Generative Adversarial Self-Imitation Learning

TL;DR

GASIL introduces a regularizer for reinforcement learning that encourages agents to imitate successful past trajectories using a generative adversarial framework, improving performance in environments with sparse and delayed rewards.

Contribution

It proposes GASIL, a novel method combining imitation learning with adversarial training to enhance reinforcement learning in challenging environments.

Findings

GASIL improves PPO performance in 2D Point Mass environments.

GASIL enhances learning in MuJoCo tasks with delayed rewards.

The method effectively handles stochastic dynamics.

Abstract

This paper explores a simple regularizer for reinforcement learning by proposing Generative Adversarial Self-Imitation Learning (GASIL), which encourages the agent to imitate past good trajectories via generative adversarial imitation learning framework. Instead of directly maximizing rewards, GASIL focuses on reproducing past good trajectories, which can potentially make long-term credit assignment easier when rewards are sparse and delayed. GASIL can be easily combined with any policy gradient objective by using GASIL as a learned shaped reward function. Our experimental results show that GASIL improves the performance of proximal policy optimization on 2D Point Mass and MuJoCo environments with delayed reward and stochastic dynamics.