Enhancing RL Generalizability in Robotics through SHAP Analysis of Algorithms and Hyperparameters
Lingxiao Kong, Cong Yang, Oya Deniz Beyan, Zeyd Boukhers
TLDR
This paper uses SHAP to analyze algorithm and hyperparameter impacts on RL generalization in robotics, enabling better configuration selection.
Key contributions
- Proposes an explainable framework using SHAP to quantify RL configuration impacts on generalization.
- Establishes a theoretical foundation connecting Shapley values to RL generalizability.
- Empirically analyzes configuration impact patterns across diverse RL algorithms and hyperparameters.
- Introduces SHAP-guided configuration selection, improving RL generalizability in robotic tasks.
Why it matters
RL models struggle with generalization across environments due to sensitive configurations. This work provides a systematic, explainable method to understand and leverage configuration impacts using SHAP, leading to more robust and generalizable robotic RL systems. Practitioners can use these insights for better deployment.
Original Abstract
Despite significant advances in Reinforcement Learning (RL), model performance remains highly sensitive to algorithm and hyperparameter configurations, while generalization gaps across environments complicate real-world deployment. Although prior work has studied RL generalization, the relative contribution of specific configurations to the generalization gap has not been quantitatively decomposed and systematically leveraged for configuration selection. To address this limitation, we propose an explainable framework that evaluates RL performance across robotic environments using SHapley Additive exPlanations (SHAP) to quantify configuration impacts. We establish a theoretical foundation connecting Shapley values to generalizability, empirically analyze configuration impact patterns, and introduce SHAP-guided configuration selection to enhance generalization. Our results reveal distinct patterns across algorithms and hyperparameters, with consistent configuration impacts across diverse tasks and environments. By applying these insights to configuration selection, we achieve improved RL generalizability and provide actionable guidance for practitioners.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.