Use of large language modeling tools to mine literature for pollution sources

Variety of LLMs are being widely used in various domains to scan existing literature/documents to answer meaningful questions. In this work, we propose to use existing LLMs to generate of a database of pollution sources in the Indian subcontinent as reported in literature. The work will involve transfer learning to retrain existing LLMs to be able to solve the problem at hand. The database, once developed, will be of immense benefit to researchers working in the area of pollution source apportionment. 

Scope of the work: Computational

Prerequisite: Interest and ability to pick up machine learning tools in the area of natural language processing and image processing

Name of Faculty