Proceedings

Don’t Do What Doesn’t Matter: Intrinsic Motivation with Action Usefulness

Mathieu Seurin, Florian Strub, Philippe Preux, Olivier Pietquin

Download BibTeX Cite

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. Main Track. Pages 2950-2956.

https://doi.org/10.24963/ijcai.2021/406

Keywords

  • Machine Learning: Deep Reinforcement Learning
  • Machine Learning: Reinforcement Learning

Abstract

Sparse rewards are double-edged training signals in reinforcement learning:

easy to design but hard to optimize. Intrinsic motivation guidances have thus been developed toward alleviating the resulting exploration problem. They usually incentivize agents to look for new states through novelty signals. Yet, such methods encourage exhaustive exploration of the state space rather than focusing on the environment's salient interaction opportunities. We propose a new exploration method, called Don't Do What Doesn't Matter (DoWhaM), shifting the emphasis from state novelty to state with relevant actions. While most actions consistently change the state when used, e.g. moving the agent, some actions are only effective in specific states, e.g., opening a door, grabbing an object. DoWhaM detects and rewards actions that seldom affect the environment.

We evaluate DoWhaM on the procedurally-generated environment MiniGrid against state-of-the-art methods. Experiments consistently show that DoWhaM greatly reduces sample complexity, installing the new state-of-the-art in MiniGrid.

Mathieu Seurin, Florian Strub, Philippe Preux, and Olivier Pietquin. 2021. Don’t Do What Doesn’t Matter: Intrinsic Motivation with Action Usefulness. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21. IJCAI, 2950–2956. https://doi.org/10.24963/ijcai.2021/406
Welcome Future Conferences AI Journal Division IJCAI Proceedings About IJCAI IJCAI Awards