Metraipolis › archive › dc_f5da0e7756ff

FoSeRL: Formal Sequential Robustness Certification for Reinforcement Learning Policies — arXiv:2610.04754

Source
arXiv
Original
https://arxiv.org/abs/2610.04754
License (as archived)
CC0-1.0
Archived
2026-10-06T05:30:42.924Z
SHA-256 of the archived text
b122b1030e67a69ae2d3046bb1feb9bd261885b1a76f58d98d8e7624683108a4
Stable URL of this page
https://metraipolis.com/documento/dc_f5da0e7756ff

How to cite

FoSeRL: Formal Sequential Robustness Certification for Reinforcement Learning Policies — arXiv:2610.04754 — arXiv, https://arxiv.org/abs/2610.04754 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f5da0e7756ff (Metraipolis archive dc_f5da0e7756ff, sha256 b122b1030e67a69ae2d3046bb1feb9bd261885b1a76f58d98d8e7624683108a4)
BibTeX
@misc{dcf5da0e7756ff,
  title = {FoSeRL: Formal Sequential Robustness Certification for Reinforcement Learning Policies — arXiv:2610.04754},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.04754},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f5da0e7756ff; sha256 b122b1030e67a69ae2d3046bb1feb9bd261885b1a76f58d98d8e7624683108a4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.04754. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Even a few action perturbations can substantially degrade the performance of a deployed decision policy. Certifying the resulting return loss is challenging in stochastic environments, where returns vary even without an attack. We introduce FoSeRL, a framework for certifying deployed RL policies against precommitted, temporally sparse action attacks. The deployed policy is unchanged, with no smoothing or retraining. Certification requires a resettable simulator supporting shared randomness and independent one-step successor queries, but no analytical dynamics model. FoSeRL certifies that an attacked episode loses no more than a prescribed amount of return relative to the same episode unattacked, with at least a target probability and at a user-specified confidence level. Both runs share the initial state and randomness, so the measured loss reflects the attack, not the episode; carrying the running return gap as a state coordinate makes it the terminal value, reducing trajectory-level certification to terminal safety. Time-dependent barrier conditions on the augmented state bound the terminal failure probability: satisfied exactly, they certify every admissible precommitted attack; learned from sampled trajectories and verified on held-out data, they certify the same guarantee under a specified attack-episode setting. Across six stochastic continuous-control environments and three RL policy families (TD3, SAC, and PPO), FoSeRL certifies non-trivial cardinality--magnitude robustness frontiers, achieves substantially larger certified budgets than policy smoothing, and reveals marked robustness differences among policies with comparable nominal performance.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.04754 — 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 2026-10-06T05:27:05.715Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).