SafeRL-Lite: A Lightweight, Explainable, and Constrained Reinforcement Learning Library

2025-06-17

AI Summary

Existing reinforcement learning (RL) toolkits generally lack native support for hard safety constraints and interpretable decision-making mechanisms. To address this, we propose the first lightweight, open-source library that deeply integrates SHAP values and Grad-CAM saliency maps into the constrained RL training pipeline—specifically within the DQN framework—to jointly optimize safety and interpretability. Our approach employs custom constraint-aware reward shaping and Gym environment wrappers, enabling zero-modification integration with existing RL codebases. It supports real-time attribution of decisions and quantitative violation analysis. Evaluated on a safety-constrained CartPole variant, our method reduces safety violation rates to <0.3%, generates visually verifiable attribution maps, achieves sub-8ms per-step inference latency, and enables one-command deployment via pip. This work bridges critical gaps in trustworthy, production-ready constrained RL.

Technology Category

• Machine Learning: Reinforcement Learning
• Natural Language Processing: Safety and Robustness
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

We introduce SafeRL-Lite, an open-source Python library for building reinforcement learning (RL) agents that are both constrained and explainable. Existing RL toolkits often lack native mechanisms for enforcing hard safety constraints or producing human-interpretable rationales for decisions. SafeRL-Lite provides modular wrappers around standard Gym environments and deep Q-learning agents to enable: (i) safety-aware training via constraint enforcement, and (ii) real-time post-hoc explanation via SHAP values and saliency maps. The library is lightweight, extensible, and installable via pip, and includes built-in metrics for constraint violations. We demonstrate its effectiveness on constrained variants of CartPole and provide visualizations that reveal both policy logic and safety adherence. The full codebase is available at: https://github.com/satyamcser/saferl-lite.

Problem

• Enforcing hard safety constraints in RL agents
• Providing human-interpretable decision rationales in RL
• Lack of lightweight, explainable RL toolkits with safety

Innovation

• Lightweight Python library for constrained RL
• Modular wrappers enable safety-aware training
• Real-time explanations via SHAP and saliency