Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18265 papers
Tastes affect the body and our emotions. We used tasteless, sweet, and bitter stimuli to induce participants' moods, and we examined the effect of mood on an emotional evaluation of pleasant, neutral, and unpleasant images using event-related potentials, N2, N400, and late positive potential (LPP), which reflect emotional evaluation in the brain. The results indicated that mood valence was most positive for sweetness and most negative for bitterness. Moreover, there was no significant mood effect on subjective valence ratings of emotional images. Furthermore, the N2 amplitude, which is related to the early semantic processing of preceding stimuli, was unaffected by the taste induced mood. In contrast, we found that the N400 amplitude, which is related to the mismatch of emotional valence between stimuli, increased significantly for unpleasant images when participants were in a positive rather than negative mood state. Also, the LPP amplitude, which is related to the emotional valence of images, showed only the main effect of the images' emotional valence. The N2's results suggest that the early semantic processing of taste stimuli might have had a negligible impact on emotional evaluation because taste stimuli minimize semantic processing that accompanies mood induction. In contrast, the N400 reflected the effects of the induced mood, and the LPP reflected the impact of the valence of emotional images. The use of taste stimuli to induce mood revealed different brain processing of taste-induced mood effects on emotional evaluation, including N2's involvement in semantic processing, N400's involvement in matching emotions between mood and stimuli, and LPP's involvement in subjective evaluations of stimuli.
The Japanese CCGBank serves as training and evaluation data for developing Japanese CCG parsers. However, since it is automatically generated from the Kyoto Corpus, a dependency treebank, its linguistic validity still needs to be sufficiently verified. In this paper, we focus on the analysis of passive/causative constructions in the Japanese CCGBank and show that, together with the compositional semantics of ccg2lambda, a semantic parsing system, it yields empirically wrong predictions for the nested construction of passives and causatives.
Vision-Language Pre-training (VLP) has advanced the performance of many visionlanguage tasks, such as image-text retrieval, visual entailment, and visual reasoning. The pre-training mostly utilizes lexical databases and image queries in English. Previous work has demonstrated that the pre-training in English does not transfer well to other languages in a zero-shot setting. However, multilingual pre-trained language models (MPLM) have excelled at a variety of single-modal language tasks. In this paper, we propose a simple yet efficient approach to adapt VLP to unseen languages using MPLM. We utilize a cross-lingual contextualized token embeddings alignment approach to train text encoders for non-English languages. Our approach does not require image input and primarily uses machine translation, eliminating the need for target language data. Our evaluation across three distinct tasks (image-text retrieval, visual entailment, and natural language visual reasoning) demonstrates that this approach outperforms the state-of-the-art multilingual vision-language models without requiring large parallel corpora. Our code is available at https://github.com/Yasminekaroui/CliCoTea.
Hoarding disorder is characterised by the acquisition of, and failure to discard large numbers of items regardless of their actual value, a perceived need to save the items and distress associated with discarding them, significant clutter in living spaces that render the activities associated with those spaces very difficult causing significant distress or impairment in functioning. To aid development of an intervention for hoarding disorder we aimed to identify current practice by investigating key stakeholders existing practice regarding identification, assessment and intervention associated with people with hoarding disorder. Two focus groups with a purposive sample of 17 (eight male, nine female) stakeholders representing a range of services from housing, health, and social care were audio recorded, transcribed verbatim and analysed thematically. There was a lack of consensus regarding how hoarding disorder was understood and of the number of cases of hoarding disorder however all stakeholders agreed hoarding disorder appeared to be increasing. The clutter image rating scale was most used to identify people who needed help for hoarding disorder, in addition to other assessments relevant to the stakeholder. People with hoarding disorder were commonly identified in social housing where regular access to property was required. Stakeholders reported that symptoms of hoarding disorder were often tackled by enforced cleaning, eviction, or other legal action however these approaches were extremely traumatic for the person with hoarding disorder and failed to address the root cause of the disorder. While stakeholders reported there was no established services or treatment pathways specifically for people with hoarding disorder, stakeholders were unanimous in their support for a multi-agency approach. The absence of an established multiagency service that would offer an appropriate and effective pathway when working with a hoarding disorder presentation led stakeholders to work together to suggest a psychology led multiagency model for people who present with hoarding disorder. There is currently a need to examine the acceptability of such a model.
In this paper, we present a grammar-based natural language framework for robot programming, specifically for pick-and-place tasks. Our approach uses a custom dictionary of action words, designed to store together words that share meaning, allowing for easy expansion of the vocabulary by adding more action words from a lexical database. We validate our Natural Language Robot Programming (NLRP) framework through simulation and real-world experimentation, using a Franka Panda robotic arm equipped with a calibrated camera-in-hand and a microphone. Participants were asked to complete a pick-and-place task using verbal commands, which were converted into text using Google's Speech-to-Text API and processed through the NLRP framework to obtain joint space trajectories for the robot. Our results indicate that our approach has a high system usability score. The framework's dictionary can be easily extended without relying on transfer learning or large data sets. In the future, we plan to compare the presented framework with different approaches of human-assisted pick-and-place tasks via a comprehensive user study.
Abstract The present study analyzes the transformation of the vowel system and especially the process of vowel mergers based on the Latin inscriptions of the Gallic and Germanic provinces. With the help of the Computerized Historical Linguistic Database of the Latin Inscriptions of the Imperial Age ( http://lldb.elte.hu/ ), it tries to draw and then compare the phonological profiles of the selected provinces and to describe the dialectal position of Gaul and the Germanic provinces regarding vocalism in three periods (AD 1–300, 301–500 and 501–700). The analysis, which also covers comparisons with certain provinces of Italy, Spain and Dalmatia, is carried out considering four aspects: the ratio of vocalic versus consonantal changes, the ratio of vowel mergers compared to vocalic changes, the ratio of e-i and o-u mergers compared to each other, and the ratio of vowel mergers by stressed and unstressed syllable. As a result of the present study, it was revealed that Gallic provinces cannot be treated as a unit or as clearly separate from the other areas studied according to either aspect of the study, especially not in the early, pre-Christian period. Gallic provinces appear to behave in the same or a levelled manner at most in the later and/or latest periods. The Germanic provinces, especially Germania Superior, have, albeit with some delay, adapted to the Gallic provinces in their late development. The present study, which continued József Herman's research, managed to explore the hitherto little-known linguistic and dialectological features of Latin in the Gallic and Germanic provinces.
<h3>BACKGROUND AND PURPOSE:</h3> Brain atrophy is an important surrogate for brain reserve, the capacity of the brain to cope with acquired injuries such as acute stroke. It is unclear how well atrophy measurements on MR imaging can be reproduced using NCCT imaging. We aimed to compare pragmatic atrophy measures on NCCT with MR imaging in patients with acute ischemic stroke. <h3>MATERIALS AND METHODS:</h3> This is a post hoc analysis, including baseline NCCT and 24-hour follow-up MR imaging data from the Safety and Efficacy of Nerinetide (NA-1) in Subjects Undergoing Endovascular Thrombectomy for Stroke (ESCAPE-NA1) trial. Cortical atrophy was measured using the global cortical atrophy scale, and subcortical atrophy was measured using the intercaudate distance-to-inner-table width (CC/IT) ratio. Agreement and correlation between these measures on NCCT and MR imaging were calculated using the Gwet agreement coefficient 1 and Pearson correlation coefficients, respectively. <h3>RESULTS:</h3> Among 1105 participants in the ESCAPE-NA1 trial, interpretable NCCT and 24-hour MR imaging were available in 558 (50.5%) patients (mean age, 67.2 [SD, 13.7] years; 282 women). Cortical atrophy assessments performed on NCCT underestimated atrophy severity compared with MR imaging (eg, patients with global cortical atrophy of ≥1 assessed on NCCT = 133/558 [23.8%] and on MR imaging = 247/558 [44.3%]; a 20.5% difference). Overall, cortical (ie, global cortical atrophy) atrophy assessments on NCCT had substantial or better agreement with MR imaging (Gwet agreement coefficient 1 of > 0.784; <i>P </i><.001). Subcortical atrophy measures (CC/IT ratio) showed strong correlations between NCCT and MR imaging (Pearson correlation = 0.746, <i>P </i><.001). <h3>CONCLUSIONS:</h3> Brain atrophy can be evaluated using simple measures in emergently acquired NCCT. Subcortical atrophy assessments on NCCT show strong correlations with MR imaging. Although cortical atrophy assessments on NCCT are strongly correlated with MR imaging ratings, there is a general underestimation of atrophy severity on NCCT.
Genre classification for text documents is useful in media monitoring and detection of misinformation. Recent work in text classification for genre has shown that advanced algorithms such as neural networks and transformers are well-suited for the purpose. However, for shorter text documents, such as those obtained from social media or news articles, training of deep learning models becomes challenging since they require a large amount of input. Furthermore, genre classification of text summaries, such as headlines of news, is an important direction which has not been explored at large. In this work, the effect of Extractive and Abstractive Summarization on classification for genre of text documents was evaluated. Gensim summarizer was used to obtain extractive summaries and the Pegasus summarizer to obtain abstractive summaries. For classification, two classes of genres, Fiction and Non-fiction, were considered while the gold standard Brown Corpus was used for experimentation. The features used for genre classification were frequencies of various Part-of-Speech (PoS) tags derived from five Penn TreeBank annotated tags. Logistic Regression (LR) and Support Vector Machines (SVM) were used for classification purposes. The results of classification were better for summaries obtained using the extractive technique, indicating that the features of extractive summaries remain in agreement with the documents from which the summary is constructed as compared to abstractive summaries. Further, the SVM classifier performed better than the LR classifier. For exhaustive coverage of the research goal, further experimentation with the number of words of the output summaries of the extractive technique was performed to arrive at a threshold value of the length of summaries. The value indicated that summaries as short as 80 words can be successfully classified using this method.
BACKGROUND AND OBJECTIVES: Previous research identified cognitive reappraisal as an adaptive emotion regulation strategy. However, theories on emotion regulation flexibility suggest that reappraisal effectiveness (RE) may depend on an individual's familiarity with stressors. In this study, we expect high reappraisal inventiveness (RI), i.e., the generation of many and categorically different reappraisals, to increase RE for individuals with low situational familiarity. Individuals with high situational familiarity, however, would be more effective with low RI. DESIGN: A total of 148 participants completed the Script-based Reappraisal Task, in which they were presented with fear- and anger-eliciting scripts. Depending on trial type, participants were instructed to reappraise (reappraisal-trial) or react naturally (control-trial) to the scripts. After each trial, participants indicated affective states and reappraisals. We assessed RI and calculated RE-scores as difference between affect ratings in reappraisal- and control-trials for valence and arousal. Finally, participants rated the familiarity with each situation. RESULTS: The results indicated a significant moderating effect of situational familiarity on the relationship between RI and RE-valence (not RE-arousal). The moderation was mainly driven by a detrimental effect of RI for individuals with high situational familiarity. CONCLUSIONS: Our results hint at the importance of individual experience with emotional content in the research of cognitive reappraisal.
The article is devoted to the study of the texts of the museum exhibition, which play the mediator role between visitors of a museum and its artifacts. The purpose of the study is to consider the text of the museum exhibition as an effective tool for intercultural communication. The review of foreign and domestic works shows that multilingual museum exposition helps to develop a dialogue and understanding between different cultures. The author studied the museum labels of the State Hermitage Museum, their grammatical and stylistic peculiarities, which should be taken into account while translating museum texts from Russian into English. It is determined that the grammatical peculiarities of the museum texts are verbless constructions, incomplete sentences, where linguistic norms are violated. The style of the texts of museum expositions is, as a rule, publicistic, some texts also have features of scientific, spoken and literary styles. While translating, the translator should aim to save the meaning and functional style.
This corpus-based study aimed to investigate the presence of context-dependent linguistic errors in a corpus of clinical reports. The data were taken from a corpus comprising more than 2 million words and made up of clinical reports from emergency medicine, intensive care unit, general surgery, and psychiatry. Quantitative and qualitative analyses were carried out. A language model based on n-grams was developed for the detection of errors, parameters for the selection of cases were defined, and a classification tool was implemented. The findings indicated that emergency medicine was the medical specialty with the highest number of context-dependent errors and that the most frequent type of error was omission of written accent. Furthermore, the analysis revealed the presence of errors of competence due to the incorrect application of the linguistic norm of Spanish, phenomena of phonetic similarity, and composition of words; it is also worth noting that performance errors occurred due to rapid typing on the keyboard. This study constituted the first analysis and creation of a typology of context-dependent errors for the medical domain in Spanish. It contributed to the design of a module based on linguistic knowledge that can be used for the development and improvement of automatic correction systems that, in turn, are used for data processing in medicine.
Social touch is crucial for human well-being, as a lack of tactile interactions increases anxiety, loneliness and need for social support. To address the detrimental effects of social isolation, we build on cutting-edge research on social touch and movement sonification to investigate whether social tactile gestures could be perceived through sounds, a sensory channel giving access to remote information. Four experiments investigated participants’ perception of auditory stimuli that were recorded with our ‘audio-touch’ sonification technique, which captures the sounds of touch. In the first experiment, participants correctly categorized sonified skin-on-skin tactile gestures (i.e., stroking, rubbing, tapping, hitting). In the second experiment, the audio-touch sample consisted of the sonification of six socio-emotional intentions conveyed through touch (i.e., anger, attention, fear, joy, love, sympathy). Participants’ valence ratings of the underlying touch were coherent with the intended emotions, while their categorization presented more variability. In two additional experiments investigating the characteristics of the surface involved in the tactile interactions (i.e skin or object), skin proved to be a critical factor in the auditory categorization of both gestural and emotional audio-touch stimuli. This result reveals that human skin-on-skin interactions convey information which sets them apart, perceptually, from object-on-object interactions. Our research thus unveils that social touch can be perceived through sounds, when they are obtained with our specific sonifying methodology, tailored for skin-on-skin interactions. This bears great promise for giving remote access, through the auditory channel, to meaningful social touch interactions.
Graph neural networks (GNNs) have achieved remarkable success in structured prediction, owing to the GNNs’ powerful ability in learning expressive graph representations. However, most of these works learn graph representations based on a static graph constructed by an existing parser, suffering from two drawbacks: (1) the static graph might be error-prone, and the errors introduced in the static graph cannot be corrected and might accumulate in later stages, and (2) the graph construction stage and graph representation learning stage are disjoined, which negatively affects the model’s running speed. In this paper, we propose a joint-learning-based dynamic graph learning framework and apply it to two typical structured prediction tasks: syntactic dependency parsing, which aims to predict a labeled tree, and semantic dependency parsing, which aims to predict a labeled graph, for jointly learning the graph structure and graph representations. Experiments are conducted on four datasets: the Universal Dependencies 2.2, the Chinese Treebank 5.1, the English Penn Treebank 3.0 in 13 languages for syntactic dependency parsing, and the SemEval-2015 Task 18 dataset in three languages for semantic dependency parsing. The experimental results show that our best-performing model achieves a new state-of-the-art performance on most language sets of syntactic dependency and semantic dependency parsing. In addition, our model also has an advantage in running speed over the static graph-based learning model. The outstanding performance demonstrates the effectiveness of the proposed framework in structured prediction.
Mornings are salient times for disrupted affect that may be impacted by prior sleep. The current study extends work linking sleep disruptions with negative affect by examining how nightly changes in sleep duration, timing, and quality relative to a person's average impact morning affect. We further tested whether depression severity moderated the relationship between nightly variations in sleep and morning affect. This is a secondary analysis of participants ages 18-65 years with varying levels of depression (N = 91) who wore an Actiwatch for 3-17 days (n = 73) while reporting morning affect using a visual analogue scale. Multilevel models tested the previous night's sleep duration, timing, or quality as a predictor of morning affect. Sleep measures were group-mean centred to account for nightly variation in participants' sleep. A cross-level interaction between depression severity and nightly sleep was entered. Sleeping longer (b = 0.1; p < 0.001) and later (b = 1.8; p = 0.01) than usual were both associated with better morning mood. There was a significant interaction between nightly actigraphic sleep duration and depression severity on morning affect (b = 0.003; p = 0.003). Participants with higher depression severity reported worse affect upon waking after sleeping less than their usual. In comparison, sleeping less than usual did not affect morning affect ratings for participants with lower depression. A similar interaction was found for sleep quality (b = 0.02; p < 0.001). There was no interaction for midsleep timing. Sleeping less than usual impacted morning affect in individuals with greater depression, potentially suggesting a pathway by which sleep disturbances perpetuate depression.
GUARDIAN-MT (GUarded by Advanced Radar technology-based Diagnostics Applied in palliative and intensive care Nursing – Music Therapy) was the music therapy subproject within a research examining the reliability of non-contact vital parameter measurements using radar, compared to validated measurement methods. The focus of GUARDIAN-MT was on evaluating live-played sounds during a cold-pressor induced pain stimulus using a mixed-method approach. In addition to direct standard measures, a content-analytical procedure was applied to analyze telephone interviews regarding the retrospective experience of the pain event according to psychologically relevant experiential categories. It was found that particularly the arousal ratings, assessed by independent evaluators, had a substantial influence on the perceived valence of the pain event. Arousal-reducing interventions, such as well-coordinated performance of individual live music, can contribute to experiencing the pain not only as milder, but also less prolonged. The analysis of qualitative data can help uncover and deepen the essential qualities and effects of music therapy interventions.
Objective: Insomnia affects 30-45% of the world population, is related to mortality (i.e., auto accidents and job-related accidents), and is related to mood and affect disorders such as anxiety and depression. Better understanding of insomnia via increased research will decrease the burden on insomnia. The neurocognitive model of sleep proposes that conditioned somatic and cognitive hyperarousal develop in response to repeated pairings of sleep-related stimuli with insomnia-related wakefulness. The purpose of this study was to examine the neurocognitive model of sleep using a novel laboratory paradigm, the Sleep Approach Avoidance Task (SAAT). It was hypothesized that individuals who report symptoms of insomnia will display a bias for negative sleep-related images from the SAAT, which is presumably a reflection of cognitive, behavioral and physiological processes associated with hyperarousal. It was also hypothesized that participants who report poor sleep would provide different subjective ratings for negative images (i.e., stronger valence and arousal) than individuals who reported better sleep. Participants and Methods: An initial sample of 66 healthy college-aged participants completed the Insomnia Severity Index (ISI), the Pittsburgh Sleep Quality Index (PSQI) the Dysfunctional Attitudes and Beliefs about Sleep (DBAS) scale and the Epworth Sleepiness Scale (ESS). Participants also completed the SAAT. The SAAT was developed to assess sleep-related bias in adults. The SAAT is a visual, joystick controlled reaction time task that measures implicit bias for positive and negative sleep-related images. At the end of the task the participants are also asked to rate each image along three dimensions included valence, arousal and dominance. Results: There was a positive correlation between the SAAT and the ISI [r(61) =.30, p =.01], indicating that symptoms of insomnia are related to negative approach-related bias for sleep-related images. No other correlations were observed between the SAAT and self-report sleep measures. With regard to rating of images, higher dominance ratings for negative images were correlated with the SAAT [r(62) =.24, p =.03], which indicates that the approach bias for negative images is associated with “being in control.” Multiple linear regression was used to test if ISI scores and dominance ratings for negative images significantly predicted SAAT bias scores. The overall regression was statistically significant [r2 =.13, F(2, 58) = 4.15, p =.02]. ISI scores significantly predicted SAAT scores (ß =.27, p =.04), whereas dominance ratings for negative images did not significantly predict SAAT scores (ß =.20, p =.11). Exploratory correlational analyses were also completed for ratings of images and other sleep self-report measures. Valence ratings for positive sleep-related images were positively correlated with the ESS [r(64) =.36, p =.01], whereas valence ratings for negative sleep-related images were negatively correlated with the ESS [r(64) = -.24, p =.03]. Conclusions: Hypotheses were partially supported with the ISI being the only self-report measure associated with negative bias for sleep-related images. While ratings of dominance are associated with bias for negative sleep-related images, these ratings do not provide unique variance. These findings indicate a cognitive processing bias for sleep-related stimuli among young adult poor sleepers. Limitations, implications for assessment and intervention are discussed.
It is widely known that English teaching practices are largely rooted in colonialism and linguistic imperialism – the belief in the superiority of the Western teaching methods and linguistic norms established in such countries as the UK and the US. The current trend to decolonize English teaching has been gaining momentum for over three decades but has not entirely penetrated into the mainstream teaching of English, with English for Academic Purposes not being an exception. Relatively little focus has been placed in the research onto non-English speaking contexts with the purpose of analysing the current state of EAP provision at universities and gauging the influence of the native-speakerist academic English norms on EAP students’ writing. In the case of the EU context, it may be particularly pertinent to investigate the potential influence that Brexit might have had in terms of rejecting the imposition of the Standard British English norms and associated teaching approaches. This paper is meant to be a reflection with an attempt to stimulate the discussion of EAP teaching practices and academic discourses in the EU higher education in the post-Brexit era. It will consider the issues in the EAP provision in the EU with the example of Portuguese HE and will reflect on the native-speakerist tendencies within the academia and ways to tackle the dominance of the Anglophone norms. This paper hopes to contribute to the argument in favor of the decolonization of EAP teaching practices in non-English speaking contexts, as decolonization can help foster a more equitable and inclusive world.
Moral decision-making is influenced by various factors, including personality and language. In this cross-sectional study, we investigated the Foreign-Language effect (FLe) in early, highly proficient, Catalan-Spanish bilinguals and examined the role of several personality dimensions in their responses to moral dilemmas. We obtained a multilevel data structure with 766 valid trials from 52 Catalan-dominant undergraduate students who read and responded anonymously to a computerized task with 16 standardized moral dilemmas, half in Catalan and half in Spanish. Results of a multilevel multivariate logistic regression analysis showed that consistent with previous research, participants gave more utilitarian responses to impersonal than personal dilemmas. The language of the dilemma had no effect on the response (dichotomous: utilitarian vs. deontological), decision time, or affective ratings, contradicting the hypothesis of shallower emotional processing of the information in the second language. Interestingly, cruelty features of psychopathy were significantly associated with an enhanced proportion of utilitarian decisions irrespective of the language or the nature of the dilemmas. Furthermore, cruelty features interacted with participants' assessment of dilemma aspects like vividness and verisimilitude. Overall, our findings suggest that early bilinguals immersed in a dual-language context using close Romance languages do not show the FLe and that personality traits like cruelty can modulate moral decisions regardless of language or dilemma type.
This paper introduces CORAE, a novel web-based open-source tool for COntinuous Retrospective Affect Evaluation, designed to capture continuous affect data about interpersonal perceptions in dyadic interactions. Grounded in behavioral ecology perspectives of emotion, this approach replaces valence as the relevant rating dimension with approach and withdrawal, reflecting the degree to which behavior is perceived as increasing or decreasing social distance. We conducted a study to experimentally validate the efficacy of our platform with 24 participants. The tool’s effectiveness was tested in the context of dyadic negotiation, revealing insights about how interpersonal dynamics evolve over time. We find that the continuous affect rating method is consistent with individuals’ perception of the overall interaction. This paper contributes to the growing body of research on affective computing and offers a valuable tool for researchers interested in investigating the temporal dynamics of affect and emotion in social interactions.
Dialogue-level dependency parsing has received insufficient attention, especially for Chinese. To this end, we draw on ideas from syntactic dependency and rhetorical structure theory (RST), developing a high-quality human-annotated corpus, which contains 850 dialogues and 199,803 dependencies. Considering that such tasks suffer from high annotation costs, we investigate zero-shot and few-shot scenarios. Based on an existing syntactic treebank, we adopt a signal-based method to transform seen syntactic dependencies into unseen ones between elementary discourse units (EDUs), where the signals are detected by masked language modeling. Besides, we apply single-view and multi-view data selection to access reliable pseudo-labeled instances. Experimental results show the effectiveness of these baselines. Moreover, we discuss several crucial points about our dataset and approach.
Introduction: Effective communication often involves expressing disagreement while maintaining social harmony, which is influenced by cultural and linguistic norms. Native speakers of English typically employ various politeness strategies in their disagreement speech acts. However, Iraqi EFL learners may navigate these strategies differently due to variations in cultural norms and language proficiency. Therefore, the current study aimed to contrastively analyze the way Iraqi EFL learners and native English speakers perform the speech act of disagreement in light of politeness. Methodology: In this regard, a discourse completion test (DCT) was administered to 66 participants, comprising 33 Iraqi EFL students studying English as a foreign language (TEFL) and 33 native English speakers. The DCT was made up of scenarios that mirrored real-life circumstances in order to provoke responses from people who disagreed with them. Brown and Levinson’s (1987) theory of politeness was employed to analyze participants’ utterances. Results: The findings indicated that while expressing disagreement with people of higher, participants in both groups were more concerned with keeping their interlocutors’ positive faces. Furthermore, the study findings indicated that despite differences in the two groups of participants, Iraqi EFL learners utilized positive indirect politeness strategies more frequently than English native speakers. On the other hand, English native speakers applied direct and negative politeness strategies. Conclusion: Generally, the findings indicated that both groups tended to use the most direct type of disagreement as the social distance and power relation decreased.
Atuy Galon is a unique fanfiction that began as a series of short stories on social media, eventually gaining prominence and being published in a comic strip format. What sets this work apart is its distinctive use of slang language, an ever-evolving linguistic phenomenon that resonates particularly with younger audiences. The central figures in the narrative, Atuy, Anton, and Sahrul, not only use slang but each has developed their own distinct slang lexicon, reflecting varied facets of cultural identity. As language remains paramount in shaping one's cultural self-awareness, this research meticulously examines the linguistic choices of these characters. Employing a descriptive qualitative approach alongside a rigorous discourse analysis methodology, the study aims to decode the linguistic intricacies within Atuy Galon and their broader implications for cultural identity formation. The significance of this exploration extends beyond literary analysis; it offers a window into the dynamic interplay between language, culture, and identity among today's youth. Additionally, it underscores the transformative power of technology in reshaping linguistic norms and practices.
In real time, the size of information resources in natural language is growing rapidly. The processing of these resources urgently requires the presence of linguistic databases and knowledge. Processing of information resources in natural language requires the presence of text corpora and thesauri. To create and process them, markup languages and ontological models of subject areas are required. Insufficient use of linguistic and ontological knowledge used in information retrieval and automatic text processing applications leads to various problems: irrelevant search, poor-quality categorization and referencing of documents. The existing markup languages mainly contain concepts of Romano-Germanic and Slavic language groups. These puzzles are considered burning in the field of computational linguistics. For these purposes, it is proposed to create a metalanguage and an ontological model of the grammar of the Kazakh language. Keywords: ontological model, Kazakh language, natural language, linguistic, Kazakh grammar, semantic. Қазіргі уақытта табиғи тілдегі ақпараттық ресурстардың көлемі тез өсуде. Бұл ресурстарды өңдеу жедел түрде лингвистикалық мәліметтер базасы мен білімнің болуын талап етеді. Ақпараттық ресурстарды табиғи тілде өңдеу мәтіндік корпус пен тезауристан құралады. Ақпаратты іздеу және мәтінді автоматты өңдеу қолданбаларында қолданылатын лингвистикалық және онтологиялық білімдерді жеткіліксіз пайдалану әртүрлі мәселелерге әкеледі. Қолданыстағы белгілеу тілдерінде негізінен роман-герман және славян тілдері топтарының ұғымдары бар. Оларды жасау үшін белгілеу тілдері және пәндік облыстардың онтологиялық үлгілері қажет болады. Бұл есептеуіш лингвистика саласында кеңінен таралған деп саналады. Сонымен қатар, мәтінді автоматты өңдеудің заманауи әдістеріне тіл мен әлем туралы қосымша білім көлемін енгізу күрделі мәселе болып табылады. Осы мақсатта қазақ тілі грамматикасының метатілі мен онтологиялық моделін жасау ұсынылады. Түйiн сөздер: онтологиялық модель, қазақ тілі, табиғи тіл, лингвистикалық, қазақ грамматикасы, семантикалық. В настоящее время объем информационных ресурсов на естественном языке стремительно растет. Развитие этих ресурсов требует наличия актуальной лингвистической базы данных и знаний. Обработка информационных ресурсов на естественном языке состоит из корпуса текстов и тезауруса. Недостаточное использование Абай атындағы ҚазҰПУ-нің ХАБАРШЫСЫ, «Физика-математика ғылымдары» сериясы, №3(79), 2022 лингвистических и онтологических знаний, используемых в приложениях для поиска информации и обработки текстов, приводит к различным проблемам. Существующие языки нотации в основном содержат понятия романо- германской и славянской языковых групп. Для их создания потребуются языки разметки и онтологические модели предметных областей. В то же время в современные методы автоматической обработки текстов трудно внедрить дополнительные знания о языке и мире. Для этого предлагается создать метамодель и онтологическую модель грамматики казахского языка. Ключевые слова: онтологическая модель, казахский язык, естественный язык, лингвистика, казахская грамматика, семантика.
Abstract This study aims to empirically test whether identifying as a supporter of either New South Wales (NSW) or Queensland (QLD) rugby league teams influences the extent that their respective team colors blue and maroon are associated with positively and negatively valenced words. We used a valence categorization experiment and affective rating task (valence and preference) to investigate if team affiliation and shared ingroup experience influenced affective associations with team colors. NSW supporters were faster and more accurate when categorizing positive words presented in blue than maroon font and negative words in maroon than blue font. While QLD supporters did not significantly differ when categorizing words in either blue or maroon, they rated blue and maroon equally positively in contrast to the NSW supporters. Results from this study give us greater insights into how color‐valence associations can be formed through subcultural ingroup affiliations.
Abstract Stress, anxiety, and depressive symptoms can be reduced by listening to music, but the underlying mechanisms remain unclear. To address this gap, we measured brain connectivity while participants listened to songs of different genres: ambient, pop, and metal. Additionally, affective ratings were obtained while participants ( n = 30) listened to the six different songs, and subjective ratings of state anxiety were solicited at the terminus of each song. Electroencephalography (EEG) connectivity combining weighted Phase Lag Index and graph theory was utilised to document brain activity during listening. Repeated-measures ANOVA indicated that listening to more pleasant and less arousing songs was associated with lower self-reported state anxiety levels than songs rated unpleasant and highly arousing. Of interest, EEG alpha connectivity differed across two ambient songs, particularly in the frontal lobes, despite being from the same genre and rated as highly pleasant and low in arousal. We also observed a sex effect on EEG results, where female participants ( n = 18) displayed stronger connectivity than male participants ( n = 12). Combined, these results suggest that ambient songs reduce state anxiety but have divergent brain responses, possibly reflecting the complex nature of music listening, including sensory processing, emotion and cognition.
The paper traces the dynamics of the interpretation of the grammatical nature of the vocative in Ukrainian grammars from the 16th century until the present. The subject of the analysis is the content and presentation of this category in two sections of Ukrainian grammar books: (1) morphological, which clarifies the status of the vocative in the inflectional paradigm of the noun, and (2) syntactic, in which the means of expressing address are characterized. Based on the findings of the research, various trends in the description of the vocative in different historical periods have been identified, in particular: (1) until the beginning of the 20th, it was unequivocally qualified as an equal member of the inflectional paradigm of the noun, equal to other cases; (2) from the beginning of the 20th century to 1933 was a period of competition between two theories (the vocative is a case the same as others or the vocative is not a true case, but a “special” form in the inflectional paradigm of the noun); (3) the canonization of the “fake case” status theory; (4) from 1991 to the present there has been an unanimity of authors in qualifying the vocative as a case. Comparing the stages of fundamental changes in the scientific definition of the vocative in grammars with defining events in the history of Ukraine provides the basis for discussions about the influence of socio–political factors on the representation of linguistic theories and the codification of the linguistic norms.
The Wall Street Journal section of the Penn Treebank has been the de-facto standard for evaluating POS taggers for a long time, and accuracies over 97\% have been reported. However, less is known about out-of-domain tagger performance, especially with fine-grained label sets. Using data from Elder Scrolls Fandom, a wiki about the \textit{Elder Scrolls} video game universe, we create a modest dataset for qualitatively evaluating the cross-domain performance of two POS taggers: the Stanford tagger (Toutanova et al. 2003) and Bilty (Plank et al. 2016), both trained on WSJ. Our analyses show that performance on tokens seen during training is almost as good as in-domain performance, but accuracy on unknown tokens decreases from 90.37% to 78.37% (Stanford) and 87.84\% to 80.41\% (Bilty) across domains. Both taggers struggle with proper nouns and inconsistent capitalization.
Numerous studies have been conducted on the interpretation and translation of English terms into other languages. The purpose of this study was to identify the adequate Indonesian equivalent terminology for hotel amenities, services, and facilities applied in English and the strategies utilized by both domestic and international hotel guests in understanding the equivalent terms in their native language. Qualitative research methodology was used. The subjects included 10 domestic guests from a 5-star hotel, 10 domestic guests from a 4-star hotel, 5 international guests from a 3-star hotel, and 2 hotel staff from a 5-star hotel, 3 staff from a 4-star hotel, and 1 staff from a 3-star hotel. The findings demonstrated that some of the English terms commonly used in hotels had Indonesian equivalents, and some did not. The international guests strategies were: 1) searching in an online dictionary or a Google search; 2) asking people they met nearby immediately; and 3) guessing the meaning. Domestic guests’ strategies included: (a) asking other guests or hotel staff for clarification; and (b) guessing the meaning. Future research should overcome the limitations of this study, considering translations and linguistic norms training strategies.
Introduction A study was conducted to investigate if an individual’s trust in law enforcement affects their perception of the emotional facial expressions displayed by police officers. Methods The study invited 77 participants to rate the valence of 360 face images. Images featured individuals without headgear (condition 1), or with a baseball cap (condition 2) or police hat (condition 3) digitally added to the original photograph. The images were balanced across sex, race/ethnicity (Asian, African American, Latine, and Caucasian), and facial expression (Happy, Neutral, and Angry). After rating the facial expressions, respondents completed a survey about their attitudes toward the police. Results The results showed that, on average, valence ratings for “Angry” faces were similar across all experimental conditions. However, a closer examination revealed that faces with police hats were perceived as angrier compared to the control conditions (those with no hat and those with a baseball cap) by individuals who held negative views of the police. Conversely, participants with positive attitudes toward the police perceived faces with police hats as less angry compared to the control condition. This correlation was highly significant for angry faces ( p &lt; 0.01), and stronger in response to male faces compared to female faces but was not significant for neutral or happy faces. Discussion The study emphasizes the substantial role of attitudes in shaping social perception, particularly within the context of law enforcement.
- output-{ciep,treebanks}-full.csv: frequency and entropy for all the categories, using four types of combinations of layers;<br> - plots.R: R script to draw plots from the output files;<br> - readReport-{CIEP+,treebanks}.R: R script to extract frequency and compute entropy from the report files (not included);<br> - ud-wordorder.py: Python script to extract word order pairs from conllu files and write them in report files. Unfortunately, I cannot include the report files, as CIEP+ is protected by copyright; the analysis can be however replicated with respect to the UD Treebanks.
The goal of this contribution is to present The Digital Rosetta Stone, which is a project developed at Leipzig University by the Chair of Digital Humanities and the Egyptological Institute/Egyptian Museum Georg Steindorff in collaboration with the British Museum and the Digital Epigraphy and Archaeology Project at the University of Florida. The aims of the project are to produce a collaborative digital edition of the Rosetta Stone, address standardization and customization issues for the scholarly community, create data that can be used by students to understand the language and content of the document, and produce a high-resolution 3D model of the stone. First, the three versions of the text were transcribed and encoded in XML according to the EpiDoc guidelines. Next, the versions were aligned with the Ugarit iAligner tool that supports the alignment of ancient texts with modern languages, such as English and German. All three texts were then parsed syntactically and morphologically through Treebank annotation. Finally, the project explored new 3D-digitization techniques of the Rosetta Stone in the British Museum in order to enhance traditional archaeological methods and facilitate the study of the artifact. The results of this work were used in different courses in Digital Humanities, Digital Philology, and Egyptology.
Embodied cognition research identifies mechanisms by which our cognitive activity is connected to body experiences. This approach encompasses not only experimental manipulations but also the quantification of variables related to group and individual differences, i.e., participant-related variables. Moreover, stimuli-related characteristics, such as sensorimotor word ratings, can either be used for the selection of experimental materials or can be the main output of a study themselves. This quantitative information about individuals or stimuli can be collected through non-experimental methods, such as questionnaires and cognitive tests. This chapter gives an overview of questionnaires and cognitive tests often used in embodied cognition research. A questionnaire is a list of questions asking participants to provide information on certain aspects, such as their sociodemographic or medical status. A test is a series of tasks which participants perform for further evaluation by researchers, such as tests of mathematical ability, reading speed, or counting direction. Rating studies collect subjective evaluations of various parameters, typically for large sets of items. The present chapter is divided into two main sections: Participant-related variables and stimuli-related characteristics. We present examples from cognitive linguistics, psycholinguistics, psychophysics, as well as from research on numerical cognition, peripersonal space, and attitudes towards social robots.
We present an approach for assessing how multilingual large language models (LLMs) learn syntax in terms of multi-formalism syntactic structures. We aim to recover constituent and dependency structures by casting parsing as sequence labeling. To do so, we select a few LLMs and study them on 13 diverse UD treebanks for dependency parsing and 10 treebanks for constituent parsing. Our results show that: (i) the framework is consistent across encodings, (ii) pre-trained word vectors do not favor constituency representations of syntax over dependencies, (iii) sub-word tokenization is needed to represent syntax, in contrast to character-based models, and (iv) occurrence of a language in the pretraining data is more important than the amount of task data when recovering syntax from the word vectors.
The purpose of the article is to consider the morphological peculiarities of the system of the nouns in New Bulgarian translation of the “Catechismos” written by Theodore the Studite, which is a part of the manuscripts no. 1/154 kept in Odessa National Scientific Library. The subject of the research is the morphological specifics of nouns in Odessa copy of the “Catechismos” dating from the 18th. The morphological peculiarities of nouns is considered in the context of the formation of a linguistic norm, which allowed the combination with different intensity of linguistic means of several language systems functioning at the time (traditional Middle Bulgarian written language, Church Slavonic Eastern recensions, and vernacular language form). The analysis proposed in this paper presents the extensive system of cases, which does not reflect the real vernacular Bulgarian speech in the 18th; the specifics of the functioning of the gramemes of the case paradigm of masculine, feminine and neuter nouns in the singular and plural forms is analyzed. Usage case endings mistakes, which indicates their artificial nature, are considered. The lack of article of nominal parts of speech is noted; the predominance of compound declension form of the adjectives and participles over short forms is revealed; the relatively high frequency of use of active present participles is registered. The results of the study make it possible to outline some probable factors that determine the writer’s preference for using the linguistic tools of the so-called “bookish”, “traditional”, “archaic” writing systems. An another reason which to some extent explains he usage of case inflections in the text of this relatively late stage of the historical development of the Bulgarian language might be the use of East Slavic copies of the Studite’s sermons by the scriber. Еhe comparison of “Catechismos” copies of South and East Slavic origin is necessary for verification of this assumption, in which we see prospects for further research.
An analysis of robots (simulators) in education is provided. Promising directions for their development are highlighted, such as realism, interactivity, adaptation and personalization. The features of using simulators in dentistry are considered. The main disadvantages of existing simulators in dentistry have been identified, namely the lack of a communicative component and imitation of patient behavior. The anthropomorphic dental simulator is based on the Robo-C robot, which is a unique combination of advanced technologies and human facial expressions, which allows it to communicate with people, reproduce movements of different parts of the body and express emotions. As dental components, the following components were created and implemented into the Robo-C control system: a Smart jaw, including cameras and a temperature sensor, and a Smart tooth, including a pressure sensor. The Robo-C control system has been upgraded taking into account the Smart jaw and Smart tooth, which made it possible to connect the dental treatment process with the robot’s servos through its linguistic base. The process of analyzing data obtained from Smart jaw cameras using a neural network is described. A two-stage classification scheme for dental defects has been proposed and its effectiveness has been proven. The linguistic base contains a set of rules with the help of which devices (microphone, speakers, servos, Smart jaw, Smart tooth) interact with each other. An example of compiling a linguistic database rule is given. The linguistic base, Smart-jaw and Smart-tooth are configured for one of four cases: caries treatment, tooth preparation for a crown, tooth extraction, endodontic treatment. Treatment quality control is carried out using a comprehensive assessment of communication interaction with the robot and analysis of Smart-jaw and Smart-tooth data. An example of work in one of the cases is given. The anthropomorphic dental simulator presented in the article allows the use of new technologies in the training of dentists, as well as the simulation of various dental procedures, which will significantly improve the practical preparation of students for working with patients.
In this paper, we propose a method for removing linguistic information from speech for the purpose of isolating paralinguistic indicators of affect. The immediate utility of this method lies in clinical tests of sensitivity to vocal affect that are not confounded by language, which is impaired in a variety of clinical populations. The method is based on simultaneous recordings of speech audio and electroglotto-graphic (EGG) signals. The speech audio signal is used to estimate the average vocal tract filter response and amplitude envelop. The EGG signal supplies a direct correlate of voice source activity that is mostly independent of phonetic articulation. These signals are used to create a third signal designed to capture as much paralinguistic information from the vocal production system as possible-maximizing the retention of bioacoustic cues to affect-while eliminating phonetic cues to verbal meaning. To evaluate the success of this method, we studied the perception of corresponding speech audio and transformed EGG signals in an affect rating experiment with online listeners. The results show a high degree of similarity in the perceived affect of matched signals, indicating that our method is effective.
BACKGROUND: Cognitive behavioral therapy (CBT) is a moderately efficacious treatment for hoarding disorder (HD), with most individuals remaining symptomatic after treatment. The Joining Forces Trial will evaluate whether 10 weeks of in-home decluttering can significantly augment the outcomes of group CBT. METHODS: A randomized controlled trial of in-home decluttering augmentation of group CBT for HD. Adult participants with HD (N = 90) will receive 12 weeks of protocol-based group CBT for HD. After group CBT, participants will be randomized to either 10 weeks of in-home decluttering led by a social services team or a waitlist. The primary endpoint is 10 weeks post-randomization. The primary outcome measures are the self-reported Saving Inventory-Revised and the blind assessor-rated Clutter Image Rating. Participants on the waitlist will cross over to receive the in-home decluttering intervention after the primary endpoint. Data will be analyzed according to intention-to-treat principles. We will also evaluate the cost-effectiveness of this intervention from both healthcare and societal perspectives. DISCUSSION: HD is challenging to treat with conventional psychological treatments. We hypothesize that in-home decluttering sessions carried out by personnel in social services will be an efficacious and cost-effective augmentation strategy of group CBT for HD. Recruitment started in January 2021, and the final participant is expected to reach the primary endpoint in December 2024. TRAIL REGISTRATION: ClinicalTrials.gov NCT04712474. Registered on 15 January 2021.
The principle of DEPENDENCY LENGTH MINIMIZATION, which seeks to keep syntactically related words close in a sentence, is thought to universally shape the structure of human languages for effective communication. However, the extent to which dependency length minimization is applied in human language systems is not yet fully understood. Preverbally, the placement of long-before-short constituents and postverbally, short-before-long constituents are known to minimize overall dependency length of a sentence. In this study, we test the hypothesis that placing only the shortest preverbal constituent next to the main-verb explains word order preferences in Hindi (a SOV language) as opposed to the global minimization of dependency length. We characterize this approach as a least-effort strategy because it is a cost-effective way to shorten all dependencies between the verb and its preverbal dependencies. As such, this approach is consistent with the bounded-rationality perspective according to which decision making is governed by "fast but frugal" heuristics rather than by a search for optimal solutions. Consistent with this idea, our results indicate that actual corpus sentences in the Hindi-Urdu Treebank corpus are better explained by the least effort strategy than by global minimization of dependency lengths. Additionally, for the task of distinguishing corpus sentences from counterfactual variants, we find that the dependency length and constituent length of the constituent closest to the main verb are much better predictors of whether a sentence appeared in the corpus than total dependency length. Overall, our findings suggest that cognitive resource constraints play a crucial role in shaping natural languages.
Nowadays, tree-structured deep learning classifier models have been widely used in different applications to ensure effective feature representation and learning. Amongst, dimensional sentiment analysis is the most interactive research field, which intends to identify continuous numerical values in the valence-arousal (VA) space. To achieve this, a tree-structured regional convolutional neural network with long short-term memory (T-CNN-LSTM) model was developed, which predicts the VA ratings of the texts for sentiment analysis. In contrast, the effect of a low prediction rate and difficulty of feature learning in a small number of class samples was not analyzed. Hence, this manuscript proposes an adversarial T-CNN-LSTM (A-T-CNN-LSTM) model for predicting the VA to achieve more fine-grained sentiment analysis. This model develops a semantic-enabled frequency-aware generative adversarial network (SFGAN) to produce more adversarial samples using the generator network and decrease the spectral data loss of the discriminator. It embeds the frequency-aware categorizer (FAC) into the discriminator to determine the input veracity in the spatial and spectral domains. Besides, semantic restricted sampling is employed in SFGAN for synthesizing the image subject to a semantic mask. Further, the created samples are classified by the T-CNN-LSTM for predicting the VA scores of sentences. Finally, the experimental results exhibit that the A-T-CNN-LSTM on stanford sentiment Treebank (SST) and CIFAR-10 databases achieves 90.12% and 91% accuracy than the other tree-structured CNNs.
OBJECTIVE: To test and initially describe a new handheld wireless ultrasound technique (TE Air) for clinical use. METHODS: In this pilot study, the new ultrasound device TE Air from Mindray was used to examine the hepatic and renal vessels of healthy volunteers for first impressions. The probe has a sector transducer with a frequency range of 1.8-4.5 MHz. The B-mode and color-coded doppler sonography (CCDS) scanning methods were used. A high-end device from the same company (Resona 9, Mindray) was used as a reference. The results were evaluated using an image rating scale ranging from 0 to 5, with 0 indicating not assessable and 5 indicating without limitations. RESULTS: Altogether, 61 participants (n = 34 female [55.7%], n = 27 male [44.3%]), age range 18-83 years, mean age 37.9±16.5 years) could be adequately studied using TE AIR and the high-end device. With one exception, the image quality score for TE Air never fell below 3 and had a mean/median scored of 4.97/5.00 for the B-mode, 4.92/5.00 for the color flow (CF) mode, and 4.89/5.00 for the pulse wave (PW) mode of the hepatic vein, 4.90/5.00 for the portal vein, 4.11/4.00 for the hepatic artery, and 4.57/5.00 for the renal segmental artery. A significant difference in the assessment of flow measurement of the hepatic artery and renal segmental arteries was found between TE AIR and the high-end device. CONCLUSIONS: TE Air represents a new dimension in point-of-care ultrasound via wireless handheld devices. Especially, its flow measurement ability offers a relevant advantage over other available handheld models. TE Air provides a formally sufficient image quality in terms of diagnostic significance.
Total-body PET/CT scanners provide increased sensitivity, enabling the adjustment of imaging parameters by reducing injected activity or shortening acquisition time. This study aimed to evaluate the limitations of reduced [18F]FDG activity doses on image quality, lesion detectabil-ity, and quantification of lesion uptake in the Biograph Vision Quadra, as well as to assess the benefits of the recently introduced ultra-high sensitivity mode in a clinical setting. A number of 26 patients who underwent [18F]FDG-PET/CT (3.0 MBq/kg, 5 min. scan time) were included in this analysis. PET raw data was rebinned for shorter frame durations to simulate 5 min. scans with lower activities in high sensitivity (HS) and ultra-high sensitivity (UHS) modes. Image quality, noise, and lesion detectability (n=82) were assessed using a 5-point Likert scale. The co-efficient of variation (CoV), signal-to-noise ratio (SNR), tumor-to-background ratio (TBR), and standardized uptake values (SUV) including SUVmean, SUVmax, and SUVpeak were evaluated. Sub-jective image ratings were generally superior in UHS compared to HS mode. At 0.5 MBq/kg, le-sion detectability decreased to 95% (HS) and 98% (UHS). SNR was comparable at 1.0 MBq/kg in HS (5.7±0.6) and 0.5 MBq/kg in UHS (5.5±0.5). With lower doses, there were negligible reductions in SUVmean and SUVpeak, whereas SUVmax increased steadily. Reducing [18F]FDG activity to 1.0 MBq/kg (HS/UHS) in a total-body PET/CT provides diagnostic image quality without statistical-ly significant changes in uptake parameters. UHS mode improves image quality, noise, and le-sion detectability compared to HS mode.
Word-centred neglect dyslexia is most commonly conceptualised as a deficit caused by attentional biases within spatially coded internal representations of words. However, recent research has suggested that at least some cases of word-centred neglect dyslexia are unrelated to visuospatial neglect and may instead be modulated by self-inhibition and lexical factors. Here, we set out to provide novel insight into potential underlying mechanisms modulating the occurrence of word-centred lateralised reading errors in healthy participants. A sample of 47 healthy readers completed a novel attentional cueing paradigm in which they sequentially identified lateral cues and read presented words under limited exposure conditions. Reading responses were analysed to determine whether word-centred neglect dyslexia could be simulated in healthy readers, to compare the strengths of induced biases, and to identify systematic differences in lexical characteristics between target words and neglect dyslexia reading errors. Healthy participants produced frequent lateralised reading errors in both horizontal and vertical reading stimuli with > 50% of errors classed as neglect dyslexic. Cues appended to word beginnings elicited significantly more reading errors than cues at word ends, illustrating the interaction between existing reading spatial attentional biases and cue-induced biases. Neglect dyslexia reading errors were found to contain significantly more letters per word and had higher concreteness ratings than target words. These findings demonstrate that word-centred neglect dyslexia can be simulated using attentional cues in healthy readers. These results provide important insight into the mechanisms underlying word-centred neglect dyslexia and further fundamental understanding of this syndrome.
Artificial neural networks open up unprecedented machine learning capabilities at the cost of ever growing computational requirements. Sparsifying the parameters, often achieved through weight pruning, has been identified as a powerful technique to compress the number of model parameters and reduce the computational operations of neural networks. Yet, sparse activations, while omnipresent in both biological neural networks and deep learning systems, have not been fully utilized as a compression technique in deep learning. Moreover, the interaction between sparse activations and weight pruning is not fully understood. In this work, we demonstrate that activity sparsity can compose multiplicatively with parameter sparsity in a recurrent neural network model based on the GRU that is designed to be activity sparse. We achieve up to $20\times$ reduction of computation while maintaining perplexities below $60$ on the Penn Treebank language modeling task. This magnitude of reduction has not been achieved previously with solely sparsely connected LSTMs, and the language modeling performance of our model has not been achieved previously with any sparsely activated recurrent neural networks or spiking neural networks. Neuromorphic computing devices are especially good at taking advantage of the dynamic activity sparsity, and our results provide strong evidence that making deep learning models activity sparse and porting them to neuromorphic devices can be a viable strategy that does not compromise on task performance. Our results also drive further convergence of methods from deep learning and neuromorphic computing for efficient machine learning.
Technology-based in-home reading and spelling programs have the potential to compensate for the lack of sufficient instructions provided at schools. However, the recent COVID-19 pandemic showed the immaturity of the existing remote teaching solutions. Consequently, many students did not receive the necessary instructions. This paper presents a model for developing intelligent reading and spelling programs. The proposed approach is based on an optimization model that includes artificial neural networks and linear regression to maximize the educational value of the pedagogical content. This model is personalized, tailored to the learning ability level of each user. Regression models were developed for estimating the lexical difficulty in the literacy tasks of auditory and visual lexical decision, word naming, and spelling. For building these regression models, 55 variables were extracted from French lexical databases that were used with the data from lexical mega-studies. Forward stepwise analysis was conducted to identify the top 10 most important variables for each lexical task. The results showed that the accuracy of the models (based on root mean square error) reached 88.13% for auditory lexical decision, 89.79% for visual lexical decision, 80.53% for spelling, and 83.86% for word naming. The analysis of the results showed that word frequency was a key predictor for all the tasks. For spelling, the number of irregular phoneme-graphemes was an important predictor. The auditory word recognition depended heavily on the number of phonemes and homophones, while visual word recognition depended on the number of homographs and syllables. Finally, the word length and the consistency of initial grapheme-phonemes were important for predicting the word-naming reaction times.