Automatic Discovery of Subgoals in Reinforcement Learning using Diverse Density
Amy McGovern, Andrew G. Barto · ScholarWorks@UMassAmherst (University of Massachusetts Amherst) · 2001
This paper presents a method by which a rein-forcement learning agent can automatically dis-cover certain types of subgoals online. By creat-ing useful new subgoals while learning, the agent is able to accelerate learning on the current task and to transfer its expertise to other, related tasks through the reuse of its ability to attain subgoals. The agent discovers subgoals based on common-alities across multiple paths to a solution. We cast the task of finding these commonalities as a multiple-instance learning problem and use the concept of diverse density to find solutions. We illustrate this approach using several gridworld tasks. 1.