Long-Term Exploration in Persistent MDPs

TL;DR

This paper introduces RbExplore, a novel exploration method for reinforcement learning in large, sparse-reward environments, leveraging persistent MDPs to improve exploration efficiency without domain knowledge.

Contribution

The paper proposes RbExplore, a new exploration algorithm utilizing rollback in persistent MDPs, outperforming existing curiosity-driven methods in hard-exploration games.

Findings

RbExplore outperforms or matches state-of-the-art methods in Prince of Persia.

The method is effective without rewards or domain knowledge.

It demonstrates improved exploration in large, sparse environments.

Abstract

Exploration is an essential part of reinforcement learning, which restricts the quality of learned policy. Hard-exploration environments are defined by huge state space and sparse rewards. In such conditions, an exhaustive exploration of the environment is often impossible, and the successful training of an agent requires a lot of interaction steps. In this paper, we propose an exploration method called Rollback-Explore (RbExplore), which utilizes the concept of the persistent Markov decision process, in which agents during training can roll back to visited states. We test our algorithm in the hard-exploration Prince of Persia game, without rewards and domain knowledge. At all used levels of the game, our agent outperforms or shows comparable results with state-of-the-art curiosity methods with knowledge-based intrinsic motivation: ICM and RND. An implementation of RbExplore can be found at https://github.com/cds-mipt/RbExplore.