Research Projects

Manipuri, also known as Meetilon/Meiteinon holds significant cultural history and is one of the 22 officially recognized languages in the Eight Schedule of India’s constitution [1]. Despite its official status, Manipuri remains a low-resource language [2] with limited digital resources and linguistic datasets, hindering the development of Natural language Processing (NLP) tools like machine translation, sentiment analysis, and intelligent chatbots. The proposed project seeks to bridge this gap by developing resources and tools to support the digital preservation of Manipuri language and its integration into emerging technologies such as Large Language Models (LLMs) and Retrieval Augmented Generation (RAG). The scarcity of data in the Manipuri language would be addressed by employing various data mining techniques such as web scraping, digitization of storybooks, legal documents, newspapers, social media content extraction, and video/audio content extraction through Audio Speech Recognition (ASR). This would be followed by (i) the creation of a Manipuri-Romanji script bi-directional transliteration tool; and (ii) the creation of Manipuri-English parallel corpora. The dataset would be made available online for public use for community-based development. The dataset would be used to fine-tune pre-trained LLMs to understand the linguistic structures and build NLP applications such as machine translation, sentiment analyzer, Part-Of-Speech (POS) tagger and smart chatbots. The outputs will enable practical applications, including an interactive English-Manipuri dictionary, a language learning platform, and an online English-Manipuri translator. This project’s outcomes extend beyond academia; they align with India’s policy initiatives for promoting the Indian Knowledge System (IKS), potentially informing preservation strategies for other low-resource and endangered Indian languages. The anticipated results include a valuable digital collection of Manipuri linguistic assets, fostering community engagement, cultural preservation, and support for future NLP applications. This project, by demonstrating viable strategies for the digital preservation of Manipuri, could provide a model for preserving other low-resource tribal languages in India that lack substantial digital resources or even written scripts.