathological fear and anxiety. Such alterations in the acquisition and extinction of conditioned fear may be related to cognitive and neurobiological factors that predict a certain vulnerability to anxiety disorders. The present study builds on previous own work by investigating the effect of an experimentally induced low perceived self-efficacy on fear acquisition, extinction and extinction retrieval in a differential fear conditioning task. Our results suggest that a negative verbal feedback, which leads to a decreased self-efficacy, is associated with changes in the acquisition of conditioned fear. During fear acquisition, the negative verbal feedback group showed decreased discrimination of fear responses between the aversive and safe conditioned stimuli (CS) relative to a group receiving a neutral feedback. The effects of the negative verbal feedback on the acquisition of fear discrimination learning were indexed by an impaired ability to discriminate the probability of receiving a shock during acquisition upon presentation of the aversive (CS+) relative to the safe stimuli (CS-). However, the effects of low self-efficacy on discrimination learning were limited to fear acquisition. No differences between the groups were observed during extinction and extinction retrieval. Furthermore, analysis of other outcome measures, i.e., skin conductance responses and CS valence ratings, revealed no group differences during the different phases of fear conditioning. In conclusion, lower perceived self-efficacy alters cognitive/expectancy components of discrimination during fear learning but not evaluative components and physiological responding. The pattern of findings suggests a selective, detrimental role of low(er) self-efficacy on the subject's ability to learn the association between ambiguous cues and threat/safety.
OBJECTIVE: Biased attention for disorder-relevant information plays a crucial role in the maintenance of different mental disorders including eating disorders and might be of use to define recovery beyond symptom-related criteria. METHOD: We assessed attention deployment using eye tracking in a cued choice viewing paradigm to two different categories of disorder-relevant stimuli in 24 individuals with acute anorexia nervosa (AN), 20 weight-recovered individuals with a history of AN (WRAN) and 23 healthy control participants (CG). Picture pairs consisted of a food stimulus or a picture depicting physical activity and a matched control stimulus (household item/physical inactivity). Participants rated the valence of stimuli afterwards. RESULTS: The groups did not differ in initial attention deployment. In later processing stages, AN patients showed a generalized attentional avoidance of food and control pictures as compared to CG, while WRAN individuals were in between. AN patients showed an attentional bias toward physical activity pictures as compared to WRAN individuals, but not the CG. AN individuals rated the food pictures and the pictures showing physical inactivity as less pleasant than the CG, while WRAN individuals were in between. DISCUSSION: Attention deployment is partly changed in WRAN as compared to the acute AN group, especially with regard to a shift away from illness-compatible stimuli (physical activity), and this might be a useful recovery criterion. Valence rating of food stimuli might be an additional useful tool to distinguish between acutely ill and weight-recovered individuals. Attentional biases for illness-compatible stimuli might qualify as a valuable approach to defining recovery in AN.
This paper is a linguistic as well as technical survey for the development of a shallow discourse parser for Czech. It focuses on long-distance discourse relations signalled by (mostly) anaphoric discourse connectives. Proceeding from the division of connectives on “structural” and “anaphoric” according to their (in)ability to accept distant (non-adjacent) text segments as their left-sided arguments, and taking into account results of related analyses on English data in the framework of the Penn Discourse Treebank, we analyze a large amount of language data in Czech. We benefit from the multilayer manual annotation of various language aspects from morphology to discourse, coreference and bridging relations in the Prague Dependency Treebank 3.0. We describe the linguistic parameters of long-distance discourse relations in Czechin connection with their anchoring connective, and suggest possible ways of their detection. Our empirical research also outlines some theoretical consequences for the underlying assumptions in discourse analysis and parsing, e.g. the risk of relying too much on different (language-specific?) part-of-speech categorizations of connectives or the different perspectives in shallow and global discourse analyses (the minimality principle vs. higher text structure).
We present a comparative analysis of PP ordering in English and (Mandarin) Chinese, two languages with distinct typological word order characteristics. Previous work on PP orderings have mainly focused on English using data of relatively small size. Here we leverage corpora of much larger scale with straightforward annotations. We use the Penn Treebank for English, which includes three corpora that cover both written and spoken domains, and the Chinese Penn Treebank for Chinese. We explore the individual effect of dependency length, the argument status of the PP (argument or adjunct) and the traditional adverbial ordering rule, Manner before Place before Time. In addition, we evaluate the predictive power of dependency length and argument status with weights estimated from logistic regression models. We show that while dependency length plays a strong role across genre for English, it only exerts a mild effect in Chinese. On the other hand, the argument status of the PP has a pronounced role in both languages, that is, there exists a strong tendency for the argument-like PP to appear closer to the head verb than the adjunct-like PP. Our work contributes empirically to the long-standing proposal in linguistic typology that crosslinguistic word ordering preference is driven by cooperating and competing principles.