Metraipolis › archive › dc_7ff556259cc1
Learning Safe Flow Matching Robot Policies from Binary Feasibility Oracles — arXiv:2610.40075
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.40075
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T06:14:45.716Z
- SHA-256 of the archived text
e287aba9a70827c27816adf327b958b443d33642282584bfef763e72d0b543f2- Stable URL of this page
https://metraipolis.com/documento/dc_7ff556259cc1
How to cite
Learning Safe Flow Matching Robot Policies from Binary Feasibility Oracles — arXiv:2610.40075 — arXiv (en), https://arxiv.org/abs/2610.40075 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7ff556259cc1 (Metraipolis archive dc_7ff556259cc1, sha256 e287aba9a70827c27816adf327b958b443d33642282584bfef763e72d0b543f2)
BibTeX
@misc{dc7ff556259cc1,
title = {Learning Safe Flow Matching Robot Policies from Binary Feasibility Oracles — arXiv:2610.40075},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.40075},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7ff556259cc1; sha256 e287aba9a70827c27816adf327b958b443d33642282584bfef763e72d0b543f2},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.40075. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Robot foundation models increasingly rely on flow-based action generation for general-purpose manipulation, yet the generated actions are not guaranteed to satisfy physical constraints such as joint limits and collision avoidance. This problem is especially challenging when the feasible set cannot be described analytically and is accessible only through black-box feasibility oracles. We study this practical oracle-only setting and propose an oracle-guided differentiable safeguard for safe flow-matching robot policies. The safeguard uses binary feasibility queries to correct unsafe updates in the forward pass, while extracting local boundary information to provide constraint-aware learning signals in the backward pass. Combined with discrete-time flow matching, the same safeguard is used during training and inference, enabling forward--backward aligned policy learning without explicit or differentiable constraint models. Experiments on obstacle-aware robotic manipulation demonstrate 100% constraint satisfaction while preserving task performance. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.40075 — 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 2026-10-31T06:07:08.754Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).