Metraipolis › archive › dc_bfaf0f037b65

Cooperative Multi-Agent Vision-Language-Action Models via Reinforced Fine Tuning — arXiv:2609.36588

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.36588
License (as archived)
CC0-1.0
Archived
2026-10-07T16:21:25.732Z
SHA-256 of the archived text
e347bc208034e2d19d942bdec396f97f7ead32394343c8cc97cc5277797ee44c
Stable URL of this page
https://metraipolis.com/documento/dc_bfaf0f037b65

How to cite

Cooperative Multi-Agent Vision-Language-Action Models via Reinforced Fine Tuning — arXiv:2609.36588 — arXiv (en), https://arxiv.org/abs/2609.36588 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bfaf0f037b65 (Metraipolis archive dc_bfaf0f037b65, sha256 e347bc208034e2d19d942bdec396f97f7ead32394343c8cc97cc5277797ee44c)
BibTeX
@misc{dcbfaf0f037b65,
  title = {Cooperative Multi-Agent Vision-Language-Action Models via Reinforced Fine Tuning — arXiv:2609.36588},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.36588},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bfaf0f037b65; sha256 e347bc208034e2d19d942bdec396f97f7ead32394343c8cc97cc5277797ee44c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36588. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We study reinforcement learning (RL) methods for cooperative multi-agent Vision-Language-Action (VLA) models. This problem is challenging because VLAs are pretrained on large-scale single-agent data and therefore lack the fine-grained coordination skills required for inter-robot collaboration. Supervised fine-tuning (SFT) on multi-robot demonstrations partially bridges this gap, but its performance is bounded by the demonstration data and cannot improve from its own experience. We present a three-stage reinforced fine-tuning (RFT) pipeline for multi-agent VLAs. First, initialization-aware data collection sweeps over initial configurations and invokes human demonstrations only when the pretrained VLA repeatedly fails, yielding robustness to initialization shift with reduced human cost. Second, offline credit-filtered tuning assigns credit to individual agents and fine-tunes on per-agent trajectories with positive advantage rather than on entire joint rollouts. Third, we find existing online RL for VLAs are less effective for hard multi-agent tasks, which we attribute to noisy co-exploration and unstable updates. We instead use online latent-space fine tuning, which freeze the VLA and perform RL in its latent noise space. We evaluate our multi-agent VLA with both $\pi_0$ and $\pi_{0.5}$ backbones across 11 tasks in RoboTwin, RoboFactory and real-world manipulation with two Franka robots. Our multi-agent VLA improves the average success rate by $+23.1\%$, $+16.4\%$, and $+44\%$ on RoboTwin, RoboFactory, and real-world tasks, respectively. Code available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.MA).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2609.36588 — 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 2026-10-07T15:54:11.827Z dalla pagina https://arxiv.org/catchup/cs.MA/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).