Learning with AMIGo: Adversarially Motivated Intrinsic Goals
This addresses the problem of sparse rewards in reinforcement learning for AI agents, offering a novel approach to intrinsic motivation that could improve learning efficiency in complex environments.
The paper tackles the challenge of reinforcement learning in environments with sparse extrinsic rewards by proposing AMIGo, an agent that uses a teacher to generate adversarially motivated intrinsic goals, enabling a student policy to learn general skills and solve challenging procedurally-generated tasks where other methods fail.
A key challenge for reinforcement learning (RL) consists of learning in environments with sparse extrinsic rewards. In contrast to current RL methods, humans are able to learn new skills with little or no reward by using various forms of intrinsic motivation. We propose AMIGo, a novel agent incorporating -- as form of meta-learning -- a goal-generating teacher that proposes Adversarially Motivated Intrinsic Goals to train a goal-conditioned "student" policy in the absence of (or alongside) environment reward. Specifically, through a simple but effective "constructively adversarial" objective, the teacher learns to propose increasingly challenging -- yet achievable -- goals that allow the student to learn general skills for acting in a new environment, independent of the task to be solved. We show that our method generates a natural curriculum of self-proposed goals which ultimately allows the agent to solve challenging procedurally-generated tasks where other forms of intrinsic motivation and state-of-the-art RL methods fail.