Ranking Policy Decisions

Abstract

Policies trained via Reinforcement Learning (RL) are often needlessly complex, making them difficult to analyse and interpret. In a run with \(n\) time steps, a policy will make \(n\) decisions on actions to take; we conjecture that only a small subset of these decisions delivers value over selecting a simple default action. Given a trained policy, we propose a novel black-box method based on statistical fault localisation that ranks the states of the environment according to the importance of decisions made in those states. We argue that among other things, the ranked list of states can help explain and understand the policy. As the ranking method is statistical, a direct evaluation of its quality is hard. As a proxy for quality, we use the ranking to create new, simpler policies from the original ones by pruning decisions identified as unimportant (that is, replacing them by default actions) and measuring the impact on performance. Our experiments on a diverse set of standard benchma

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keypouget2020ranking

Related papers

• Clustered Policy Decision Ranking (2023)0.00
• Test Where Decisions Matter: Importance-driven Testing For Deep Reinforcement Learning (2024)0.00
• General Policy Evaluation And Improvement By Learning To Identify Few But Crucial States (2022)0.00
• "so, Tell Me About Your Policy...": Distillation Of Interpretable Policies From Deep Reinforcement Learning Agents (2025)0.00
• What Matters In On-policy Reinforcement Learning? A Large-scale Empirical Study (2020)0.00
• Policy Agnostic RL: Offline RL And Online RL Fine-tuning Of Any Class And Backbone (2024)0.00
• Improving Sample Efficiency In Evolutionary RL Using Off-policy Ranking (2022)0.00
• Which Rewards Matter? Reward Selection For Reinforcement Learning Under Limited Feedback (2025)0.00