CARL: Chat Assisted Reinforcement Learning

Sanjiv Das, Harshad Khadilkar, Sukrit Mittal +2

Failure of applied Reinforcement Learning (RL) to solve a given problem predominantly stems from unaddressed specification gaps, where the intent of human experts in the problem's…