Metraipolis › archive › dc_997b3f5df11f

Online Evolution Strategy for Flow-Matching VLA Policies via Self-Supervised Trajectory Distribution Optimization — arXiv:2609.38855

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.38855
License (as archived)
CC0-1.0
Archived
2026-10-07T12:44:24.733Z
SHA-256 of the archived text
0995b7edaed18a8d112d4598eb2aa621d59d675b35056c59a9cd201a3e417858
Stable URL of this page
https://metraipolis.com/documento/dc_997b3f5df11f

How to cite

Online Evolution Strategy for Flow-Matching VLA Policies via Self-Supervised Trajectory Distribution Optimization — arXiv:2609.38855 — arXiv (en), https://arxiv.org/abs/2609.38855 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_997b3f5df11f (Metraipolis archive dc_997b3f5df11f, sha256 0995b7edaed18a8d112d4598eb2aa621d59d675b35056c59a9cd201a3e417858)
BibTeX
@misc{dc997b3f5df11f,
  title = {Online Evolution Strategy for Flow-Matching VLA Policies via Self-Supervised Trajectory Distribution Optimization — arXiv:2609.38855},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.38855},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_997b3f5df11f; sha256 0995b7edaed18a8d112d4598eb2aa621d59d675b35056c59a9cd201a3e417858},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.38855. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-Language-Action (VLA) models based on generative frameworks, such as Flow Matching, have recently achieved impressive performance in robotic manipulation. Unlike deterministic policies, Flow Matching enables VLA models to learn conditional action trajectory distributions, where latent noise vectors induce different actions under the same task scenario. However, we observe that these distributions are often ill-formed, with successful and failed behaviors coexisting while considerable probability mass remains in unfavorable regions. To this end, we propose Online-ES, an online adaptation framework for Flow Matching VLAs based on Evolution Strategy (ES), which refines the learned action trajectory distribution through interaction feedback. Instead of pruning the latent noise space, our method performs evolutionary exploration directly in the action trajectory space, where diverse trajectories generated by Flow Matching provide candidate solutions for adaptation. By perturbing sampled trajectories and evaluating their execution outcomes, we derive a self-supervised MSE objective that transfers the evolution direction from trajectory space into model parameter space. Mathematically, we prove that the proposed objective provides an unbiased estimator of the optimal evolution direction. Moreover, we also incorporate failure experiences as negative feedback to regularize the evolution direction, steering the policy away from previously explored failure regions. Experiments in both simulation and real-world environments demonstrate that Online-ES achieves policy improvement comparable to reinforcement fine-tuning, without learning a value model or computing advantages.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.AI).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI); Systems and Control (eess.SY).
Pagina del lavoro: https://arxiv.org/abs/2609.38855 — 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 2026-10-07T12:14:13.958Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).