Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18625 papers
Recent approaches to Word Sense Disambiguation (WSD) have profited from the enhanced contextualized word representations coming from contemporary Large Language Models (LLMs).This advancement is accompanied by a renewed interest in WSD applications in Humanities research, where the lack of suitable, specific WSD-annotated resources is a hurdle in developing ad-hoc WSD systems.Because they can exploit sentential context, LLMs are particularly suited for disambiguation tasks.Still, the application of LLMs is often limited to linear classifiers trained on top of the LLM architecture.In this paper, we follow recent developments in non-parametric learning and show how LLMs can be efficiently fine-tuned to achieve strong few-shot performance on WSD for historical languages (English and Dutch, date range: 1450-1950).We test our hypothesis using (i) a large, general evaluation set taken from large lexical databases, and (ii) a small real-world scenario involving an ad-hoc WSD task.Moreover, this paper marks the release of GysBERT, a LLM for historical Dutch.
We present substructure distribution projection (SUBDP), a technique that projects a distribution over structures in one domain to another, by projecting substructure distributions separately. Models for the target domain can then be trained, using the projected distributions as soft silver labels. We evaluate SUBDP on zeroshot cross-lingual dependency parsing, taking dependency arcs as substructures: we project the predicted dependency arc distributions in the source language(s) to target language(s), and train a target language parser on the resulting distributions. Given an English treebank as the only source of human supervision, SUBDP achieves better unlabeled attachment score than all prior work on the Universal Dependencies v2.2 (Nivre et al., 2020) test set across eight diverse target languages, as well as the best labeled attachment score on six languages. In addition, SUBDP improves zeroshot cross-lingual dependency parsing with very few (e.g., 50) supervised bitext pairs, across a broader range of target languages.
Purpose: To identify whether exposing medical students to a multimodal curriculum of complementary and alternative medicine (CAM) practices improves their understanding of CAM clinical applications. Background: A significant portion of the U.S. population uses CAM: 34% of adults and 12% of children. Integrative medicine combines the best of conventional and CAM practices. Despite the increased clinical acceptance of CAM, medical education has been lagging, leaving gaps in learners' knowledge. It is important for medical education to keep pace with these developments by educating students and expanding the view of interprofessional care. Methods: A total of 101 first-year medical students at the University of Connecticut participated in a multimodal CAM curriculum. This included (1) an hour lecture, (2) an online research assignment for a continuity patient, and (3) 2 of 4 modules: acupuncture, hypnotherapy, Reiki, or pet therapy. Pre- and post-tests were administered 1 week apart to assess familiarity with CAM practices and the perceived safety and efficacy of each modality. The familiarity was rated on a scale of 0 (not familiar) to 10 (very familiar). Paired Student's t-tests assessed changes from pre- to post-tests at significant levels (p < 0.01). Results: Overall, the mean percentage of students who were able to identify 1 of the top 8 CAM modalities increased from 38% to 49%. The average familiarity rating of CAM significantly increased from 4.7 pretest to 6.6 post-test (p < 0.01). The top 8 CAM modalities, as selected by students, included acupuncture, meditation, yoga, massage, Reiki, chiropractic, hypnosis, and pet therapy. Overall, the familiarity ratings increased for both safety and effectiveness with intermodule variability from pre- and post-test (p < 0.01). Larger increases in effectiveness familiarity were found than of safety familiarity (p < 0.01). Conclusions: This multimodal curriculum significantly improved medical students' familiarity with CAM modalities and the perceived safety and effectiveness of the modalities.
Automatically inferring drivers’ emotions during driver-pedestrian interactions to improve road safety remains a challenge for designing in-vehicle, empathic interfaces. To that end, we carried out a lab-based study using a combination of camera and physiological sensors. We collected participants’ (N=21) real-time, affective (emotion self-reports, heart rate, pupil diameter, skin conductance, and facial temperatures) responses towards non-verbal, pedestrian crossing videos from the Joint Attention for Autonomous Driving (JAAD) dataset. Our findings reveal that positive, non-verbal, pedestrian crossing actions in the videos elicit higher valence ratings from participants, while non-positive actions elicit higher arousal. Different pedestrian crossing actions in the videos also have a significant influence on participants’ physiological signals (heart rate, pupil diameter, skin conductance) and facial temperatures. Our findings provide a first step toward enabling in-car empathic interfaces that draw on behavioural and physiological sensing to in situ infer driver emotions during non-verbal pedestrian interactions.
Objective: Adverse childhood experiences, ranging from childhood trauma to neglect or mistreatment, show associations with alcohol dependence in adulthood. Alcohol researchers have not yet clearly demonstrated the potential impact of childhood maltreatment on everyday drinking in alcohol consumers who do not have an alcohol use disorder (AUD). This study examined whether a history of childhood neglect results in differential ratings of stress, affect, and desire to drink, during typical alcohol consumption in moderate to heavy drinkers without an AUD. Methods: The parent study from which these analyses were generated recruited overall healthy, albeit moderate to heavy alcohol users who fell above National Institute on Alcohol Abuse and Alcoholism (NIAAA) classifications for low-risk drinking. Childhood Trauma Questionnaire (CTQ) responses were collected, and real-time surveys were collected in participants' natural environments approximately every three hours between 9 a.m. and 9 p.m. using iPhones equipped with a study-created application probing stress, affect, and desire for alcohol, while participants followed their typical drinking routine (3 days) and during a brief period of imposed abstinence (3 days). Results: Thirty-six participants averaging 41 years of age and consuming an average of 17 (±5) drinks per week participated in this study. CTQ responses showed low prevalence of childhood abuse, but moderate to high emotional (M = 17.39, SD = 6.77) and physical neglect (M = 11.11, SD = 3.31) scores. Multilevel modeling revealed significantly higher stress and lower affect ratings among participants reporting higher physical neglect. Alcohol consumption was significantly associated with decreased stress, and increased affect and desire for alcohol. A significant interaction showed that as childhood physical neglect ratings increased, the benefits observed following drinking (of decreased stress, and increased affect and desire for alcohol) were reduced. Conclusions: These results suggest that moderate to heavy healthy daily drinkers with histories of greater childhood physical neglect experience poorer mood and higher stress on a daily basis, with smaller improvement experienced from drinking alcohol. Among moderate to heavy daily drinkers without an AUD, those with greater childhood physical neglect experience poorer mood and higher stress on a daily basis, and have smaller improvements in stress, affect and desire while/following drinking alcohol than those with less childhood physical neglect.
Abstract Introduction Decreased rapid eye movement (REM) frontal theta spectral activity during sleep has been identified following intense negative experiences in both naturalistic and experimental settings. It is unknown, however, whether this spectral marker a) also appears following lower intensity naturalistic negative affective experiences, and b) whether it is responsive to arousal, to negative valence, or to both. Methods In a community sample of adults (N = 50, age range 21-49 [M = 30.30, SD = 7.85]) with no mental health or sleep disorders except bruxism, we separately measured naturalistic experiences of negative valence and arousal 40-60 minutes prior to bedtime, then assessed at-home sleep using a single night of ambulatory polysomnography with 8 electroencephalography channels. REM frontal theta was calculated as the mean relative spectral power density between 4.9-8 Hz at F3 and F4 during REM. Affective experience was assessed with a 0-100 slider rating filled out on participants’ smartphones. Negative valence was assessed by asking, “How negative are you feeling right now?” Arousal was assessed using a composite of somatic and cognitive arousal that reverse scored and averaged two ratings, “How calm is your body?” and “How quiet is your mind?” Analyses were conducted with Bayesian general linear models using flat priors. Results Contrary to our hypothesis, pre-sleep arousal was linearly associated with increased REM frontal theta (ß = 2.06 [0.95, 3.20], pMAP &lt; 0.001). A quadratic association fit the data better than the linear model (Bayes Factor = 250.47), such that, for low to moderate pre-sleep arousal, REM frontal theta power increased as arousal ratings increased, but for moderate to high pre-sleep arousal, REM frontal theta power decreased as arousal ratings increased (ß = -2.04 [-3.18, -0.90], pMAP &lt; 0.001). Both associations remained significant even adjusting for negative valence. Contrary to expectations, pre-sleep negative valence was not associated with REM frontal theta. Conclusion Separately measuring affect dimensions revealed that REM frontal theta power was selectively responsive to naturalistic pre-sleep arousal. At higher arousal ratings, our results matched prior findings. However, at lower arousal ratings, the association was reversed, suggesting that spectral markers may vary across affective intensities and contexts. Support (If Any) National Science Foundation Graduate FellowshipR56, National Institute of Dental and Craniofacial Research, “Emotion dysregulation and sleep-time masticatory muscle activity in sleep bruxism,” DE025321
Abstract Nouns and verbs are known to differ in the types of grammatical information they encode. What is less well known is the relationship between verbal and nominal coding within and across languages. The equi-complexity hypothesis holds that all languages are equally complex overall, which entails trade-offs between coding in different domains. From a diachronic point of view, this hypothesis implies that the loss and gain of coding in different domains can be expected to balance each other out. In this study, we test to what extent such inverse coevolution can be observed in a sample of 244 languages, using data from a comprehensive cross-linguistic database (Grambank) and applying computational phylogenetic modelling to control for genealogical relatedness. We find evidence for coevolutionary relationships between specific features within nominal and verbal domains on a global scale, but not for overall degrees of grammatical coding between languages. Instead, these amounts of nominal and verbal coding are positively correlated in Sino-Tibetan languages and inversely correlated in Indo-European languages. Our findings indicate that accretion and loss of grammatical information in nominal words and verbs are lineage-specific.
Probing has become an important tool for analyzing representations in Natural Language Processing (NLP). For graphical NLP tasks such as dependency parsing, linear probes are currently limited to extracting undirected or unlabeled parse trees which do not capture the full task. This work introduces DepProbe, a linear probe which can extract labeled and directed dependency parse trees from embeddings while using fewer parameters and compute than prior methods. Leveraging its full task coverage and lightweight parametrization, we investigate its predictive power for selecting the best transfer language for training a full biaffine attention parser. Across 13 languages, our proposed method identifies the best source treebank 94% of the time, outperforming competitive baselines and prior work. Finally, we analyze the informativeness of task-specific subspaces in contextual embeddings as well as which benefits a full parser’s non-linear parametrization provides.
We propose a morphology-based method for low-resource (LR) dependency parsing. We train a morphological inflector for target LR languages, and apply it to related rich-resource (RR) treebanks to create cross-lingual (xinflected) treebanks that resemble the target LR language. We use such inflected treebanks to train parsers in zero-(training on x-inflected treebanks) and few-shot (training on x-inflected and target language treebanks) setups. The results show that the method sometimes improves the baselines, but not consistently.
As part of its ambitious Belt and Road initiative, the Chinese government, cognizant of the fact that establishing strong people-to-people bonds locally, regionally, and internationally will be key to the initiative’s success and that languages will play a pivotal role in facilitating this, has called for the diversification of language programmes beyond English at the tertiary and pre-tertiary levels. The aim is to boost the plurilingual and pluricultural competence (PPC) of Chinese citizens, who are often already multilingual because of their knowledge of Chinese dialects and minority languages. This article reports on a study involving 248 university students from over 20 provinces in China that explored the interactions between their reported PPC, translingual dispositions, interculturalism, orientations toward cultural differences, and identity-related views. The results revealed that the participants’ translingual dispositions entailed engaging in language negotiation practices to a greater extent than in resisting linguistic norms or exploring language. Moreover, the more multilingual the participants were, the more fluidly they reported moving between languages and the greater was their PPC, which was also positively and statistically significantly predicted by their age, ethnic background, acceptance of cultural differences, and views about their personal identity.
This paper addressed the problem of structured sentiment analysis using a bi-affine semantic dependency parser, large pre-trained language models, and publicly available translation models. For the monolingual setup, we considered: (i) training on a single treebank, and (ii) relaxing the setup by training on treebanks coming from different languages that can be adequately processed by cross-lingual language models. For the zero-shot setup and a given target treebank, we relied on: (i) a word-level translation of available treebanks in other languages to get noisy, unlikely-grammatical, but annotated data (we release as much of it as licenses allow), and (ii) merging those translated treebanks to obtain training data. In the post-evaluation phase, we also trained cross-lingual models that simply merged all the English treebanks and did not use word-level translations, and yet obtained better results. According to the official results, we ranked 8th and 9th in the monolingual and cross-lingual setups.
The present study investigates the relations between L2-English proficiency and L1-Turkish lexical property evaluations. We asked whether L2 proficiency affects lexical properties, including imageability and concreteness ratings of 600 Turkish words selected from the Word Frequency Dictionary of Written Turkish. Seventy-two participants (L1-Turkish - L2-English) provided ratings of concreteness and imageability for 600 words on a 7-point scale. In order to assess their L2 proficiency, we administered Peabody Picture Vocabulary Test-IV (PPVT-IV). We divided categories into two subcategories as high and low for the frequency, concreteness, imageability, and age of acquisition (AoA). The relationship between these subcategories and imageability-concreteness was examined by mixed effects linear regression analyses. We found that L2 proficiency and imageability ratings were positively correlated and specifically, this positive association was evident for low-frequency words and later acquired words. Results are in line with the interaction of bilingual representation under the dual-coding theory which suggests that bilinguals develop an interconnected imaginal representation for two languages as opposed to separate verbal representations. As L2 proficiency increased, the imageability also increased. These findings have implications for literature investigating the relationship between L2 proficiency and linguistic outcomes. Additionally, findings point to the importance of considering the L2 proficiency of participants when lexical tasks that involve cue words or word lists are used.
Abstract This paper investigates linguistic complexity across natural languages from a corpus-based perspective and relies on the assumptions of linguistic profiling as a methodological framework. We focus in particular on the domain of syntactic complexity and analyze the distribution of a set of features taken as proxies of complexity phenomena at sentence level, which were extracted from 63 treebanks annotated according to the Universal Dependencies formalism. This dataset guarantees that the features considered are modeling the same linguistic phenomena in different treebanks, allowing reliable comparison among languages. We show that our approach is able to identify tendencies of structural proximity between languages not necessarily in line with typologically-supported classification, thus shedding light on new corpus-based findings.
Downy mildew is a major disease of grapevine. Conventional methods for assessing crop diseases are time-consuming and require trained personnel. This work aimed to develop and validate a new method to automatically estimate the severity of downy mildew in grapevine leaves using fuzzy logic and computer vision techniques. Leaf discs of two grapevine varieties were inoculated with Plasmopara viticola and subsequently, RGB images were acquired under indoor conditions. Computer vision techniques were applied for leaf disc location in Petri dishes, image pre-processing and segmentation of pre-processed disc images to separate the pixels representing downy mildew sporulation from the rest of the leaf. Fuzzy logic was applied to improve the segmentation of disc images, rating pixels with a degree of infection according to the intensity of sporulation. To validate the new method, the downy mildew severity was visually evaluated by eleven experts and averaged score was used as the reference value. A coefficient of determination (R2) of 0.87 and a root mean squared error (RMSE) of 7.61 % was observed between the downy mildew severity obtained by the new method and the visual assessment values. Classification of the severity of the infection into three levels was also attempted, achieving an accuracy of 86 % and an F1 score of 0.78. These results indicate that computer vision and fuzzy logic can be used to automatically estimate the severity of downy mildew in grapevine leaves. A new method has been developed and validated to assess the severity of downy mildew in grapevine. The new method can be adapted to assess the severity of other diseases and crops in agriculture.
Classical views suggest that experienced affect is related to a specific bodily response, whereas recent perspectives challenge this view postulating that similar affective experiences rather evoke different physiological responses. To further advance this debate in the field, we used representational similarity analysis to investigate the correspondence between subjective affect (arousal and valence ratings) and physiological reactions (skin conductance response [SCR], startle blink response, heart rate, and corrugator activity) across various emotion induction contexts (picture viewing task, sound listening task, and imagery task). Significant similarities were exclusively observed between SCR and arousal in the picture viewing task. However, none of the other physiological measures showed a significant relation with valence and arousal ratings in any of the tasks. These findings are discussed within the framework of the Populations hypothesis, suggesting that physiological responses do not depend on the experienced affect but are directly associated with the context in which they are evoked.
Exploration of the physiological signals associated with subjective emotional dynamics has practical significance. Previous studies have reported that the dynamics of subjective emotional valence and arousal can be assessed using facial electromyography (EMG) and electrodermal activity (EDA), respectively. However, it remains unknown whether other methods can assess emotion dynamics. To investigate this, EMG of the trapezius muscle and fingertip temperature were tested. These measures, as well as facial EMG of the corrugator supercilii and zygomatic major muscles, EDA (skin conductance level) of the palm, and continuous ratings of subjective emotional valence and arousal, were recorded while participants (n = 30) viewed emotional film clips. Intra-individual subjective–physiological associations were assessed using correlation analysis and linear and polynomial regression models. Valence ratings were linearly associated with corrugator and zygomatic EMG; however, trapezius EMG was not related, linearly or curvilinearly. Arousal ratings were linearly associated with EDA and fingertip temperature but were not linearly or curvilinearly related with trapezius EMG. These data suggest that fingertip temperature can be used to assess the dynamics of subjective emotional arousal.
BACKGROUND: The understanding of the cerebral neurobiology of anorexia nervosa (AN) with respect to state- versus trait-related abnormalities is limited. There is evidence of restitution of structural brain alterations with clinical remission. However, with regard to functional brain abnormalities, this issue has not yet been clarified. METHODS: We compared women with AN (n = 31), well-recovered female participants (REC) (n = 18) and non-patients (NP) (n = 27) cross-sectionally. Functional magnetic resonance imaging was performed to compare neural responses to food versus non-food images. Additionally, affective ratings were assessed. RESULTS: Functional responses and affective ratings did not differ between REC and NP, even when applying lenient thresholds for the comparison of neural responses. Comparing REC and AN, the latter showed lower valence and higher arousal ratings for food stimuli, and neural responses differed with lenient thresholds in an occipital region. CONCLUSIONS: The data are in line with some previous findings and suggest restitution of cerebral function with clinical recovery. Furthermore, affective ratings did not differ from NP. These results need to be verified in intra-individual longitudinal studies.
Modern Irish is a minority language lacking sufficient computational resources for the task of accurate automatic syntactic parsing of usergenerated content such as tweets. Although language technology for the Irish language has been developing in recent years, these tools tend to perform poorly on user-generated content. As with other languages, the linguistic style observed in Irish tweets differs, in terms of orthography, lexicon, and syntax, from that of standard texts more commonly used for the development of language models and parsers. We release the first Universal Dependencies treebank of Irish tweets, facilitating natural language processing of user-generated content in Irish. In this paper, we explore the differences between Irish tweets and standard Irish text, and the challenges associated with dependency parsing of Irish tweets. We describe our bootstrapping method of treebank development and report on preliminary parsing experiments.
Reviewed by: Postclassical Greek: Contemporary Approaches to Philology and Linguistics ed. by Dariya Rafiyenko and Ilja A. Seržant William A. Ross dariya rafiyenko and ilja a. seržant (eds.), Postclassical Greek: Contemporary Approaches to Philology and Linguistics (Trends in Linguistics: Studies and Monographs 335; Berlin: de Gruyter, 2020). Pp. viii + 339. $114.99. This volume brings together a slate of scholars to examine linguistic aspects of postclassical Greek, including the biblical corpus, in an interdisciplinary context. It begins with an essay by the editors Dariya Rafiyenko and Ilja A. Seržant that offers an overview of postclassical Greek itself, which they define as "the entire set of spoken and written varieties of the period from 323 BC up to 1453 AD" (p. 1). Their essay briefly surveys issues of periodization and grammar, with an emphasis on the importance of variation in linguistic description, for which reason Rafiyenko and Seržant join others—rightly so in my opinion—in questioning the separation of linguistics from philology that has prevailed throughout much of the last century. They advocate instead what they call the "rephilologization of historical linguistics" (p. 12). The volume is divided into two sections. The first is "Grammatical Categories," which begins with "Purpose and Result Clauses: ἵνα-hína and ὥστε-hṓste in the Greek Documentary Papyri of the Roman Period," by Giuseppina di Bartolo, who surveys the disappearance of certain semantic distinctions in different moods and a reduction of conjunctions to introduce purpose and result clauses. In "Syntactic Factors in the Greek Genitive-Dative Syncretism: The Contribution of New Testament Greek," Chiara Gianollo considers postnominal genitives in the NT, which allowed for the propagation of an external possession construction. Then, in "Future Periphrases in John Malalas," Daniel Kölligan looks at the earliest Byzantine world chronicler, in whose work certain novel, but not idiosyncratic, expressions of future reference appear alongside older patterns. The next essay is "Combining Linguistics, Paleography and Papyrology: The Use of eis, pros and epí in Greek Papyri," by Joanne Stolk, who contrasts these prepositions with the bare dative in expressions of animate goal of motion and transfer verbs, providing new interpretations for uses otherwise considered exceptional. In "Future Forms in Postclassical Greek: Some Remarks on the Septuagint and the New Testament," Liana Tronci examines diachronic morphology to mark future tense, taking external factors such as register variation within the biblical corpus into account. The next essay, by Brian D. Joseph, is entitled "Greek Infinitive-Retreat versus Grammaticalization: An Assessment"; he argues that the morphosyntactic changes in infinitive use represent degrammaticalization and thus that grammatical change is not always unidirectional. The last essay in the first section of this volume is "Postclassical Greek and Treebanks for a Diachronic Analysis," by Nikolaos Lavidas and Dag Trygve Truslew Haug. They employ diachronic analysis to examine syntactic phenomena such as backward control and thus build a linguistic profile of earlier texts. The second section in this volume includes topics related to sociolinguistic aspects and variation in postclassical Greek. The first essay, by Marina Benedetti, is "The Perfect [End Page 538] Paradigm in Theodosius' Κανόνες: Diathetically Indifferent and Diathetically Non-Indifferent Forms," in which she examines the discussion of the perfect in a fourth-century c.e. grammatical treatise. Next, in "Forms of the Directive Speech Act: Evidence from Early Ptolemaic Papyri," Carla Bruno considers strategies available for inducing others to action, including performative utterances and indirect implicatures, among other possibilities. The next essay is one of the longest in the book: "What's in a (personal) Name? Morphology and Identity in Jewish Greek Literature in the Hellenistic and Roman Periods," by Robert Crellin. He challenges the notion that non-nativized morphology of Hebrew personal names implies low-level Greek, arguing instead that sociolinguistic and literary considerations are at work in the decision to adapt personal names or not. The next essay is "Confusion of Mood or Phoneme? The Impact of L1 Phonology on Verb Semantics," by Sonja Dahlgren and Martti Leiwo, who question whether the extensive nonstandard vowel use in Egyptian Greek texts is evidence of poor command of Greek. They suggest instead that such spellings represent underdifferentiated Greek phonemes and transfer elements from non-Greek prosodic systems. The last...
Reviewed by: Por un lenguaje inclusivo: Estudios y reflexiones sobre estrategias no sexistas en la lengua española by Tina Escaja y María Natalia Prunes Emilia Alonso-Marks Por un lenguaje inclusivo: Estudios y reflexiones sobre estrategias no sexistas en la lengua española Academia Norteamericana de la Lengua Española, 2021 Editado por Tina Escaja y María Natalia Prunes This edited volume in the Colección El árbol de las palabras, No. 5, that the Academia de la Lengua Norteamericana (ANLE) puts together, provides an overview of the background of inclusive language and research into language inclusivity and gender neutrality, an approach whose practices have met with resistance from the Royal Spanish Academy of Language (RAE) and other associations concerned with the purity and rigor of linguistic forms. Language inclusion is the practice of mindfulness with regards to gender when writing. Academics, writers, teachers and students from different affiliations and communities advocate for implementing forms that avoid linguistic sexism and integrate those in our society today, where women and other gender identities are demanding to have their presence acknowledged and their voices heard. Nevertheless, the language inclusivity initiative has met with resistance, as "invisibility and exclusion continue to permeate the Spanish language for the sake of a tradition that has been questioned in recent years" (17). This tradition is protected zealously by institutions like the RAE, mentioned above, whose founders and followers continue to be mostly men. For the editors, being mindful and intentional about language inclusivity is not a strictly linguistic topic, "but through it, it [End Page 193] manifests in [the language] an explicit political militance" (25). It is a power struggle between the traditional forces that deny the social dimension of language, and the innovation forces that demand an end not only to gender binarism, reflected on the generic masculine as the unmarked gender, but also to blatant gender discrimination forms that still exist in society today. So, the struggle is not only linguistic in nature, but socio-political as well. Escaja and Prunes (2021) have edited a volume containing 10 academic contributions and 10 contributions of free creation into an anthology of critical research studies and reflections about gender inclusive strategies in the Spanish language. The volume does not offer controversial perspectives around the topic of language inclusivity, but rather takes a position clearly pro the use of inclusive language, focusing on introducing alternatives to language inclusion. The short papers are supposed to add an intellectual counterpoint to the strictly academic essays. Accordingly, the 20 contributions have been distributed into two sections. Part I houses the strictly academic contributions, and Part II contains the short, creative pieces. Part I is not organized by themes, theoretical approaches or analytical frameworks. According to the editors, the idea behind this apparent disorganization is for readers "to enjoy the most pluralistic, pleasant reading possible, where interdisciplinarity gets intermingled and complemented" (31). The first study focuses on how language is the vehicle people use to imagine the world. It is used to build "subjectivities," that is, things pertaining to the first person singular. However, one cannot lose sight that language—the word—is a foundational piece of the individual within society; and society utilizes it, among other things, as an instrument to decide what is acceptable or unacceptable, and only those aspects that are deemed acceptable get normalized. The second study centers around the use of gender-neutral forms in the Spanish Constitutions and how these forms meet criteria having to do with linguistic norm and precision on the one hand, and language economy and what is natural, on the other. In the third article, the contributor takes the reader on a historic journey, where the use of language-inclusive forms gets documented through examples found in newspaper articles and teaching files dating back to the beginning of the twentieth century through today. The contributor concluded that it would be ideal to share this experience in Spanish language, literature and culture courses in the USA, so that students could make connections with English and the social movements that have advocated for language inclusivity. Similarly, the fourth study centers around providing the reader a historic overview based on five...
Traditional recurrent neural networks (RNNs) have a fixed, finite number of memory cells. In theory (assuming bounded range and precision), this limits their formal language recognition power to regular languages, and in practice, RNNs have been shown to be unable to learn many context-free languages (CFLs). In order to expand the class of languages RNNs recognize, prior work has augmented RNNs with a nondeterministic stack data structure, putting them on par with pushdown automata and increasing their language recognition power to CFLs. Nondeterminism is needed for recognizing all CFLs (not just deterministic CFLs), but in this paper, we show that nondeterminism and the neural controller interact to produce two more unexpected abilities. First, the nondeterministic stack RNN can recognize not only CFLs, but also many non-context-free languages. Second, it can recognize languages with much larger alphabet sizes than one might expect given the size of its stack alphabet. Finally, to increase the information capacity in the stack and allow it to solve more complicated tasks with large alphabet sizes, we propose a new version of the nondeterministic stack that simulates stacks of vectors rather than discrete symbols. We demonstrate perplexity improvements with this new model on the Penn Treebank language modeling benchmark.
We introduce a method for unsupervised parsing that relies on bootstrapping classifiers to identify if a node dominates a specific span in a sentence. There are two types of classifiers, an inside classifier that acts on a span, and an outside classifier that acts on everything outside of a given span. Through self-training and co-training with the two classifiers, we show that the interplay between them helps improve the accuracy of both, and as a result, effectively parse. A seed bootstrapping technique prepares the data to train these classifiers. Our analyses further validate that such an approach in conjunction with weak supervision using prior branching knowledge of a known language (left/right-branching) and minimal heuristics injects strong inductive bias into the parser, achieving 63.1 F 1 on the English (PTB) test set. In addition, we show the effectiveness of our architecture by evaluating on treebanks for Chinese (CTB) and Japanese (KTB) and achieve new state-of-the-art results. 1
Ten years ago, when the META-NET Network of Excellence conducted a study on language technology support for European languages, Latvian was included in the category of languages with little or no support. During the last decade, notable progress has been made in the development of language resources and tools for Latvian, particularly regarding the creation of advanced datasets like speech corpora and treebanks, state-of-the-art neural language models, machine translation systems, speech technology, and technologies for natural language understanding and human-computer interaction. This paper provides an overview of the most recent activities in the language technology field in Latvia: national and international initiatives, key language resources and tools, key projects and initiatives. We summarize both the recent activities and the most significant achievements after the publication of the META-NET White Paper on Latvian.
We describe Turkish Discourse Bank 1.2, the latest version of a discourse corpus annotated for explicitly or implicitly conveyed discourse relations, their constitutive units, and senses in the Penn Discourse Treebank style. We present an evaluation of the recently added tokens and examine three commonly occurring dependency patterns that hold among the constitutive units of a pair of adjacent discourse relations, namely, shared arguments, full embedding and partial containment of a discourse relation. We present three major findings: (a) implicitly conveyed relations occur more often than explicitly conveyed relations in the data; (b) it is much more common for two adjacent implicit discourse relations to share an argument than for two adjacent explicit relations to do so; (c) both full embedding and partial containment of discourse relations are pervasive in the corpus, which can be partly due to subordinator connectives whose preposed subordinate clause tends to be selected together with the matrix clause rather than being selected alone. Finally, we briefly discuss the implications of our findings for Turkish discourse parsing.
Summary This study explores the everyday Otherness experienced by Southeast Asian marriage-migrant women in South Korea. South Korea is increasingly ethnically diverse due to the dramatic rise in international marriages between foreign women and Korean men, most of which are facilitated by marriage brokers. Yet little research has been conducted on marriage-migrants’ experiences of communicating with local Koreans. Drawing on data collected through in-depth interviews with five participants from Cambodia and Vietnam, this study focuses on specific factors that cause conflicts between these women and local Koreans in various social contexts, including the household, workplaces, and wider communities, and how the women respond to such conflicts and manage challenging interactions. The participants’ narratives demonstrate the tensions and conflicts they encounter, which can be divided into three categories: the imposition of Korean ways of living, negative stereotyping, and language use. The women describe being perceived as deviating from Korean society’s cultural and linguistic norms and facing pressure to conform to these norms, which sometimes conflict with their own sense of identity. In addition, they experience marginalization through Othering and negative stereotyping in their interactions with Koreans and struggle to develop a sense of belonging to the host society. The results of this study provide implications for second language programs designed for marriage-migrants, which have the potential to enable marriage-migrants to achieve sustainable development in their second language learning and to support their development of multilingual and multicultural identities.
This paper deals with the evolution of nominal expressions with prenominal possessives in Romance languages. Until now, the prevalent view has been that their historical development follows one of the two diachronic paths: either they stop co-occurring with determiners (French, Spanish) or they start requiring their presence (European Portuguese). One common analysis associates the first case with a transition from a modifier to a determiner semantics of possessives and the second with a retention of a modifier semantics (e.g. Alexiadou 2004). It has also been observed, however, that documented Romance languages each make use of two morphologically distinct possessive paradigms, the so-called short and long forms, and, at least for French, that the short forms co-occur with determiners in historical documents less frequently than the long ones (Butet 2018). We adduce novel quantitative evidence from treebanks of French, Spanish, and Portuguese to the latter observation and show for the first time that the two paradigms followed very different evolutionary trajectories: frequency of determiners goes monotonically up with long possessives and goes up and then down with the short possessives across the three languages. Given these new data, we argue that the only point of evolutionary divergence in Romance possessive DPs is the choice of either the long or the short prenominal possessive in each of the languages, while the semantics of long and short possessives is stable both diachronically and across the three languages we examine. We also argue that the observed diachronic developments are consequences of a pan-Romance rise of a DP grammar. The paper thus contributes to our understanding of the emergence of DP in general and definiteness marking in particular, of the semantico-syntactic architecture of possessive DPs, as well as of the evolutionary divergence and convergence within a language family.
Neural networks (NNs) have been successfully deployed in various applications of artificial intelligence. However, architectural design of NNs is still a challenging problem. This is due to the need to navigate a search space based on a large number of hyperparameters. This forces the search space of possible architectures to grow exponentially. Using a trial-and-error design approach is very time consuming and leads to suboptimal architectures. In addition, approaches, such as neural architecture search based on reinforcement learning and differentiable gradient-based architecture search, often incur huge computational costs or significant memory requirements. To address these challenges, we propose the CURIOUS NN synthesis methodology. It uses a performance predictor to efficiently navigate the architectural search space with an evolutionary search process. The predictor is built using quasi Monte-Carlo sampling, boosted decision tree regression, and an intelligent iterative sampling method. It is designed to be sample efficient. CURIOUS starts from a base architecture and explores the architectural search space to obtain a variant of the base architecture with the highest performance. This search framework is general and covers all important NN architecture types, e.g., feedforward NNs (FFNNs), convolutional NNs (CNNs), recurrent NNs (RNNs), and transformers. We evaluate the performance of CURIOUS on various datasets and base architectures. Through these experiments, we demonstrate significant performance improvements over the baseline architectures. For the MNIST dataset, our CNN architecture achieves an error rate of 0.66%, with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$8.6\times $ </tex-math></inline-formula> fewer parameters compared to the LeNet-5 baseline. For the CIFAR-10 dataset, we use the ResNet architectures and residual networks with Shake-Shake regularization as the baselines. Our synthesized ResNet-18 has a 2.52% accuracy improvement over the original ResNet-18, 1.74% over ResNet-101, and 0.16% over ResNet-1001, while requiring comparable number of parameters and floating-point operations to the original ResNet-18. This result shows that instead of just increasing the number of layers to increase accuracy, an alternative is to use a better NN architecture with a small number of layers. In addition, CURIOUS achieves an error rate of just 2.69% with a variant of the residual architecture with Shake-Shake regularization. We also use the set of optimized hyperparameters found for ResNet-18 on the CIFAR-10 dataset to train and evaluate the model on the ImageNet dataset, and show 3.43% (1.83%) improvement in the top-1 (top-5) error rate compared to the original ResNet-18 model. CURIOUS also obtains the highest accuracy for various other FFNNs that are geared toward edge devices and IoT sensors. In addition, we use CURIOUS to search for deep RNN architectures for the SICK dataset for sentence similarity evaluation. It achieves a mean-squared error of only 0.2060, improving upon the base network performance, without the need to stack multiple long short-term memories. We also use CURIOUS to search for a better NN classifier for the sentiment analysis task on the Stanford sentiment treebank dataset using a pretrained BERT model and again demonstrate improvements in performance.
This article discusses the process of translating Carolina Maria de Jesus’s autobiographical narrative “Favela”, published in Onde estaes felicidade? in 2014, a translation project carried out as an experiment within the SELCS Brazilian Translation Club workshops in October 2021 at University College London. The study identifies points of convergence within the translation, and divergences from expected English cultural and linguistic norms, as well as detailed analyses of several strategies for dealing with problems and concepts such as equivalence, foreignisation, acculturation, and untranslatability. The reflection on translation procedures was guided by the principle of “negative revision”, taking into account the most recent research on the teaching and the practice of literary translation, and paying particular attention to Carolina Maria de Jesus’ autographs, which in themselves dispel several racial, social and cultural stigmas conveyed by the first editions and translations of the author’s work.
Vocal and facial cues typically co-occur in natural settings, and multisensory processing of voice and face relies on their synchronous presentation. Psychological research has examined various facial and vocal cues to attractiveness as well as to judgements of sexual dimorphism, health, and age. However, few studies have investigated the interaction of vocal and facial cues in attractiveness judgments under naturalistic conditions using dynamic, ecologically valid stimuli. Here, we used short videos or audio tracks of females speaking full sentences and used a manipulation of voice pitch to investigate cross-modal interactions of voice pitch on facial attractiveness and related ratings. Male participants had to rate attractiveness, femininity, age, and health of synchronized audio-video recordings or voices only, with either original or modified voice pitch. We expected audio stimuli with increased voice pitch to be rated as more attractive, more feminine, healthier, and younger. If auditory judgements cross-modally influence judgements of facial attributes, we additionally expected the voice pitch manipulation to affect ratings of audiovisual stimulus material. We tested 106 male participants in a within-subject design in two sessions. Analyses revealed that voice recordings with increased voice pitch were perceived to be more feminine and younger, but not more attractive or healthier. When coupled with video recordings, increased pitch lowered perceived age of faces, but did not significantly influence perceived attractiveness, femininity, or health. Our results suggest that our manipulation of voice pitch has a measurable impact on judgements of femininity and age, but does not measurably influence vocal and facial attractiveness in naturalistic conditions.
BACKGROUND AND AIMS: Poor quality of life is a main complaint among individuals with irritable bowel syndrome (IBS). Self-rated health (SRH) is a powerful predictor of clinical outcomes, and also reflects psychological and social aspects of life and an overall sense of well-being. This population-based twin study evaluates how IBS affects ratings of physical and mental health, and influences perceptions of hindrance of daily activity by physical or mental health. Further, we examine how IBS is related to these SRH measures. METHODS: The sample included 5288 Norwegian twins aged 40-80, of whom 575 (10.9%) suffer from IBS. Hierarchical regressions were used to estimate the impact of IBS on perceptions of health, before and after accounting for other chronic physical and mental health conditions. Two dimensions of SRH, physical and mental, and two aspects of functional limitations, the extent to which physical or mental health interferes with daily activities, were included as outcomes in separate models. Co-twin control analyses were used to explore whether the relationships between IBS and the four measures of SRH are causal, or due to shared genetic or shared environment effects. RESULTS: IBS was an independent predictor of poor self-rated physical health (OR = 1.82 [1.41; 2.33]), the size of this effect was comparable to that predicted by chronic somatic conditions. However, in contrast to somatic diseases, IBS was associated with the perception that poorer ratings of mental health (OR = 1.45 [1.02; 2.06]), but not physical health (OR = 1.23 [0.96; 1.58]), interfered with daily activity. The co-twin control analyses suggest that causal mechanisms best explain the relationships between IBS with self-rated physical health and with hindrance of daily activities. In contrast, the relationship between IBS and self-rated mental health was consistent with shared genetic effects. CONCLUSION: IBS is predictive of poor self-rated physical health. The relationship between IBS and self-rated mental health is best explained by shared genetic effects which might partially explain why mental health interferes with daily activity to a larger degree among those with IBS.
On the CIFAR-10 (Canadian Institute for Advanced Research-10), ImageNet, and Penn Treebank datasets, Neural Architecture Search (NAS) algorithms obtained better results by computerizing the process of architectural design on the CIFAR-10, ImageNet, and Penn Treebank datasets. Even though the search time has been simplified, search algorithms count on performance prediction or controllers. When used on a new task with a fresh dataset, this may necessitate optimal structuring. The problem of architecture search is not solved because this is done by hand. Using continuous relaxation and gradient descent methods, Differentiable Architecture Search (DARTS) [1] avoids this issue. There are, however, plenty of intriguing methods to make DARTS better. In this paper, we first split the DARTS’s supernet into three (03) sub-supernets and applied neural message passing so that each node in the graph has information from other nodes. The three (03) sub-supernets by using gradient descent to find the best graph represent the whole search space. By adding parameter sharing and transfer learning, our method enhances the final accuracy of one-shot-based DARTS systems consistently. On CIFAR-10, it achieves 98.25% test set accuracy, according to the results.
The article considers the role of tax administration digitalization bodies in the context of current challenges and threats caused by the pandemic in the world and war in Ukraine. The importance of developing the electronic interaction channels between the State Tax Service of Ukraine, tax administrations of OECD countries and taxpayers in the context of Covid-19 and war action is outlined. The focus is on expanding the list of electronic services (for business, private entrepreneurs, IT services), which will increase opportunities for the implementation of the principle of convenience in fulfilling the tax obligation to the state. Continuous development and improvement of services provided by state tax authorities increase the image rating of state institutions, in particular, carry out electronic taxation, which allows: automate internal tax functions; to build electronic information interaction between taxpayers and state tax authorities in the field of taxation; to form effective online communication and ensure fast and secure data exchange between government agencies in the field of taxation; to ensure effective international cooperation in electronic format. The directions for the tax authorities digitalization strategy change for the purpose of effective taxes and fees administration process are offered: granting equal access of citizens, business representatives, tax administrators to digital technologies and new opportunities (to reduce digital gaps); advanced training of personnel for the full development of digitalization of the State Tax Service; increasing the level of automation and digitalization of public services together with the motivation of government agencies.
Background/Aims: Exposure toward positive emotional cues with – and without – reproductive significance plays a crucial role in daily life and regarding well-being as well as mental health. While possible adverse effects of oral contraceptive (OC) use on female mental and sexual health are widely discussed, neural processing of positive emotional stimuli has not been systematically investigated in association with OC use. Considering reported effects on mood, well-being and sexual function, and proposed associations with depression, it was hypothesized that OC users showed reduced neural reactivity toward positive and erotic emotional stimuli during early as well as later stages of emotional processing and also rated these stimuli as less pleasant and less arousing compared to naturally cycling (NC) women. Method: Sixty-two female subjects (29 NC and 33 OC) were assessed at three time points across the natural menstrual cycle and corresponding time points of the OC regimen. Early (early posterior negativity, EPN) and late (late positive potential, LPP) event-related potentials in reaction to positive, erotic and neutral stimuli were collected during an Emotional Picture Stroop Paradigm (EPSP). At each appointment, subjects provided saliva samples for analysis of gonadal steroid concentration. Valence and arousal ratings were collected at the last appointment. Results: Oral contraceptive users had significantly lower endogenous estradiol and progesterone concentrations compared to NC women. No significant group differences in either subjective stimulus evaluations or neural reactivity toward positive and erotic emotional stimuli were observed. For the OC group, LPP amplitudes in reaction to erotic vs. neutral pictures differed significantly between measurement times across the OC regimen. Discussion: In this study, no evidence regarding alterations of neural reactivity toward positive and erotic stimuli in OC users compared to NC was found. Possible confounding factors and lines for future research are elaborated and discussed.
RST-style discourse parsing plays a vital role in many NLP tasks, revealing the underlying semantic/pragmatic structure of potentially complex and diverse documents. Despite its importance, one of the most prevailing limitations in modern day discourse parsing is the lack of large-scale datasets. To overcome the data sparsity issue, distantly supervised approaches from tasks like sentiment analysis and summarization have been recently proposed. Here, we extend this line of research by exploiting distant supervision from topic segmentation, which can arguably provide a strong and oftentimes complementary signal for high-level discourse structures. Experiments on two human-annotated discourse treebanks confirm that our proposal generates accurate tree structures on sentence and paragraph level, consistently outperforming previous distantly supervised models on the sentence-to-document task and occasionally reaching even higher scores on the sentence-to-paragraph level.
Sequence labeling, in which a class or label is assigned to each token in a given input order, is a fundamental task in natural language processing. Many advanced neural network architectures have recently been proposed to solve the sequential labeling problem affecting this task. By contrast, only a few approaches have been proposed to address the sequential ensemble problem. In this paper, we resolve the sequential ensemble problem by applying the sequential alignment method in a proposed ensemble framework. Specifically, we propose a simple but efficient ensemble candidate generation framework with which multiple heterogeneous systems can easily be prepared from a single neural sequence labeling network. To evaluate the proposed framework, experiments were conducted with part-of-speech (POS) tagging and dependency label prediction problems. The results indicate that the proposed framework achieved accuracy values that were higher by 0.19 and 0.33 than those achieved by the hard-voting method on the Penn-treebank POS-tagged and Universal dependency-tagged datasets, respectively.
We created a Slovak language model in the Spacy library. When creating the model, we used pretraining and training in the Spacy library. During the pretraining, we used Fasttext vectors and the text we collected. We used data from Slovak Dependency Treebank during the training. We created a model with an accuracy of 86.249%. We also examined the effect of pretraining on the model. All achieved results are described in the final part.
Leichte Sprache (LS; easy-to-read German) defines a variety of German characterized by simplified syntactic constructions and a small vocabulary. It provides barrier-free information for a wide spectrum of people with cognitive impairments, learning difficulties, and/or a low level of literacy in the German language. The levels of difficulty of a range of syntactic constructions were systematically evaluated with LS readers as part of the recent LeiSA project ( Bock, 2019 ). That study identified a number of constructions that were evaluated as being easy to comprehend but which fell beyond the definition of LS. We therefore want to broaden the scope of LS to include further constructions that LS readers can easily manage and that they might find useful for putting their thoughts into words. For constructions not considered in the LeiSA study, we performed a comparative treebank study of constructions attested to in a collection of 245 LS documents from a variety of sources. Employing the treebanks TüBa-D/S (also called VERBMOBIL) and TüBa-D/Z, we compared the frequency of such constructions in those texts with their incidence in spoken and written German sources produced without the explicit goal of facilitating comprehensibility. The resulting extension is called Extended Leichte Sprache (ELS). To date, text in LS has generally been produced by authors proficient in standard German. In order to enable text production by LS readers themselves, we developed a computational linguistic system, dubbed ExtendedEasyTalk. This system supports LS readers in formulating grammatically correct and semantically coherent texts covering constructions in ELS. This paper outlines the principal components: (1) a natural-language paraphrase generator that supports fast and correct text production while taking readership-design aspects into account, and (2) explicit coherence specifications based on Rhetorical Structure Theory (RST) to express the communicative function of sentences. The system’s writing-workshop mode controls the options in (1) and (2). Mandatory questions generated by the system aim to teach the user when and how to consider audience-design concepts. Accordingly, users are trained in text production in a similar way to elementary school students, who also tend to omit audience-design cues. Importantly, we illustrate in this paper how to make the dialogues of these components intuitive and easy to use to avoid overtaxing the user. We also report the results of our evaluation of the software with different user groups.
Artificial voices are nowadays embedded into our daily lives with latest neural voices approaching human voice consistency (naturalness). Nevertheless, behavioral, and neuronal correlates of the perception of less naturalistic emotional prosodies are still misunderstood. In this study, we explored the acoustic tendencies that define naturalness from human to synthesized voices. Then, we created naturalness-reduced emotional utterances by acoustic editions of human voices. Finally, we used Event-Related Potentials (ERP) to assess the time dynamics of emotional integration when listening to both human and synthesized voices in a healthy adult sample. Additionally, listeners rated their perceptions for valence, arousal, discrete emotions, naturalness, and intelligibility. Synthesized voices were characterized by less lexical stress (i.e., reduced difference between stressed and unstressed syllables within words) as regards duration and median pitch modulations. Besides, spectral content was attenuated toward lower F2 and F3 frequencies and lower intensities for harmonics 1 and 4. Both psychometric and neuronal correlates were sensitive to naturalness reduction. (1) Naturalness and intelligibility ratings dropped with emotional utterances synthetization, (2) Discrete emotion recognition was impaired as naturalness declined, consistent with P200 and Late Positive Potentials (LPP) being less sensitive to emotional differentiation at lower naturalness, and (3) Relative P200 and LPP amplitudes between prosodies were modulated by synthetization. Nevertheless, (4) Valence and arousal perceptions were preserved at lower naturalness, (5) Valence (arousal) ratings correlated negatively (positively) with Higuchi's fractal dimension extracted on neuronal data under all naturalness perturbations, (6) Inter-Trial Phase Coherence (ITPC) and standard deviation measurements revealed high inter-individual heterogeneity for emotion perception that is still preserved as naturalness reduces. Notably, partial between-participant synchrony (low ITPC), along with high amplitude dispersion on ERPs at both early and late stages emphasized miscellaneous emotional responses among subjects. In this study, we highlighted for the first time both behavioral and neuronal basis of emotional perception under acoustic naturalness alterations. Partial dependencies between ecological relevance and emotion understanding outlined the modulation but not the annihilation of emotional integration by synthetization.
While 360° videos watched in a VR headset are gaining in popularity, it is necessary to lower the required bandwidth to stream these immersive videos and obtain a satisfying quality of experience. Doing so requires predicting the user's head motion in advance, which has been tackled by a number of recent prediction methods considering the video content and the user's past motion. However, human motion is a complex process that can depend on many more parameters, including the type of attentional phase the user is currently in, and their emotions, which can be difficult to capture. This is the first article to investigate the effects of user emotions on the predictability of head motion, in connection with video-centric parameters. We formulate and verify hypotheses, and construct a structural equation model of emotion, motion and predictability. We show that the prediction error is higher for higher valence ratings, and that this relationship is mediated by head speed. We also show that the prediction error is lower for higher arousal, but that spatial information moderates the effect of arousal on predictability. This work opens the path to better capture important factors in human motion, to help improve the training process of head motion predictors.
Many studies in the literature attempt recognition of emotions through the use of videos or images, but very few have explored the role that sounds have in evoking emotions. In this study we have devised an experimental protocol for elicitation of emotions by using, separately and jointly, images and sounds from the widely used International Affective Pictures System and International Affective Digital Sounds databases. During the experiments we have recorded the skin conductance and pupillary signals and processed them with the goal of extracting indices linked to the autonomic nervous system, thus revealing specific patterns of behavior depending on the different stimulation modalities. Our results show that skin conductance helps discriminate emotions along the arousal dimension, whereas features derived from the pupillary signal are able to discriminate different states along both valence and arousal dimensions. In particular, the pupillary diameter was found to be significantly greater at increasing arousal and during elicitation of negative emotions in the phases of viewing images and images with sounds. In the sound-only phase, on the other hand, the power calculated in the high and very high frequency bands of the pupillary diameter were significantly greater at higher valence (valence ratings > 5). Clinical relevance- This study demonstrates the ability of physiological signals to assess specific emotional states by providing different activation patterns depending on the stimulation through images, sounds and images with sounds. The approach has high clinical relevance as it could be extended to evaluate mood disorders (e.g. depression, bipolar disorders, or just stress), or to use physiological patterns found for sounds in order to study whether hearing aids can lead to increased emotional perception.
Focus on language-specific properties with insights from formal minimalist syntax can improve universal dependency (UD) parsing. Such improvements are especially sensitive for low-resource African languages, like Wolof, which have fewer UD treebanks in number and amount of annotations, and fewer contributing annotators. For two different UD parser pipelines, one parser model was trained on the original Wolof treebank, and one was trained on an edited treebank. For each parser pipeline, the accuracy of the edited treebank was higher than the original for both the dependency relations and dependency labels. Accuracy for universal dependency relations improved as much as 2.90%, while accuracy for universal dependency labels increased as much as 3.38%. An annotation scheme that better fits a language's distinct syntax results in better parsing accuracy.