Summary of Bet: Explaining Deep Reinforcement Learning Through the Error-prone Decisions, by Xiao Liu et al.
BET: Explaining Deep Reinforcement Learning through The Error-Prone Decisions
by Xiao Liu, Jie Zhao, Wubing Chen, Mao Tan, Yongxing Su
First submitted to arxiv on: 14 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research proposes a novel approach called Backbone Extract Tree (BET) to improve the interpretability of Deep Reinforcement Learning (DRL) agents in safety-sensitive domains. BET identifies error-prone states by analyzing uniform decision-making patterns and expressing them within neighborhoods defined by representative states. The method is evaluated in various popular RL environments, showing superiority over existing self-interpretable models in terms of explanation fidelity. A use case is demonstrated for providing explanations for agents in StarCraft II, a complex multi-agent cooperative game. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary BET helps us understand how DRL agents make decisions by finding patterns where they consistently make the same choices. This can help us identify situations where an agent might make mistakes. The researchers tested BET in different scenarios and showed that it works better than other methods for explaining why an agent made a certain decision. |
Keywords
* Artificial intelligence * Reinforcement learning
