Adaptive Probabilistic Shielding by Learning MDPs for Safe Reinforcement Learning
On the map: Deep reinforcement learning methods → Safe and constrained RL
Abstract (arXiv)
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.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Safe and constrained RL, one of 15 directions in Deep reinforcement learning methods. Builds on 20 of its 40 references that are in the map. First posted on arXiv 20 Aug 2026.
Builds on
20 of 40 references.
Closest ancestors · 6
Same direction, newest first
- 2026Robust Shielding for Safe Reinforcement Learning29 May
- 2025Compositional shield synthesis for safe reinforcement learning in partial observability15 Sep
- 2025Probabilistic Shielding for Safe Reinforcement Learning9 Mar
- 2024Learning-Based Shielding for Safe Autonomy under Unknown Dynamics7 Oct
- 2024Realizable Continuous-Space Shields for Safe Reinforcement Learning2 Oct
- 2024Safe Reinforcement Learning in Black-Box Environments via Adaptive Shielding28 May
Other references · 14
- 2017Safe Reinforcement Learning via Shieldingcited by 449
- 2021Safe Multi-Agent Reinforcement Learning via Shieldingcited by 61
- 2019PAC Statistical Model Checking for Markov Decision Processes and Stochastic Gamescited by 32
- 2020Adaptive Shielding under Uncertaintycited by 18
- 2024Robust Markov Decision Processes: A Place Where AI and Formal Methods Meetcited by 18
- 2022Robust Anytime Learning of Markov Decision Processescited by 16
- 2022Automata Learning meets Shieldingcited by 9
- 2023Shielded Reinforcement Learning for Hybrid Systemscited by 8
- 2024Compositional Shielding and Reinforcement Learning for Multi-Agent Systemscited by 8
- 2023Approximate Model-Based Shielding for Safe Reinforcement Learningcited by 5
- 2025Adaptive Shielding via Parametric Safety Proofscited by 5
- 2020Approximating Euclidean by Imprecise Markov Decision Processescited by 4
2 more
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2025ProSh: Probabilistic Shielding for Model-free Reinforcement Learning
- 2025Adaptive Shielding for Safe Reinforcement Learning under Hidden-Parameter Dynamics Shifts
- 2024Leveraging Approximate Model-based Shielding for Probabilistic Safety Guarantees in Continuous Environments
- 2024Think Smart, Act SMARL! Analyzing Probabilistic Logic Shields for Multi-Agent Reinforcement Learning
- 2023Model-based Dynamic Shielding for Safe and Efficient Multi-Agent Reinforcement Learning
- 2026Safe Reinforcement Learning via Recovery-based Shielding with Gaussian Process Dynamics Models
- 2025Provably Optimal Reinforcement Learning under Safety Filtering
- 2026Easy-to-Use Shielding for Reinforcement Learning