MITRA - Machine Translation and Language Analysis for Premodern Asian Low Resource Languages

Watch a research talk exploring the Dharmamitra project's innovative approach to developing language models for classical Asian low-resource languages including Sanskrit, Tibetan, Classical Chinese, and Pali. Learn how these models address both basic NLP tasks like word segmentation and advanced applications such as semantic search and machine translation. Discover the unique challenges of working with premodern language data and the strategies employed to overcome them. Explore a comprehensive application that combines various tools to provide detailed grammatical explanations and corpus-wide search capabilities, serving both beginning language learners and experienced researchers. Presented by Sebastian Nehrdich, a research assistant at UC Berkeley leading the MITRA project and developer of the BuddhaNexus platform, this talk demonstrates how modern NLP techniques can enhance the study and accessibility of classical Asian texts.