Search

With the increasing practice of using regional languages in social media platforms, hate speech detection in regional languages has received the attention of researchers. In India, hundreds of languages are spoken in various forms, which are dependent on their geography, culture, etc. Recently the number of active internet users has been rapidly increasing in India, and therefore social media has penetrated the common Indian population. Though the need for proper detection and timely removal of abusive or offensive texts has increased, well-organized and labeled data for Indian languages are scarce. Almost all the regional languages in India are low-resource languages. Hence, the objective of this study is to develop an approach that will learn from relatively small volumes of Indian language data and provide state-of-the-art results. A fusion of features extracted from a fined-tuned multilingual BERT (Bidirectional Encoder Representations from Transformers) and a fine-tuned Indic BERT has been proposed in this study. Since the BERT models that we have used for this work are pre-trained using a large volume of texts in multiple Indian languages, transfer learning solves the problem of low training data volume, and this makes the proposed model more generic. Three datasets for three different Indian languages namely, Bengali, Marathi, and Hindi have been considered in this study to evaluate the proposed approach. The proposed model achieved a weighted F1 score of 0.923, 0.815, and 0.924 for the Bengali, Hindi, and Marathi datasets respectively. In the Bengali and Marathi datasets, the obtained results are better than the existing best results. [ABSTRACT FROM AUTHOR]

Introduction: Various tools for measuring health literacy are designed to assess reading comprehension and numeracy in English speakers. There is a need to develop a tool in the vernacular language and estimate health literacy levels in Indian settings. The present study was conducted with the objectives to develop a Marathi version of a 14-item health literacy scale (HLS-14) to test the reliability and validity of its Marathi version and to estimate the health literacy among patients attending the out-patient department at a tertiary care centre. Methodology: The present study was conducted among 50 adult patients attending the out-patient department of a tertiary hospital from July 2022 to December 2022. The 14-Item Health Literacy Scale available in English was translated into Marathi and back-translated to English, and the final version was developed. Bilingual study subjects were asked to fill the scales on day 0 and on day 7. Cronbach’s alpha was calculated for internal validity, and the correlation coefficient was calculated for the reliability of the tool and health literacy was estimated. Results: When items of the Health Literacy Scale were analysed, all the items barring 2, 6, and 10 gave an r-value of more than 0.70, which shows good reliability of each translated item. The Cronbach’s alpha value found for the current translated Marathi questionnaire is 0.66. Internal consistency is good. The mean total health literacy score was 51.16 ± 6.81. Conclusions: A translated Marathi version of HLS-14 is developed, which is valid and reliable. The health literacy among the study participants is marginal.

Code-switching or code-mixing in spoken language is quite common. Can code-mixing also occur in a written form of a language, such as in a newspaper article, in a short story, or in a translation? This paper examines the use of code-mixing in the translation of a German short story into Marathi. Marathi is an Indo-European language that is recognized as one of the 22 scheduled languages of India. The article analyses whether code-mixing is used in the Marathi translation of the German short story when no other equivalents are available in the target language i.e., in Marathi or is it more of a thoughtful choice even if equivalents exist in the target language. With relevant examples, this empirical analysis further investigates the role of code-mixing in the Marathi translation of the German short story, “Der hellgraue Frühjahrsmantel” written by Wolfgang Hildesheimer (1962). The title of the translated text reads “राखाडी स्प्रिंगजाकीट” (rākhādī springjākīṭ, 2017). With this analysis, the article presents some remarkable characteristics of both the Indian language Marathi, its linguistic context and culture. The analysis can stimulate interest in Marathi and could also be useful in further research in the areas of code mixing or code switching in translation.

The advent of Multilingual Language Models (MLLMs) and Large Language Models (LLMs) has spawned innovation in many areas of natural language processing. Despite the exciting potential of this technology, its impact on developing high-quality Machine Translation (MT) outputs for low-resource languages remains relatively under-explored. Furthermore, an open-source application, dedicated to both fine-tuning MLLMs and managing the complete MT workflow for low-resources languages, remains unavailable. We aim to address these imbalances through the development of adaptMLLM, which streamlines all processes involved in the fine-tuning of MLLMs for MT. This open-source application is tailored for developers, translators, and users who are engaged in MT. It is particularly useful for newcomers to the field, as it significantly streamlines the configuration of the development environment. An intuitive interface allows for easy customisation of hyperparameters, and the application offers a range of metrics for model evaluation and the capability to deploy models as a translation service directly within the application. As a multilingual tool, we used adaptMLLM to fine-tune models for two low-resource language pairs: English to Irish (EN ↔ GA) and English to Marathi (EN ↔ MR). Compared with baselines from the LoResMT2021 Shared Task, the adaptMLLM system demonstrated significant improvements. In the EN → GA direction, an improvement of 5.2 BLEU points was observed and an increase of 40.5 BLEU points was recorded in the GA → EN direction representing relative improvements of 14% and 117%, respectively. Significant improvements in the translation performance of the EN ↔ MR pair were also observed notably in the MR → EN direction with an increase of 21.3 BLEU points which corresponds to a relative improvement of 68%. Finally, a fine-grained human evaluation of the MLLM output on the EN → GA pair was conducted using the Multidimensional Quality Metrics and Scalar Quality Metrics error taxonomies. The application and models are freely available. [ABSTRACT FROM AUTHOR]

Copyright of Linguistische Treffen in Wrocław is the property of Oficyna Wydawnicza ATUT and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Critiques of vulgarity, sexuality and class featured prominently in the social, cultural and intellectual life of post-Independence India. This article focuses on the discourse of chumban bandi (banning kissing) in the 1950s and 1960s Maharashtra and analyses how it became a particularly unique index of heightened transgressive pleasure. I situate this discourse within a larger public debate, where dominant caste middle-class elites took upon the responsibility to shepherd supposedly recalcitrant dominated castes and low-class masses towards decency, civilized action and citizenship. Many elites energetically worked on their ideology, which was rooted in high-caste, middle-class and patriarchal values to create Marathi manus and nation. Drawing upon hitherto neglected Marathi language texts, I show how elites policed the kiss to both ban on-screen kissing and paradoxically harness its energy to engage in the politics of Marathikaran (creating Marathi regional identity) and create a new Marathi identity as modern, moral and decent. [ABSTRACT FROM AUTHOR]