Large language model in bio-medical domain

Abstract

The rapid advancement of biomedical research has led to an overwhelming influx of scholarly articles, making efficient retrieval and comprehension of domain-specific knowledge a significant challenge. This project presents a novel implementation of a Retrieval-Augmented Generation (RAG)-based chatbot designed specifically for the biomedical domain. Unlike existing systems, this chatbot dynamically processes user-provided keywords in real-time, fetches relevant research papers from sources such as ArXiv, and generates contextually accurate explanations using domain-specific models like BioBERT. The chatbot eliminates the need for pre-uploading documents to an LLM interface by leveraging advanced retrieval techniques and transformer-based natural language models. Instead, it autonomously retrieves, processes, and summarises relevant information, significantly reducing time and effort. This innovation holds immense significance for researchers, clinicians, and educators by providing a seamless interface for accessing and understanding biomedical literature across various subfields.