Metraipolis › archive › dc_7991c8458074

CORMA: Constraint-Routed Decision Mamba for Generalizable Multi-Constraint Offline Safe Reinforcement Learning — arXiv:2610.34208

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.34208
License (as archived)
CC0-1.0
Archived
2026-10-27T08:01:13.329Z
SHA-256 of the archived text
20c3311444c93869de77be7679e4b2e00703c6b225b161428628ee469d4421b7
Stable URL of this page
https://metraipolis.com/documento/dc_7991c8458074

How to cite

CORMA: Constraint-Routed Decision Mamba for Generalizable Multi-Constraint Offline Safe Reinforcement Learning — arXiv:2610.34208 — arXiv (en), https://arxiv.org/abs/2610.34208 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7991c8458074 (Metraipolis archive dc_7991c8458074, sha256 20c3311444c93869de77be7679e4b2e00703c6b225b161428628ee469d4421b7)
BibTeX
@misc{dc7991c8458074,
  title = {CORMA: Constraint-Routed Decision Mamba for Generalizable Multi-Constraint Offline Safe Reinforcement Learning — arXiv:2610.34208},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.34208},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7991c8458074; sha256 20c3311444c93869de77be7679e4b2e00703c6b225b161428628ee469d4421b7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.34208. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Offline safe reinforcement learning (RL) learns safety-aware policies from static datasets while satisfying safety constraints during deployment. Motivated by the strong trajectorymodeling capabilities of Transformers, recent Decision Transformer (DT)-based methods formulate offline safe RL as conditional sequence modeling by conditioning action generation on return-to-go (RTG) and cost-to-go (CTG) tokens, thereby enabling adaptation to varying safety budgets. However, existing DT-based methods are developed for single-constraint tasks. For tasks with multiple constraints, generalizing learned policies to unseen combinations of safety budgets remains challenging due to the conflicting requirements imposed by heterogeneous constraints. Moreover, the quadratic complexity of Transformer architectures significantly increases the computational overhead in multi-constraint tasks, as representing additional constraints requires more CTG tokens. Recently, Mamba has emerged as an efficient sequence-modeling backbone, offering linear computational complexity and strong generalization capabilities. In this paper, we propose CORMA, a constraint-routed decision Mamba framework for generalizable multi-constraint offline safe RL. CORMA decomposes policy inference into constraint-aware RTG estimation and adaptive action generation, where constraint-specific experts model feasible RTG distributions and a router-guided product-of-experts mechanism fuses these distributions to provide safety-aware guidance for a Mamba-based action decoder. A three-stage optimization strategy is designed to improve training stability. Extensive experiments on OSRL benchmarks demonstrate that CORMA achieves superior reward-safety trade-offs, reduces inference latency by 37% compared with DT-based methods, and exhibits stronger zero-shot generalization to unseen safety-budget combinations than offline safe RL baselines.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.34208 — 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 2026-10-27T07:55:30.883Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).