cs.AINov 6, 2026

Safe Reinforcement Learning via Probabilistic Option Shielding

Authors: Gian Luca Ascari, Isidoro Tamassia, Ying Jiao, Luc De Raedt, Giuseppe Marra

Organizations: Apixa · KU Leuven · KU Leuven, Örebro University

Abstract

Shielding is a safe reinforcement learning technique that provides safety guarantees while learning, by blocking unsafe actions. Traditional shielding relies on a complete model of the environment to provide long-term guarantees. Probabilistic logic shielding (PLS) lifts the requirement of having such model, but only considers safety in the short term, typically looking only one or few steps ahead. We propose Probabilistic Option Shielding (POS), which extends PLS to the option level, recovering long-horizon safety without reintroducing the need for a complete environment model. POS learns a model-free critic directly from one-step PLS signals, to estimate the probability that an entire option will complete safely under uncertainty. We evaluate POS on safe navigation tasks in the LogiCity environment, and show that it achieves considerably safer training while matching the task performance of standard step-level PLS and other safety related reinforcement learning baselines.

Figures & tables

Appendix figures & tables13 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Adaptive Probabilistic Shielding by Learning MDPs for Safe Reinforcement Learning

    Aug 20, 2026Astrid Horn Brorholt, Maris F. L. Galesloot, Nils Jansen +2Safe RLRL Shielding

  2. Robust Probabilistic Shielding for Safe Offline Reinforcement Learning

    May 11, 2026Maris F. L. Galesloot, Thomas Rhemrev, Nils JansenOffline RLSafe RL

  3. Robust Shielding for Safe Reinforcement Learning

    May 29, 2026Edwin Hamel-De le Court, Thom Badings, Alessandro Abate +2Robust Markov Decision ProcessesFormal Verification