Arabic dialect identification in social media: A hybrid model with transformer models and BiLSTM.

Arabic Dialect Identification (ADI) is a challenging task in natural language processing applications due to its diversity and regional variations. Despite previous efforts, this task is still difficult. Therefore, this study aims to use transformers to address the issue of ADI on social media. A combination of two hybrid models is proposed in this study: one that combines Bidirectional Long Short-Term Memory (BiLSTM) with CAMeLBERT, and the second model that combines the BiLSTM model with AlBERT. In addition, a novel dataset comprising 121,289 user-generated comments from various social media network platforms and four major Arabic dialects (Egyptian, Jordanian, Gulf and Yemeni) was introduced. Several experiments have been conducted using conventional Machine Learning Classifiers (MLCs) and Deep Learning Models (DLMs) as baselines to measure the performance and effectiveness of the proposed models. In addition, binary classification is performed between two dialects to determine which are closest to each other. The performance of the model is measured using common metrics such as precision, recall, F-score and F-measure. Experiment results demonstrate the superior efficiency of the proposed hybrid models in ADI, CAMeLBERT with BiLSTM and ALBERT with BiLSTM, which both recorded an accuracy of 87.67 % and 86.51 %, respectively.

Female African elephant rumbles differ between populations and sympatric social groups.

Vocalizations often vary in structure within a species, from the individual to population level. Vocal differences among social groups and populations can provide insight into biological processes such as vocal learning and evolutionary divergence, with important conservation implications. As vocal learners of conservation concern, intraspecific vocal variation is of particular interest in elephants. We recorded calls from individuals in multiple, wild elephant social groups in two distinct Kenyan populations. We used machine learning to investigate vocal differentiation among individual callers, core groups, bond groups (collections of core groups) and populations. We found clear evidence for vocal distinctiveness at the individual and population level, and evidence for much subtler vocal differences among social groups. Social group membership was a better predictor of call similarity than genetic relatedness, suggesting that subtle vocal differences among social groups may be learned. Vocal divergence among populations and social groups has conservation implications for the effects of social disruption and translocation of elephants.

Dialect and mathematics performance in African American children who use AAE: Insights from explanatory IRT and error analysis.

Little is known about mismatches between the language of mathematics testing instruments and the rich linguistic repertoires that African American children develop at home and in the community, in part because research paradigms with African American English (AAE) dialect speakers face complex challenges in measurement, historical exclusion, and other social, economic, cultural, and linguistic confounds. The current study aims to provide a proof of concept and novel explanatory item response design that uses error analysis to investigate the relationship between AAE child language and children's mathematics assessment outcomes. Here, we illustrate 2nd and 3rd grade children's qualitative patterns of performance on arithmetic tasks in relation to their AAE dialect use and elaborate a unified framework for examining child and item level linguistic characteristics. Results suggest that children draw upon their emerging (bi)dialectal repertoire with arithmetic problems when selecting appropriate problem-solving strategies on language-formatted problems. The mismatch of assessment language formatting with children's repertoires may disadvantage AAE speakers' strategy selections and result in a language-based performance disadvantage unrelated to mathematical ability. Research designs that look beyond correct/incorrect scoring to examine qualitative patterns of performance in AAE speaking children can provide valuable and oft-overlooked evidence when considering equity in mathematics assessment formats.

Widespread cultural change in declining populations of Amazon parrots.

Species worldwide are experiencing anthropogenic environmental change, and the long-term impacts on animal cultural traditions such as vocal dialects are often unknown. Our prior studies of the yellow-naped amazon (Amazona auropalliata) revealed stable vocal dialects over an 11-year period (1994-2005), with modest shifts in geographic boundaries and acoustic structure of contact calls. Here, we examined whether yellow-naped amazons maintained stable dialects over the subsequent 11-year time span from 2005 to 2016, culminating in 22 years of study. Over this same period, this species suffered a dramatic decrease in population size that prompted two successive uplists in IUCN status, from vulnerable to critically endangered. In this most recent 11-year time span, we found evidence of geographic shifts in call types, manifesting in more bilingual sites and introgression across the formerly distinct North-South acoustic boundary. We also found greater evidence of acoustic drift, in the form of new emerging call types and greater acoustic variation overall. These results suggest cultural traditions such as dialects may change in response to demographic and environmental conditions, with broad implications for threatened species.

Orientation towards the vernacular and style-shifting as language behaviours in speech of first-generation Polish migrant communities speaking Norwegian in Norway.

This study describes the patterns of dialect use among L3 Norwegian speakers born in Poland who have migrated to Norway. We collected the data in the form of sociolinguistic interviews recorded in Tromsø and Oslo, two different dialect regions, in order to examine potential differences in acquisition of two dissimilar dialects in Norwegian by L3 speakers. The analyses focus on dialectal and accentual variation in their speech, and whether frequency of dialect use is dependent on selected sociocultural factors. We have found that some speakers, especially those scoring high for overall dialect use, also display style-shifting, i.e. they use dialect features from the region more frequently in unscripted speech as opposed to in more formal speech styles elicited through reading tasks or the wordlist reading tasks. This demonstrates that language learners are capable of developing sensitivity towards the vernacular form in an L3. Moreover, it shows that first-generation migrant communities in fact may be capable of developing their L2/L3/L4 language competencies in a similar way to L1 speakers, including at the level of sociolinguistic variation.

Speaking your language: The psychological impact of dialect integration in artificial intelligence systems.

As the popularity and adoption of Artificial Intelligence (AI) systems continue to rise, this article presents a promising proposition: the use of AI dialects to enhance AI perception. By delving into the potential of personalized AI dialects to augment user perceptions of warmth, competence, and authenticity, the article underscores the pivotal role of anthropomorphism in fortifying trust, satisfaction, and loyalty to AI systems. A comprehensive research framework is put forth to explore these potential mechanisms and outcomes of AI dialect introduction, shedding light on how these impacts might vary based on AI modality (text, voice, and video), industry adoption, and user demographics.

The U-shape Association between Population Agglomeration and Individual Depression: the Role of Dialect Diversity.

Depression is a relevant mental illness affecting hundreds of millions of people worldwide. As urbanization accelerates, agglomeration of populations has altered individual social network distances and life crowding, which in turn affects depressive prevalence. However, the association between depression and population agglomeration (PA) remains controversial. This study aims to explore whether and how PA could influence individual depression. Based on the China Health and Retirement Longitudinal Study (CHARLS) 2018, the empirical results showed that there was a U-shaped association between PA and individual CES-D scores. As PA increases, the risk of depression first decreases and then increases. CES-D was lowest at moderate aggregation. Dialect diversity (DD) was positively related to the incidence of individual depression. The higher the DD, the higher the risk of depression. Meanwhile, DD also played a moderating role in the association between PA and individual depression. Our observations suggest that the optimistic level of agglomeration for individual mental health is within 1500 to 2000 persons per square kilometer.

Robot-based solution for helping Alzheimer patients.

Alzheimer's is a progressive and debilitating neurological disorder characterized by cognitive decline, memory loss, and impaired daily functioning. It is an irreversible brain disease that destroys memory, thinking, and the ability to carry out daily activities. It poses significant challenges for patients and healthcare providers. Modern societies are trying to enhance the quality of people's lives, including Alzheimer's patients. In this study, we explored the potential of social robots to provide emotional support, improve cognitive function, and facilitate communication among Alzheimer's patients. This was achieved by initiating conversations on various topics such as family, relationships, and daily activities. This paper contributes to the literature by introducing a novel and well-organized framework for building an Alzheimer's care robot. Further, this study enriches the literature by introducing the Alzheimer Care Companion Robot (ACCR), designed to identify Alzheimer's patients. The ACCR initiates conversations in the native Arab-Kuwaiti dialect, displaying relevant memories through images and videos on its screen to assist in memory recall based on the individuals' life experiences. The proposed ACCR consists of 271 conversations belonging to three main categories: active, proactive, and graphical user interface (GUI) dialogs comprising 112 dialogs, 109 dialogs, and 50 dialogs for active, proactive, and GUI, respectively. The experimental result illustrated the success of the proposed solution.

Dataset for the recognition of Kurdish sound dialects.

Dialect recognition System (DRS) is a highly significant subject within the field of speech analysis. The performance of speech recognition systems is adversely impacted by factors such as the age, gender, and dialect features of the speaker. In order to address variations in dialect, it is possible to incorporate DRS into speech recognition systems. The system can be configured to utilize the appropriate speech recognition model based on the identification of the spoken dialect. Currently, there is a lack of available datasets suitable for the development of automatic dialect recognition systems specifically tailored for the Kurdish language. The proposed dataset under consideration is assessed using experimental data that has been gathered by personnel associated with the Computer Science Department at the University of Halabja. As the Kurdish language has three main dialects: Northern Kurdish (Badini variation), Central Kurdish (Sorani variant), and Hawrami, three dialects are included in the dataset.

Different divergence processes of isoglosses of folk nomenclature between wild trees and rice landraces imply the need for different conservation planning based on the type of plant resources.

BACKGROUND: The intensification of production and socio-economic changes have accelerated the loss of local traditional knowledge and plant resources. Understanding the distribution and determinants of such biocultural diversity is essential in planning efficient surveys and conservation efforts. Because the concept of biocultural diversity in socio-ecological adaptive systems comprises biological, cultural, and linguistic diversity, linguistic information should serve as a surrogate for the distribution of local biological and cultural diversity. In this study, we spatio-linguistically evaluated the names of local trees and rice landraces recorded in Ehime Prefecture, southwestern Japan. METHODS: Hierarchical clustering was performed separately for the names of local trees and rice landraces. By considering innate flora differences and species having multiple local names, a novel distance index was adopted for local tree names. For the names of rice landraces, Jaccard distance was adopted. V-measure and factor detector analysis were used to evaluate the spatial association between the isogloss maps of the folk nomenclature derived from the clustering and multiple thematic maps. RESULTS: Local tree names showed stronger spatial association with geographical factors than rice landrace names. One folk nomenclature group of trees overlapped well with the slash-and-burn cultivation area, suggesting a link between the naming of trees and the traditional production system. In contrast, rice landraces exhibited stronger associations with folklore practices. Moreover, influences of road networks and pilgrimages on rice landraces indicated the importance of human mobility and traditional rituals on rice seed transfer. High homogeneity and low completeness in the V-measure analysis indicated that the names of local trees and rice landraces were mostly homogenous within current municipalities and were shared with a couple of adjacent municipalities. The isogloss maps help to illustrate how the biological and cultural diversity of wild trees and rice landraces are distributed. They also help to identify units for inter-municipal collaboration for effective conservation of traditional knowledge related to those plant resources and traditional rice varieties themselves. CONCLUSIONS: Our spatio-linguistic evaluation indicated that complex geographical and sociological processes influence the formation of plant folk nomenclature groups and implies a promising approach using quantitative lexico-statistical analysis to help to identify areas for biocultural diversity conservation.

Dialect identification, intelligibility ratings, and acceptability ratings of dysarthric speech in two American English dialects.

The current study explored the intelligibility and acceptability ratings of dysarthric speakers with African American English (AAE) and General American English (GAE) dialects by listeners who identify as GAE or AAE speakers, as well as listener ability to identify dialect in dysarthric speech. Eighty-six listeners rated the intelligibility and acceptability of sentences extracted from a passage read by speakers with dysarthria. Samples were used from the Atlanta Motor Speech Disorders Corpus and ratings were collected via self-report. The listeners identified speaker dialect in a forced-choice format. Listeners self-reported their dialect and exposure to AAE. AAE dialect was accurately identified in 63.43% of the the opportunities; GAE dialect was accurately identified in 70.35% of the opportunities. Listeners identifying as AAE speakers rated GAE speech as more acceptable, whereas, listeners identifying as GAE speakers rated AAE speech as more acceptable. Neither group of listeners demonstrated a difference in intelligibility ratings. Exposure to AAE had no effect on intelligibility or acceptability ratings. Listeners can identify dialect (AAE and GAE) with a better than chance degree of accuracy. One's dialect may have an effect on intelligibility and acceptability ratings. Exposure to a dialect did not affect listener ratings of intelligibility or acceptability.

Lexical diversity in kinship across languages and dialects.

Languages are known to describe the world in diverse ways. Across lexicons, diversity is pervasive, appearing through phenomena such as lexical gaps and untranslatability. However, in computational resources, such as multilingual lexical databases, diversity is hardly ever represented. In this paper, we introduce a method to enrich computational lexicons with content relating to linguistic diversity. The method is verified through two large-scale case studies on kinship terminology, a domain known to be diverse across languages and cultures: one case study deals with seven Arabic dialects, while the other one with three Indonesian languages. Our results, made available as browseable and downloadable computational resources, extend prior linguistics research on kinship terminology, and provide insight into the extent of diversity even within linguistically and culturally close communities.

Considerations of dialect on the identification of speech sound disorder in Vietnamese-speaking children.

BACKGROUND: The dialect spoken by children influences diagnostic decision-making regarding the identification and severity of speech sound disorder (SSD). AIMS: The primary objective was to review papers that examined the influence of dialect on the identification of SSD in Vietnamese-speaking children. METHODS & PROCEDURES: Five studies of mono- and multilingual Vietnamese-speaking children living in Vietnam and Australia were reviewed to examine the influence of dialect on the assessment and analysis children's speech. The main Vietnamese dialects (Standard, Northern, Central, Southern) differ in the production of consonants, vowels and tones. MAIN CONTRIBUTION: Most speech assessments define correct production using the standard dialect of a language. Insights from recent studies of Vietnamese provide recommendations for also considering dialect in diagnostic decision-making. First, we recommend adding column(s) to the assessment score sheet that includes the dialectal variants spoken by adults in the child's family or community. Second, we calculate the accuracy of production twice, based on the standard form and dialectal form. Third, we report the percentage of consonants correct-standard (PCC-S) and percentage of consonants correct-dialect (PCC-D). CONCLUSIONS & IMPLICATIONS: Diagnostic decision-making is influenced by dialectal variation in children's speech, so speech and language therapists need to compare standard and dialectal productions when undertaking assessments, analysis and diagnostic decision-making. WHAT THIS PAPER ADDS: What is already known on the subject Most speech assessments use the standard form of a dialect as the correct production. The standard dialect is used for diagnosis of SSD and identification of intervention targets. What this paper adds to the existing knowledge This paper examines five research studies of Vietnamese to identify ways to consider dialect in speech and language therapy assessment and analysis. Vietnamese provides a complex environment for this examination since there are numerous Vietnamese dialects and they differ according to consonants, vowels and tones. What are the potential or actual clinical implications of this work? Speech and language therapists are encouraged to add column(s) to their assessment score sheet that includes the dialectal variants spoken by adults and to report the PCC-S and PCC-D.

Coaching Opera Secco Recitatives by Using Singers' Native Dialects: A Pilot Study.

OBJECTIVES: This study investigated whether individual coaching can improve role portrayal and emotional expression in opera secco recitatives. STUDY DESIGN: An experimental comparative study. METHODS: Eighteen students (mean age 24 years) from Norway and Sweden majoring in opera singing volunteered as test persons. As singing tasks each test person studied a short opera secco recitative in their Voice Fach during 14 days. All recitatives were picked up from renowned Italian operas from the Classical period. The texts were translated in the singers' mother tongue but performed in Italian, a foreign language to all test persons Thirteen test persons were enrolled in a Test group, and they were coached individually for 1h in interpretation of their singing tasks. During the intervention the test persons combined speaking the translated text in their native dialects with singing the recitative in Italian. The test persons were asked to distinguish various emotions and motivation in the singing task, and rehearsed how to express the singing task equally distinctly in Italian as in their native dialects. The singing task was recorded a capella immediately before and after the intervention. The test persons in the Control group recorded their singing tasks pre and post 1 h self-directed rehearsing on their singing tasks, without any coaching.. The recordings of all test persons were presented in randomized order for ten expert listeners to rate the expression of emotions and role portrayal. The samples were also analysed acoustically. RESULTS: Post coaching samples were rated significantly better in expression of various emotions and motivation compared to pre coaching samples. The Control group singers showed no significant changes in perceptual variables. Pausing increased in both groups after rehearsing. CONCLUSIONS: A short individual coaching in performance of an opera recitative seemed to stimulate young opera singers` emotional and motivational expression.

Home language variation in the narratives of urban First Nations Australian children in their first year of school.

First Nations children may speak a dialect of English that has different grammatical rules from Standard Australian English (school language). Limited studies have investigated Aboriginal English (home language) dialect in First Nations children and its impact on differential diagnosis of language disorder. This study measured the density of home language dialect and grammatical accuracy in oral narratives produced by typically developing First Nations children. Non-standardised assessment narrative protocols were used to elicit language samples from 27 Australian First Nations children aged 4.5-6 years. Local home language dialectal features were coded into the sample and grammatical accuracy was calculated separately for school language and home language. All children displayed some use of home language features. The most common home language features used were alternative use of regular past tense and irregular past tense, zero use of regular and irregular past tense, and alternative use of pronouns. Dialect density varied highly amongst participants. Grammatical accuracy was higher for home language than school language. Speech pathologists and teachers need to be aware of differences between home and school language for First Nations children to avoid misdiagnosis of language disorder. More research is required to gain normative data that informs culturally appropriate assessment practices for this population.

Cue Weighting in Perception of the Retroflex and Non-Retroflex Laterals in the Zibo Dialect of Chinese.

This study investigated cue weighting in the perception of the retroflex and non-retroflex lateral contrast in the monosyllabic words /ɭə/ and /lə/ in the Zibo dialect of Chinese. A binary forced-choice identification task was carried out among 32 natives, using computer-modified natural speech situated in a two-dimensional acoustic space. The results showed that both acoustic cues had a significant main effect on lateral identification, with F1 of the following schwa being the primary cue and the consonant-tos-vowel (C/V) duration ratio as a secondary cue. No interaction effect was found between these two acoustic cues. Moreover, the results indicated that acoustic cues were not equally weighted in production and perception of the syllables /ɭə/ and /lə/ in the Zibo dialect. Future studies are suggested involving other acoustic cues (e.g., the F1 of laterals) or adding noise in the identification task to better understand listeners' listening strategies in their perception of the two laterals in the Zibo dialect.

Fine-scale cultural variation reinforces genetic structure in England.

OBJECTIVES: Genes and languages both contain signatures of human evolution, population movement, and demographic history. Cultural traits like language are transmitted by interactions between people, and these traits influence how people interact. In particular, if groups of people differentiate each other based on some qualities of their cultures, and if these qualities are passed to the next generation, then this differentiation can result in barriers to gene flow. Previous work finds such barriers to gene flow between groups that speak different languages, and we explore this phenomenon further: can more subtle cultural differences also produce genetic structure in a population? We focus on whether subtle, dialect-level linguistic differences in England have influenced genetic population structure, likely by affecting mating preferences. MATERIALS AND METHODS: We analyze spatially dense linguistic and genetic data-both of which independently contain spatially structured variation in England-to examine whether the cultural differences represented by variation in English phonology colocalize with higher genetic rates of change. RESULTS: We find that genetic variation and dialect markers have similar spatial distributions on a country-wide scale, and that throughout England, linguistic boundaries colocalize with the boundaries of genetic clusters found using fineSTRUCTURE. DISCUSSION: This gene-language covariation, in the absence of geographic barriers that could coordinate cultural and genetic differentiation, suggests that similar social forces influenced both dialect boundaries and the genetic population structure of England.

The voice as a material clue: a new forensic Algerian Corpus.

Dialects have received bigger interest in recent years as they are increasingly used on the web and social media. Because Algerian Arabic dialects suffer from a lack of appropriate speech corpora for speech recognition, a rich dialect corpus is needed to approach Algerian Accent recognition. The latter remains a key feature in the field of Forensic Voice Comparison (FVC) systems. This paper presents a new large-scale forensic Algerian speech corpus called Sawt El-Djazaïr. An important criterion in dealing with forensic corpora is the presence of session variability. For this purpose, we collected celebrity recordings in various regions of Algeria, from different social networks, in various scenarios, and at different times. In addition, we also recorded 87 participants using cellular calls and voice over IP (VoIP) applications including Viber, WhatsApp, and Google Meet. The corpus of approximately 50 hours covers various speech topics and is spoken in twelve Algerian sub-dialects. The design guidelines of the proposed corpus are described along with the grouping of dialects across different geographical locations. Sawt El-Djazaïr is available to the research community upon request.

Cross-cultural adaptation and validation of the Moroccan version of the Glaucoma Quality of Life-15 (GQL-15) questionnaire in patients with glaucoma.

BACKGROUND: Glaucoma is the second leading cause of blindness in the world. In addition, visual impairment and psychological strain have been shown to have a significant impact on quality of life (QoL) in glaucoma patients. Maintaining the quality of life of glaucoma patients has become an important component of treatment. The goal of this study is to develop a Moroccan Arabic dialect version of the Glaucoma Quality of Life-15 questionnaire and to examine its psychometric properties. METHODS: The Glaucoma Quality of Life-15 questionnaire was translated and cross-culturally adapted into Moroccan Arabic dialect and administered to glaucoma patients recruited from the ophthalmology department of the Omar Drissi Hospital, Hassan II University Hospital, Fez. Sociodemographic and other clinical data were collected. Psychometric properties were performed, including internal consistency, which was tested using Cronbach's alpha (α), test-retest reliability using intraclass correlation coefficients (ICC). Construct validity was assessed by examining the convergent and divergent validity of items. RESULTS: The questionnaire was administered to 148 patients with a mean age of 60.91 ± 15.10 years. Over half of the patients were female (58.1%), patients were married (77.7%), illiterate (62.2%) and unemployed (82.3%). The majority of patients had primary open angle glaucoma (68.9%). The mean time to complete the GQL-15 was 3.26±0.51min. The mean summary score for the GQL-15 was 39.50±16.76. Cronbach's alpha for the entire scale was 0.95 (central and near vision 0.58; peripheral vision 0.94; glare and dark adaptation 0.87). CONCLUSION: The Moroccan Arabic dialect version of the GQL-15 demonstrates adequate reliability and validity. Therefore, this version could be used as a reliable and valid tool for quality-of-life assessment in Moroccan glaucoma patients.

Investigating a shared-dialect effect between raters and candidates in English speaking tests.

This study set out to examine existence of a shared-dialect effect, a phenomenon that when a rater shares the same dialect with a candidate, the rater is more likely to give the candidate a higher score in English speaking tests. Ten Cantonese-speaking raters and ten Mandarin-speaking raters were selected to assess forty Cantonese-accented and forty Mandarin-accented candidates' oral performance in the retelling task of the Computer-based English Listening and Speaking Test (CELST). Besides, seven raters from each group participated in the stimulated recall stage aiming to reveal their thought process. Quantitative results suggested that the two rater groups were comparable in terms of internal consistency. There were no significant differences in the scores of both candidate groups awarded by both rater groups. The effect of interaction between candidates' dialect and raters' dialect was not statistically significant, indicating non-existence of such effect. Qualitative results showed that some raters attended to candidates' accents, and indicated that awareness of accents and their familiarity with the accents affected their comprehension of the speech samples and potentially influenced their scoring process. The findings are discussed with reference to rater training, rating scale, raters' familiarity with candidates' accents, raters' attitudes toward candidates' accents and the task type. The main implication of this study is that recruiting both group raters in domestic English speaking tests is warranted if the shared-dialect effect could be duly managed.