Metraipolis › archive › dc_cd02fad4d140

A Universal Policy Wrapper With Guarantees — arXiv:2611.02888

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.02888
License (as archived)
CC0-1.0
Archived
2026-11-03T18:38:45.555Z
SHA-256 of the archived text
a8c0abdbab6c9b164f571fffd971138c849d5385a4fce2c3cc78717dab4d6f6b
Stable URL of this page
https://metraipolis.com/documento/dc_cd02fad4d140

How to cite

A Universal Policy Wrapper With Guarantees — arXiv:2611.02888 — arXiv (en), https://arxiv.org/abs/2611.02888 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cd02fad4d140 (Metraipolis archive dc_cd02fad4d140, sha256 a8c0abdbab6c9b164f571fffd971138c849d5385a4fce2c3cc78717dab4d6f6b)
BibTeX
@misc{dccd02fad4d140,
  title = {A Universal Policy Wrapper With Guarantees — arXiv:2611.02888},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02888},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cd02fad4d140; sha256 a8c0abdbab6c9b164f571fffd971138c849d5385a4fce2c3cc78717dab4d6f6b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02888. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Many control tasks require driving the system into a prescribed goal set and subsequently keeping it there. The considered setting assumes that a working controller is available that provably reaches and maintains the goal, albeit with suboptimal performance. This controller is referred to as the fallback policy. Conversely, although state-of-the-art reinforcement-learning (RL) algorithms can discover policies with impressive empirical performance, the neural networks at their core act as opaque black boxes learned through data-driven exploration and, as such, provide no theoretical mechanism to guarantee that the system will ever reach---and stay within---the prescribed goal set. This article presents CALF-Wrapper, a method that combines the fallback policy with a pretrained RL base policy and uses the critic estimate associated with the base policy to select between them during deployment. The theoretical analysis establishes that CALF-Wrapper inherits the goal-reaching property of the fallback policy under the stated assumptions and provides probability bounds for the settling time and the maximum distance from the goal set over a prescribed time interval. The experimental study evaluates CALF-Wrapper in four simulated control environments of different complexity. Comparisons include the base policy, fallback policy, and constrained reinforcement-learning methods formulated for the same objective. Sensitivity and ablation studies provide direct hyperparameter recommendations. The experiments show how CALF-Wrapper can provide effective task performance while attaining the maximum possible observed goal-reaching rate. CALF-Wrapper requires no additional dynamics model, online learning, or constrained optimization during execution.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Systems and Control (eess.SY); Dynamical Systems (math.DS).
Pagina del lavoro: https://arxiv.org/abs/2611.02888 — 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 2026-11-03T18:12:54.599Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).