Adaptive Probabilistic Shielding by Learning MDPs for Safe Reinforcement Learning
cs.LG, cs.AI, cs.LO
Submitted: 2026-08-20
Updated: 2026-08-20
Comments: 19 pages, 3 figures, 3 tables. To be published in the proceedings of RV 2026
Journal ref: RV 2026
DOI: 10.1007/978-3-032-41160-0_4
License: http://creativecommons.org/licenses/by/4.0/
The gist: Probabilistic shielding is a technique for safe reinforcement learning (RL).
Terminology
Abstract
Probabilistic shielding is a technique for safe reinforcement learning (RL). Typically, a static observer -- called the shield -- constrains the learning agent's actions to those for which acting safely remains feasible. Traditionally, the shield is computed from the transition probabilities of the underlying Markov decision process (MDP). Thus, this technique is not applicable when the MDP model is not given a priori, which, unfortunately, is the case in typical RL applications. In this paper, we study the problem of computing a shield in the setting where the transition graph of the MDP is known, but the transition probabilities are unknown. Our approach integrates probabilistic shielding with online model learning: as the RL agent explores the environment, we estimate the transition probabilities. From this estimate, we compute a shield. While the shield may be conservative initially, it adapts as the model estimate becomes more precise. Thus, the shield improves in tandem with the RL agent. This paradigm of adaptive probabilistic shielding raises a number of challenges, such as when to recompute the shield and how to balance between exploration and safety during learning. We empirically evaluate multiple variants of this paradigm across several environments.
Sources
- Compositional shield synthesis for safe reinforcement learning in partial observability
- Robust Probabilistic Shielding for Safe Offline Reinforcement Learning
- Robust Shielding for Safe Reinforcement Learning
- Shields to Guarantee Probabilistic Safety in MDPs
- Optimization-Based Robust Permissive Synthesis for Interval MDPs
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis