Metraipolis › archive › dc_ebfe78e8546d
Temporal GRPO: Beyond Trajectory-Level Credit in Vision-Language-Action Reinforcement Learning — arXiv:2608.13026
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.13026
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T04:56:29.345Z
- SHA-256 of the archived text
9b19fd3b22eca6101dc9c53ebc25e893412d842c7dfd396c4f82db38fdf5dc8a- Stable URL of this page
https://metraipolis.com/documento/dc_ebfe78e8546d
How to cite
Temporal GRPO: Beyond Trajectory-Level Credit in Vision-Language-Action Reinforcement Learning — arXiv:2608.13026 — arXiv (en), https://arxiv.org/abs/2608.13026 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ebfe78e8546d (Metraipolis archive dc_ebfe78e8546d, sha256 9b19fd3b22eca6101dc9c53ebc25e893412d842c7dfd396c4f82db38fdf5dc8a)
BibTeX
@misc{dcebfe78e8546d,
title = {Temporal GRPO: Beyond Trajectory-Level Credit in Vision-Language-Action Reinforcement Learning — arXiv:2608.13026},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.13026},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ebfe78e8546d; sha256 9b19fd3b22eca6101dc9c53ebc25e893412d842c7dfd396c4f82db38fdf5dc8a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13026. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Outcome-driven reinforcement learning offers a scalable way to post-train vision-language-action (VLA) policies from sparse task-success feedback. In common GRPO-based VLA post-training, one rollout-level advantage is applied to every action in the trajectory. A rollout that completes several valid stages but fails later can therefore penalize the actions that produced its earlier progress. We call this trajectory-level credit aliasing. Temporal GRPO addresses this problem by constructing detectable task stages, aligning each rollout with stage-specific action intervals, and comparing only rollouts that have entered the same stage. The resulting stage advantages are applied to their corresponding intervals in a single policy update. On RoboTwin 2.0, Temporal GRPO improves task success and sample efficiency, with consistent gains across task horizons. Controlled updates on LIBERO-Long preserve shared prerequisite stages and concentrate improvement at the first stage where rollout outcomes diverge. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-14 in cs.RO). Categorie arXiv: Robotics (cs.RO). Pagina del lavoro: https://arxiv.org/abs/2608.13026 — 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 2026-10-19T04:47:48.655Z dalla pagina https://arxiv.org/catchup/cs.RO/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).