Summary of Explaining Reinforcement Learning: a Counterfactual Shapley Values Approach, by Yiwei Shi et al.
Explaining Reinforcement Learning: A Counterfactual Shapley Values Approach
by Yiwei Shi, Qi Zhang, Kevin McAreavey, Weiru Liu
First submitted to arxiv on: 5 Aug 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 paper proposes a novel approach called Counterfactual Shapley Values (CSV) to enhance explainability in reinforcement learning (RL). The CSV method integrates counterfactual analysis with Shapley Values to quantify and compare the contributions of different state dimensions to various action choices. To achieve this, new characteristic value functions are introduced, including the “Counterfactual Difference Characteristic Value” and the “Average Counterfactual Difference Characteristic Value.” These functions help calculate Shapley values to evaluate the differences in contributions between optimal and non-optimal actions. The paper demonstrates the effectiveness of CSV across several RL domains, including GridWorld, FrozenLake, and Taxi, showing that it not only improves transparency but also quantifies differences across decisions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way to understand how things work together in computer games and decision-making systems. It’s called Counterfactual Shapley Values (CSV), and it helps us see what makes certain choices better than others. The idea is to figure out which parts of the situation make a difference, like if taking a different action would have led to a better outcome. They tested this new approach on several games, including GridWorld, FrozenLake, and Taxi, and showed that it can help us understand complex systems better. |
Keywords
» Artificial intelligence » Reinforcement learning
