Metraipolis › archive › dc_30fffe6ae2b4
EvoHIL: Self-Evolving Reward and Flow-Matched Policy Optimization for Robust Human-in-the-Loop Reinforcement Learning — arXiv:2608.03872
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.03872
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T13:08:25.746Z
- SHA-256 of the archived text
915cfbc73c8c7ae80e1133579de86019abf6b284a9ae7ec852bdcf4dae4e1c24- Stable URL of this page
https://metraipolis.com/documento/dc_30fffe6ae2b4
How to cite
EvoHIL: Self-Evolving Reward and Flow-Matched Policy Optimization for Robust Human-in-the-Loop Reinforcement Learning — arXiv:2608.03872 — arXiv (en), https://arxiv.org/abs/2608.03872 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_30fffe6ae2b4 (Metraipolis archive dc_30fffe6ae2b4, sha256 915cfbc73c8c7ae80e1133579de86019abf6b284a9ae7ec852bdcf4dae4e1c24)
BibTeX
@misc{dc30fffe6ae2b4,
title = {EvoHIL: Self-Evolving Reward and Flow-Matched Policy Optimization for Robust Human-in-the-Loop Reinforcement Learning — arXiv:2608.03872},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.03872},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_30fffe6ae2b4; sha256 915cfbc73c8c7ae80e1133579de86019abf6b284a9ae7ec852bdcf4dae4e1c24},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.03872. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Human-in-the-loop reinforcement learning (HIL-RL) enables robots to learn contact-rich manipulation from limited real-world interaction, but deployment exposes three coupled limitations: static visual reward models fail under scene changes; independently sampled actions cause temporally inconsistent motion; and vision-based policies remain sensitive to appearance shifts. We present EvoHIL, a unified framework that adapts the reward model, action generator, and visual do main within a staged human-in-the-loop learning process. First, self-evolving reward (SER) adapts the success classifier from human-confirmed positives and provisional weak negatives. Second, Action Flow Stabilization (AFS) generates temporally coherent action chunks through flow matching, grounding policy updates in executed action prefixes and demonstrated behavior. Third, retention-aware offline fine-tuning replays relit interaction data while anchoring the AFS actor-critic to prior behavior, adapting the visual domain without additional robot interaction. Across six manipulation tasks on Franka FR3 and SO-101 arms under a controlled lighting shift, EvoHIL improves task success, agreement with human-confirmation labels, motion smoothness, and completion time relative to human-in-the-loop and imitation this http URL page: this https URL Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-05 in cs.RO). Categorie arXiv: Robotics (cs.RO). Pagina del lavoro: https://arxiv.org/abs/2608.03872 — 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 2026-10-31T12:55:56.567Z dalla pagina https://arxiv.org/catchup/cs.RO/2026-08-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).