English

On Hard Exploration for Reinforcement Learning: a Case Study in Pommerman

Machine Learning 2019-07-30 v1 Artificial Intelligence Machine Learning

Abstract

How to best explore in domains with sparse, delayed, and deceptive rewards is an important open problem for reinforcement learning (RL). This paper considers one such domain, the recently-proposed multi-agent benchmark of Pommerman. This domain is very challenging for RL --- past work has shown that model-free RL algorithms fail to achieve significant learning without artificially reducing the environment's complexity. In this paper, we illuminate reasons behind this failure by providing a thorough analysis on the hardness of random exploration in Pommerman. While model-free random exploration is typically futile, we develop a model-based automatic reasoning module that can be used for safer exploration by pruning actions that will surely lead the agent to death. We empirically demonstrate that this module can significantly improve learning.

Keywords

Cite

@article{arxiv.1907.11788,
  title  = {On Hard Exploration for Reinforcement Learning: a Case Study in Pommerman},
  author = {Chao Gao and Bilal Kartal and Pablo Hernandez-Leal and Matthew E. Taylor},
  journal= {arXiv preprint arXiv:1907.11788},
  year   = {2019}
}

Comments

AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE) 2019

R2 v1 2026-06-23T10:32:25.548Z