Metraipolis › archive › dc_bfc480f47ed3
Safe Reinforcement Learning via Probabilistic Option Shielding — arXiv:2611.09197
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.09197
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-09T22:33:55.504Z
- SHA-256 of the archived text
5f77f1e965e3ca6527c8e28ffda21cde5dd1dce74c3750e2eb84de0b817bae11- Stable URL of this page
https://metraipolis.com/documento/dc_bfc480f47ed3
How to cite
Safe Reinforcement Learning via Probabilistic Option Shielding — arXiv:2611.09197 — arXiv (en), https://arxiv.org/abs/2611.09197 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bfc480f47ed3 (Metraipolis archive dc_bfc480f47ed3, sha256 5f77f1e965e3ca6527c8e28ffda21cde5dd1dce74c3750e2eb84de0b817bae11)
BibTeX
@misc{dcbfc480f47ed3,
title = {Safe Reinforcement Learning via Probabilistic Option Shielding — arXiv:2611.09197},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.09197},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bfc480f47ed3; sha256 5f77f1e965e3ca6527c8e28ffda21cde5dd1dce74c3750e2eb84de0b817bae11},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.09197. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.09197 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-09T22:30:47.267Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).