publications

2024

1. Book Chapter

   Large Language Models: A technological approach for teaching Hindi

   Vivek Tripathi

   2024

   Abs Bib PDF

   Multiple large language models (LLMs) have been developed in recent years. In the research phase are GPT-3 and GPT-4 from OpenAI, LLaMA from Meta, and PaLM 2 from Google, among many others. From the inception of AI (artificial intelligence) technology, I have been experimenting with the use of AI instruments in the Hindi classroom. I contend for their value by providing three justifications for the use of LLMs in language instruction. The paper begins by investigating the potential use of LLMs in teaching Hindi. It emphasizes the role they can play in vocabulary development, grammar acquisition, and conversational practice. Next, it investigates how these models can help learners improve their proficiency by providing vocabulary-building definitions, synonyms, and example sentences. Thirdly, it shows how the models can provide correct sentence structures, grammatical constructions, and cultural insights, thereby facilitating students’ comprehension of Hindi grammar and cultural nuances. Overall, the study highlights the value of combining language models with conventional learning techniques. It sets out a technological approach that can supplement traditional language teaching methodologies, and thus enhance the learning experience of Hindi language students.

   @book{advancesinhinditeaching,
     title = {Large Language Models: A technological approach for teaching Hindi},
     author = {Tripathi, Vivek},
     pages = {69--114},
     year = {2024},
     url = {https://lincom-shop.eu/epages/57709feb-b889-4707-b2ce-c666fc88085d.sf/de_DE/?ObjectPath=%2FShops%2F57709feb-b889-4707-b2ce-c666fc88085d%2FProducts%2F%22ISBN+9783969391990%22},
     publisher = {LINCOM Studies in Asian Linguistics 100},
   }

2. Journal

   

   Semantic Model for Fragment of Hindi (Part 2)

   Vivek Tripathi, and Dinesh Rathod

   2024

   Abs DOI Bib PDF

   This paper proposes a formal model for semantic analysis of a fragment of the Hindi language. This paper uses referential noun phrases, transitive and intransitive verb phrases and logical constants to compute the meaning of its sentences generated from the Hindi part-of-speech-tagged corpus features. The paper presents cases of conjunction and negation enriched with idempotent laws that provide semantic computation of simple and complex well-formed formulas. Our system works for any model, with one such model described in our glossary. It deals with the set-theoretic study of essential syntactic categories of Hindi, suggesting the suitability of our rule-based syntactic arrangement and model-based semantic computation by implementing them through an in-house software tool.

   @article{Tripathi2024b,
     title = {Semantic Model for Fragment of Hindi (Part 2)},
     author = {Tripathi, Vivek and Rathod, Dinesh},
     year = {2024},
     doi = {10.21659/rupkatha.v16n2.02},
     publisher = {Rupkatha Journal on Interdisciplinary Studies in Humanities},
   }

3. Journal

   

   Semantic Model for Fragment of Hindi (Part 1)

   Vivek Tripathi, and Dinesh Rathod

   2024

   Abs DOI Bib PDF

   This paper proposes a formal model for syntactic and semantic analysis for the Hindi language using context-free grammar. In this paper, we developed a syntactic parser that generates syntactic trees for Hindi sentences based on rules of propositional logic, and gender conventions. The context-free rules we have written follow a top-down approach with a sentence that goes on self-arrangement. A set of experiments were run based on the corpus we have created, and significant results are presented in this paper. In addition to the above, the model characterizes lexical items in terms of individuals and sets for the syntactic distribution for well-formed formulas.

   @article{Tripathi2024a,
     title = {Semantic Model for Fragment of Hindi (Part 1)},
     author = {Tripathi, Vivek and Rathod, Dinesh},
     year = {2024},
     doi = {10.21659/rupkatha.v16n1.03g},
     publisher = {Rupkatha Journal on Interdisciplinary Studies in Humanities},
   }