Metraipolis › archive › dc_f8cf040eee8a

REFACTOR-VLA: Unsupervised Library Learning of Typed Motor Programs — arXiv:2609.01215

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.01215
License (as archived): CC0-1.0
Archived: 2026-10-17T06:34:00.228Z
SHA-256 of the archived text: ef5c2ecf86369ce5b899b8b4dd25f9ec5913460c1f4971eb0623f59ba3607e5f
Stable URL of this page: https://metraipolis.com/documento/dc_f8cf040eee8a

How to cite

REFACTOR-VLA: Unsupervised Library Learning of Typed Motor Programs — arXiv:2609.01215 — arXiv (en), https://arxiv.org/abs/2609.01215 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f8cf040eee8a (Metraipolis archive dc_f8cf040eee8a, sha256 ef5c2ecf86369ce5b899b8b4dd25f9ec5913460c1f4971eb0623f59ba3607e5f)

BibTeX

@misc{dcf8cf040eee8a,
  title = {REFACTOR-VLA: Unsupervised Library Learning of Typed Motor Programs — arXiv:2609.01215},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.01215},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f8cf040eee8a; sha256 ef5c2ecf86369ce5b899b8b4dd25f9ec5913460c1f4971eb0623f59ba3607e5f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01215. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Most vision-language-action (VLA) models -- OpenVLA, $\pi_0$, RT-2, RDT-1B -- are monolithic: they emit raw motor commands or short action chunks without organizing behavior into reusable abstractions, so they degrade on long-horizon tasks and resist interpretation. Existing skill-discovery methods sidestep the core question of when two action sequences are behaviorally equivalent, either clustering contrastive embeddings or delegating the judgment to a language model uncalibrated to the robot's dynamics. We introduce REFACTOR-VLA, a wake/sleep system for learning reusable skills. Its sleep phase clusters motor-program fragments under a Behavioral-Equivalence Kernel (BEK) computed from rollouts of a learned latent world model $M_\phi$; its wake phase emits typed lambda terms over a Hindley--Milner-inspired vocabulary, consumed by a library-conditioned rectified-flow action decoder. Abstractions are admitted only if they pass Minimum Description Length and return-preservation gates. On LIBERO we report two findings. First, enlarging the world model from 188M to 430M parameters worsened performance on 4 of 4 suites, so capacity alone does not help. Second, the training objective matters far more: adding an auxiliary supervised contrastive (InfoNCE) loss during world-model warmup substantially improves sleep-phase clustering, giving Normalized Mutual Information at $n=3$ seeds of $0.462 \pm 0.021$ (object), $0.867 \pm 0.025$ (spatial), $0.915 \pm 0.013$ (goal) and $0.754 \pm 0.010$ (LIBERO-10), and beating the strongest published baseline on all 4 suites by a mean $\Delta = +0.184$. Across providers ($n=12$) the 95% bootstrap confidence interval for mean pairwise NMI is $[0.683, 0.729]$ (mean $0.705$). The sleep phase also yields the first real-LIBERO task-language library: the decoder uses 2 of 3 admitted abstractions and rewrites all 256 sampled demonstrations.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2609.01215 — 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 2026-10-17T06:15:12.871Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).