Enabling Personalized Learning in Large Language Models via Reinforcement Learning

Hamsa Bastani, Operations, Information and Decisions, The Wharton School; Angel Chung, PhD Candidate, The Wharton School; Ling-Chieh Kung, National Taiwan University; Osbert Bastani, Penn Engineering

Abstract: This project combines GPT-4 and reinforcement learning to develop a personalized learning platform for an introductory Python programming class. In collaboration with the National Taiwan University, this project will deploy and conduct a randomized control trial (RCT) to understand how to effectively use generative AI to improve student’s learning outcomes and skill development.