Teaching Assistant Bot (TA BOT) for Sri Lankan Indigenous Medical Education

Our research group is working on a series of projects aimed at enhancing the learning experience for students in various domains. One such project is the development of a Teaching Assistant Bot (TA Bot) that supports Sinhala language prompts, specifically designed for Sri Lankan Indigenous Medical Education.

The TA Bot leverages the power of Large Language Models (LLMs), like OpenAI’s ChatGPT and Google’s Bard/Gemini, which are renowned for their ability to generate human-like text and provide intelligent responses across a wide range of topics. However, these tools often fall short when it comes to providing quality, accurate, and meaningful answers in specific domains and languages. Our project aims to bridge this gap, focusing on the domain of Sri Lankan Indigenous Medicine and the Sinhala language.

The TA Bot is designed to assist students enrolled in the Bachelor of Ayurvedic Medicine and Surgery (BAMS) degree program at the Institute of Indigenous Medicine (IIM), University of Colombo. The program includes a course in Sri Lankan Indigenous Medicine (Deshiya Chikitsa), covering a broad spectrum of topics from the history and evolution of Sri Lankan Indigenous Medicine (SLIM) to traditional therapeutic techniques. The reference materials for these topics are primarily in the Sinhala language.

Despite the potential of LLMs, they currently do not provide accurate responses to questions in these topics. Furthermore, most of the academic and professional practices in Sri Lankan Indigenous Medicine are carried out in the Sinhala language, and the digital presence of knowledge related to these subjects is minimal.

This project proposes to develop a Generative-AI-based TA Bot focused on the 'Deshiya Chikitsa' course of the BAMS degree, by fine-tuning an LLM. This innovative approach aims to enhance the learning experience for students of Sri Lankan Indigenous Medicine, providing them with a valuable tool for understanding complex medical concepts in their native language. This is just one of the many projects our research group is undertaking to push the boundaries of AI in education.