Metraipolis › archive › dc_44c8accfd64d

How Should Vision-Language-Action Models Use Proprioceptive State? — arXiv:2608.03052

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.03052
License (as archived)
CC0-1.0
Archived
2026-10-31T13:07:50.894Z
SHA-256 of the archived text
cddb8183e6b6c7fb5988498ef6a602b86f422d7afc0c5a1e376e942592bc1e9d
Stable URL of this page
https://metraipolis.com/documento/dc_44c8accfd64d

How to cite

How Should Vision-Language-Action Models Use Proprioceptive State? — arXiv:2608.03052 — arXiv (en), https://arxiv.org/abs/2608.03052 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_44c8accfd64d (Metraipolis archive dc_44c8accfd64d, sha256 cddb8183e6b6c7fb5988498ef6a602b86f422d7afc0c5a1e376e942592bc1e9d)
BibTeX
@misc{dc44c8accfd64d,
  title = {How Should Vision-Language-Action Models Use Proprioceptive State? — arXiv:2608.03052},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.03052},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_44c8accfd64d; sha256 cddb8183e6b6c7fb5988498ef6a602b86f422d7afc0c5a1e376e942592bc1e9d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.03052. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recent Vision-Language-Action (VLA) models almost universally take robot proprioceptive state as input, yet wire it in incompatible ways -- serialized into text prompts, projected into the vision-language prefix, or fed directly to the action expert -- and almost always as a single current frame. Three questions remain open: (1) whether, and on which tasks, current state actually improves closed-loop control; (2) how much state history helps, and whether its benefit reflects genuine temporal variation rather than added conditioning capacity; and (3) where state should enter the model -- the vision-language backbone or the action-generation module. We answer these questions through controlled experiments on a flow-matching VLA, fixing the backbone, training data, action representation, and evaluation protocol throughout. We implement five representative interfaces -- discrete state prompt, VLM prefix, action prefix, state expert, and feature modulation -- under matched implementation details, and evaluate them on 45 atomic tasks spanning three task families plus 20 composite tasks; we then sweep the state-history length from 1 to 96 frames to examine how historical state information affects model performance. The experiments yield systematic answers to all three questions, distilled into testable design principles for state-aware VLAs.

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/2608.03052 — 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 2026-10-31T12:55:56.550Z 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).