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

arXiv: https://arxiv.org/abs/1907.11788 · HTML: https://mlsanity.com/paper/1907.11788/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 26 Jul 2019.

## Built on by (most cited first, 4 of 4)

- [Exploration in Deep Reinforcement Learning: A Survey](https://mlsanity.com/paper/2205.00824/) (2022)
- [Multi-Agent Advisor Q-Learning](https://mlsanity.com/paper/2111.00345/) (2021)
- [Multi-Agent Training for Pommerman: Curriculum Learning and Population-based Self-Play Approach](https://mlsanity.com/paper/2407.00662/) (2024)
- [Know your Enemy: Investigating Monte-Carlo Tree Search with Opponent Models in Pommerman](https://mlsanity.com/paper/2305.13206/) (2023)

## Abstract (arXiv)

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.