Classification Of Languages Research Articles

The classification of Gurage languages

The classification of Gurage languages

NLPGuard: A Framework for Mitigating the Use of Protected Attributes by NLP Classifiers

AI regulations are expected to prohibit machine learning models from using sensitive attributes during training. However, the latest Natural Language Processing (NLP) classifiers, which rely on deep learning, operate as black-box systems, complicating the detection and remediation of such misuse. Traditional bias mitigation methods in NLP aim for comparable performance across different groups based on attributes like gender or race but fail to address the underlying issue of reliance on protected attributes. To partly fix that, we introduce NLPGuard, a framework for mitigating the reliance on protected attributes in NLP classifiers. NLPGuard takes an unlabeled dataset, an existing NLP classifier, and its training data as input, producing a modified training dataset that significantly reduces dependence on protected attributes without compromising accuracy. NLPGuard is applied to three classification tasks: identifying toxic language, sentiment analysis, and occupation classification. Our evaluation shows that current NLP classifiers heavily depend on protected attributes, with up to 23% of the most predictive words associated with these attributes. However, NLPGuard effectively reduces this reliance by up to 79%, while slightly improving accuracy.

Evaluation and Use of Natural Language Processing (NLP) Reasoning and Classification Models to Support Clinical Trial Patient Identification and Enrollment in the Community Oncology Setting

Clinical trial research in oncology relies heavily on clinical documentation within the electronic medical record (EMR) to ascertain patient eligibility in clinical trials based on inclusion and exclusion criteria. The structured data elements within the EMR serve as the primary information source for defining patient cohorts, with clinical cancer stage and performance status being two pivotal criteria determining trial eligibility. The challenge arises from the inconsistent availability of clinical stage and performance status data within the structured fields of the EMR despite their consistent presence in clinical notes. Additionally, there is a deficiency of standardization of this data that exists in the unstructured field. Hence, due to lack of structured data and standardization of said data, there are limitations in developing artificial intelligence (AI) models. To increase the comprehensiveness of clinical records, a clinical research team at a community oncology practice was consulted to identify requirements and extract essential clinical features from de-identified data. The methods outlined in this paper focused on eliminating false positives to allow future development of Large Language Models (LLM) using the outputted structured fields which resulted in an increase in patient record completeness with high accuracy. The accuracy ranged from 97.5-97.75% for the models that were developed. Out of the 60,000+ patients, the numerical staging, TNM (tumor, node, metastasis) staging, and Karnofsky performance score models added a structured field for 29.62%, 21.01%, and 40.64% patients respectively. Additionally, a semi-supervised NLP algorithm was applied on the performance status algorithm which achieved a mean absolute error (MAE) of 1.57. This work demonstrates the use case of natural language processing (NLP) in optimizing the clinical research enrollment process by providing an efficient and accurate method to detect key clinical values in unstructured patient data. Similar methodology with more advanced algorithms such as LLM can be employed to detect additional patient elements such as molecular biomarkers, imaging reports, postoperative surgical outcomes (i.e., clear margins etc.) and patient treatment outcomes using the extracted structured fields.

Black Caregiver Perspectives During a Developmental Diagnostic Interview.

Despite increasing autism prevalence rates across racial and ethnic groups, research has traditionally overlooked the influence of culture on developmental diagnostic conversations. Addressing this gap in research is crucial to understanding and mitigating potential disparities in diagnostic experiences, especially among Black caregivers. Black caregivers encounter frustration and discrimination during the diagnostic process, citing dismissiveness and a lack of cultural competence from healthcare providers. To address these issues, this study centers the perspectives of Black caregivers as they participate in a developmental diagnostic program for toddlers under three years of age. This retrospective qualitative study, conducted exclusively by a Black research team, involved nineteen Black caregivers of thirteen toddlers as they participated in semi-structured interviews conducted by a multidisciplinary clinical team. The research team employed a conceptual content analysis to examine caregivers' descriptions of their children's behaviors and characteristics, as well as the factors and experiences associated with their descriptions. Caregivers provided distinct descriptions of behaviors and characteristics in the categories of language and communication, learning, and temperament, associating them with situational and dispositional factors and experiences. The findings offer valuable insights for implementing culturally responsive diagnostic practices. Additionally, these results may enhance the quality of caregiver-provider interactions during developmental assessment, leading to earlier and more accurate diagnoses for Black children.

A hybrid method for enhancing aspect sentiment classification in the french language: combining rule-based and learning methods

A hybrid method for enhancing aspect sentiment classification in the french language: combining rule-based and learning methods

The Dual Status of Judeo-Georgian

The purpose of this paper is to show the dual nature of the status of Judeo-Georgian based on its studies, theories, and methods adopted from variationist sociolinguistics, distinctive linguistic repertoire approach to Jewish languages, using descriptive/synchronic, diachronic, and comparative-contrastive methods. Based on the traditional studies in Georgia in the twentieth century, Judeo-Georgian was considered one of the varieties of Georgian, and the main works were about presenting its features and searching for similarities from standard Georgian or its dialects. The purpose of such studies was to give Judeo-Georgian a status in the Georgian linguistic space (be it a dialect of Georgian, a sub-dialect of Georgian dialects, an ethnolect, an ethno-socilect). Today, in the 21st century, Judeo-Georgian studies have acquired new dimensions. It is already considered as one of the varieties not only of Georgian but also of Jewish languages. In the works of comparative linguistics of Jewish languages, there exist common features characteristic of all Jewish languages and Judeo-Georgian too is familiar with these criteria. Therefore, searching for its similarities and differences with other Jewish languages has begun. From this point of view, the understanding of Judeo-Georgian has acquired a new dimension at the modern stage and has become interesting in two ways: 1. as an ethno-dialect of Georgian with its sub-dialects; two as one of the Jewish languages, which has a certain place in the modern classification of Jewish languages.

Phrasing Crisis Communication: How are Distinct Crisis Response Strategies Put Into Words?

This study employs a quantitative content analysis to investigate how reputation-restoring crisis responses (i.e., denial, diminishment, and apology) from organizations and public figures are verbalized. In addition, this study examines if distinct crisis response strategies are systematically associated with particular language categories (i.e., language abstraction, uncertain language, passive voice, and emotional language). By analyzing 179 audiovisual crisis messages (e.g., press conference broadcasts, television interviews, social media videos), this study shows that denials are generally expressed in the most concrete way (i.e., precise numbers), diminish strategies in the most uncertain way (i.e., hedges), and apologies are the most emotionally charged, primarily referring to the emotions of sadness and shame. These findings highlight the significant role of language in crisis communication and help practitioners in the field to become aware of key linguistic features that could influence the effectiveness of a crisis message.

Possession and syntactic categories: An argument from Äiwoo

AbstractThis paper argues that possession is syntactically category-flexible. While it is clear that in many languages possession is mostly grounded in and operates in the nominal extended projection (Szabolcsi 1983; Kayne 1993), I show that this cannot be universal. The empirical part of this article is a case study of Äiwoo, which I argue has an inherently verbal counterpart of English ’s, an abstract transitive verb I label poss. This verb can be used by itself to form clausal possession: ‘I poss this boat’ ≈ ‘this boat is mine.’ Possessed DPs also contain the verb poss: the object of this verb is extracted, forming a relative clause. Informally, ‘my boat’ really is ‘the boati"Equation missing"’ ≈ ‘the boat that is mine.’ Given this, Äiwoo simply lacks true nominal possessives. The theoretical consequence is that possession can be mapped onto different syntactic categories in different languages. This is a welcome result, as it makes the syntax-semantics mapping as flexible as it needs to be: if possession is just a tool to assert that a certain relation holds between two entities, nothing in our theory of grammar predicts that such a notion should only be limited to a specific syntactic category.

A Hybrid Ensemble Approach for Greek Text Classification Based on Multilingual Models

The present study explores the field of text classification in the Greek language. A novel ensemble classification scheme based on generated embeddings from Greek text made by the multilingual capabilities of the E5 model is presented. Our approach incorporates partial transfer learning by using pre-trained models to extract embeddings, enabling the evaluation of classical classifiers on Greek data. Additionally, we enhance the predictive capability while maintaining the costs low by employing a soft voting combination scheme that exploits the strengths of XGBoost, K-nearest neighbors, and logistic regression. This method significantly improves all classification metrics, demonstrating the superiority of ensemble techniques in handling the complexity of Greek textual data. Our study contributes to the field of natural language processing by proposing an effective ensemble framework for the categorization of Greek texts, leveraging the advantages of both traditional and modern machine learning techniques. This framework has the potential to be applied to other less-resourced languages, thereby broadening the impact of our research beyond Greek language processing.

Effects of Language Policies of Nigeria and South Africa on Linguistic Diversity and Classification of Indigenous Languages

This study investigated the effects of language policies implemented in Nigeria and South Africa on linguistic diversity and how these two nations categorized their indigenous languages. In this present study, qualitative research methodology was used, and a purposive sampling method was used to select Nigeria’s 2013 Language Policy from the National Policy of Education and South Africa’s 1996 Language in Education Policy. Thematic analysis was carried out using Latent Dirichlet Allocation (LDA) software which is a topic modeling methodology that assists in identifying basic themes within wide textual data. The latent level of analysis converged on the implicit ideas, suppositions, notions, and orientations that describe or point out the meaning of the data. The action of incorporating LDA software intensifies the efficiency and level of the qualitative research methodology used to critically examine the language policies of Nigeria and South Africa. The results revealed that the language policies in Nigeria and South Africa have a notable influence on linguistic diversity, supporting the principles of translanguaging and multilingualism. Also, the policies give priority to the inclusion of indigenous languages in education to improve unity and preserve cultural heritage. With a careful survey of language policies and their significance, the results give important and useful recommendations on the increase of inclusivity in language policies for the preservation and promotion of linguistic diversity in different cultural environments. Native languages should be given a notable preference in language policies so that there will be room for the use of translanguaging and multilingual pedagogies in the school systems. Keywords: Inclusive Education, Indigenous Languages, Language Policy, Linguistic Diversity, Multilingualism, Translanguaging

Dialectology Landscape: Duano Sea Tribe Communication at Jambi Province

Duano language in Sabak and Tungkal is included into the extinct language category. One way to avoid language extinction is through language revitalization or language documentation. The purpose of this research is to ascertain and explain the proto language and the kinship of Duano tribe in Sabak and Tungkal through dialectology. Then, language documentation is carried out to look at all related aspects from cognate to morpheme positioning. The writers used Proto Malayic from Adelaar (1992) as IPD instrument, The data sources are Duano language from Sabak and Tungkal regencies spoken by 10 respondents. This is a qualitative study. The data were collected through interview, note-taking, recording, and documentation. Then, they were analyzed comparatively. This research found 920 zero data, 510 lexical differences data, 1210 cognate, and 570 phonological differences data. This showed that almost 50% of the data has zero difference, which means they have the same vocabulary. Then, phonological difference is 7.01% indicating that the two languages are different Speech. Meanwhile, lexical difference is 35.29% implying that the two are different subdialects, The relationship between the two regions is very close., the number of vocabularies that belong to vocabulary shift was higher than the vocabulary maintenance. This means that language extinction is very high, especially in the Tungkal area.

Trends in Mutual Transpositions of Lexical and Grammatical Noun Categories in Russian Language

This study addresses the gap in understanding the lexico-grammatical-structural types of nouns, particularly how lexico-semantic variants gain new connotations during transposition. Drawing upon Babaytseva’s theory of transitivity phenomena and building on Mel’chuk’s text semantic theory, we identify the schemes and mechanisms of transposition relations concerning the lexical and grammatical categories of modern Russian nouns, by analyzing transposition phenomena and their semantic mechanisms in 660 Russian nouns with indefinite denotation and lexical and grammatical status. The research details all possible immigration transpositions of the lexical and grammatical categories of nouns at the minimal context level and their mechanisms. Findings reveal that approximately 27% of the examined nouns carry connotative meanings, encompassing functional-stylistic, cultural-ethnic, and emotional-evaluative facets. Additionally, the transposition of nouns can occur within or between categories. The principal mechanisms of transposition at the syntagmatic level include metaphors and metonymy. Transpositions at the lexical and semantic categories of nouns coincide with several phenomena: changes in the numerical paradigm, phraseologization, connotation, and the unveiling of the noun’s stylistic potential.

Application of Deep Learning for Voice Command Classification in Turkish Language

In this study, a deep learning model was developed for the recognition and classification of voice commands using the Turkish Speech Command Dataset. The division of training, validation, and test sets was carried out on an individual basis. This approach aims to prevent the model from memorizing and to enhance its generalization capability. The model was trained using Mel-Frequency Cepstral Coefficients (MFCC) features extracted from voice files, and its classification performance was evaluated in detail. The findings indicate that the model successfully classifies voice commands with a high accuracy rate, achieving an overall accuracy of 92.3% on the test set, highlighting the potential of deep learning approaches in voice recognition technologies

Occurrence of Transposition and Substantivization Phenomena Between Adjectival and Noun Word Categories in Uzbek

This study investigates the phenomena of transposition and substantivization between adjectival and noun word categories in the Uzbek language, an area that remains underexplored in linguistic research. By analyzing selected texts, the study identifies shifts where adjectives become nouns (substantivization) and nouns assume adjectival roles (transposition), focusing on their contextual usage in proverbs, literary works, and everyday speech. The research identifies 150 instances of these shifts, demonstrating the flexibility and dynamic nature of word categories in Uzbek. The findings provide new insights into language evolution and structural adaptability, with broader implications for understanding the linguistic characteristics of Turkic languages.

The political theater of August 1917: stage and backstage

Analyzing the monograph by A.B. Nikolaev "The State Conference of 1917: convocation, composition, activity", the author raises the question of the nature of the emerging political system of Russia in 1917, about its internal contradictions, which contributed to the holding of crowded forums, the establishment of new collegial authorities. In fact, there is a significant discrepancy between the institutional design of power and the categorical language of its description, between the social experience of those who tried to determine the political agenda, and the agenda itself, the popular slogans of that time. For this reason, the course of the Moscow State Conference, which became the subject of A.B. Nikolaev's close attention, is very indicative for understanding the political players of that time: they did not fully represent the "game" in which they participated. Moreover, they were unable to determine its rules.

Quantifying decision support level of explainable automatic classification of diagnoses in Spanish medical records

Background and Objective:In the realm of automatic Electronic Health Records (EHR) classification according to the International Classification of Diseases (ICD) there is a notable gap of non-black box approaches and more in Spanish, which is also frequently ignored in clinical language classification. An additional gap in explainability pertains to the lack of standardized metrics for evaluating the degree of explainability offered by distinct techniques. Methods:We address the classification of Spanish electronic health records, using methods to explain the predictions and improve the decision support level. We also propose Leberage a novel metric to quantify the decision support level of the explainable predictions.We aim to assess the explanatory ability derived from three model-independent methods based on different theoretical frameworks: SHapley Additive exPlanations (SHAP), Local Interpretable Model-agnostic Explanations (LIME), and Integrated Gradients (IG). We develop a system based on longformers that can process long documents and then use the explainability methods to extract the relevant segments of text in the EHR that motivated each ICD. We then measure the outcome of the different explainability methods by implementing a novel metric. Results:Our results beat those that carry out the same task by 7%. In terms of explainability degree LIME appears as a stronger technique compared to IG and SHAP. Discussion:Our research reveals that the explored techniques are useful for explaining the output of black box models as the longformer. In addition, the proposed metric emerges as a good choice to quantify the contribution of explainability techniques.

Telehealth general case parent training for children at risk for autism.

Parent-mediated interventions for infants and young children with an increased likelihood of autism may help ameliorate developmental concerns; however, generalization of parents' teaching strategies to novel child target skills has not been consistently demonstrated. This study expanded our parent training program, Parent Intervention for Children at-Risk for Autism (PICARA), by incorporating telehealth general case training (PICARA-TGCT) to promote generalization of teaching skills. Five parent-child dyads participated. Child target skills were chosen from the categories of imitation, receptive language, and expressive language. A concurrent multiple-baseline-across-participants design was used to evaluate the effect of training across two cohorts of parent-child dyads. Dependent variables included the percentage of correct parent teaching skills and the percentage of child correct responses. Parent teaching skills increased across all participants for both trained and untrained child target skills, as did child skills. This study provides support for PICARA-TGCT as an efficacious and efficient early intervention model.

Hybrid Convolution-Based Efficientnet-Based Hand Gesture Recognition Framework with Optimized Algorithm

The difficulties in communication and hearing are an important concern for deaf–dumb people, which stop access to their essential and basic needs. Many findings have been made to address sign languages even though this challenging problem is not still solved. Many methods aimed to propose vision-based classifiers through identical pattern investigation tasks by obtaining the difficult handcraft feature descriptions of gestures from the gathered images. However, the efficacy of all those models is less for performing with a huge signbook captured from uncontrolled and complex background conditions. So, an effective Indian Sign Language (ISL) classification method is developed by an advanced deep learning approach. At first, the hand gesture images are obtained from the data source. Only the image of the hand, even from a complicated background, is extracted from the obtained image. The features are extracted using the Scale-Invariant Feature Transform (SIFT) method and Multiscale Vision Transformer (MVT). Then, the extracted features are fed to the Hybrid Convolution-based EfficientNet (HCEN) model. The hyper-parameters in the developed HCEN model are tuned using the implemented Adaptive Political Optimizer (APO) algorithm. The recognized hand signs are obtained from the suggested HCEN model. Various experiments are conducted to determine the performance of the suggested deep learning-based hand gesture recognition model.

From color naming to color perception: Cross‐linguistic differences of the chromatic information processing in monolingual and bilingual speakers

AbstractThere has been long running debate about the interaction of language and perception. In this context, bilingual people have often shown benefits, due to their double‐active linguistic system, in cognitive functions, like inhibition, attention, and memory, which are central for visual perception. Color naming and categorization are domains for studying cross‐linguistic effects, which arise from conceptual and perceptual variations across speakers of different languages. In this study, we compared the ‘blue’ lexicon of highly proficient French‐Italian bilinguals to monolingual speakers of the corresponding languages. Prior studies have shown that Italian has two basic color terms for the blue area of color space: one denotes light blue hues (azzurro) and the other dark blue hues (blu), whilst French, with only one basic term bleu, lacks this distinction. We used a Stroop test to probe differences in perception and categorization of blue in bi‐ and monolingual speakers. We found that Italian monolinguals name the ink color more accurately and more rapidly when the word blu is rendered in dark blue ink (corresponding to the word blu) than when it is printed in light blue ink (corresponding to azzurro), since the latter represents an incongruent condition for them. This ‘category effect’ does not exist for French monolingual speakers. Our bilinguals' results demonstrate that, despite the emergence of a specific in‐between perceptual behavior, bilinguals generally performed like Italian monolinguals. These outcomes confirm the hypothesis that their second language categories (Italian) dominate their native language (French), attesting that lexical distinctions influence perceptual faculties in general. However, the ratio of the interference effect (longer reaction times for incongruent stimuli compared to control stimuli) and the facilitation effect (shorter reaction times for congruent stimuli compared to control stimuli) is not the same for bilinguals and monolinguals. The highest magnitude in the facilitation effect was revealed for bilinguals, whereas the highest magnitude in the interference effect was revealed for Italian monolingual speakers. This phenomenon adds evidence to the existence of enhanced bilingual cognitive control abilities.

CELEBRATE Feeding: A Responsive Approach to Food and Feeding in Early Learning Settings.

Early learning and child care (ELCC) settings in Canada follow nutrition standards that outline food provisions, with many also encouraging responsive feeding practices that help to create a supportive environment for children. Caregivers who lack confidence in children's ability to regulate their own intake, or those who feel stressed about mealtime, may unknowingly engage in less responsive feeding practices. The CELEBRATE Feeding Approach is a flexible framework, driven by behaviour change theory, that builds on previous definitions and concepts of responsive feeding in ELCC environments. Through this approach, there is an intentional focus on supporting early childhood educators to implement feeding practices that are more responsive. The approach incorporates 13 target educator behaviours related to the three overlapping categories of CELEBRATE language, CELEBRATE Mealtime, and CELEBRATE Play. These practices recognize and support the development of a child's sense of autonomy, confidence, and self-regulation not only at mealtimes but also through play-based exploration and language that is used throughout the day around food and feeding. The goal is that children will be open to a wide variety of food, develop their self-regulation skills, and build the foundation for a positive relationship with food throughout their lifetime.