Metraipolis › archive › dc_354d30f096fa

Does a Learned Corrector Beat a Simple Retreat? Evidence from a Frozen VLA — arXiv:2610.06921

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.06921
License (as archived)
CC0-1.0
Archived
2026-10-07T13:11:37.846Z
SHA-256 of the archived text
2c1312c9b39f792c715d27d10351aae54234a2c238de5b67ddeac64f57ba3210
Stable URL of this page
https://metraipolis.com/documento/dc_354d30f096fa

How to cite

Does a Learned Corrector Beat a Simple Retreat? Evidence from a Frozen VLA — arXiv:2610.06921 — arXiv (en), https://arxiv.org/abs/2610.06921 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_354d30f096fa (Metraipolis archive dc_354d30f096fa, sha256 2c1312c9b39f792c715d27d10351aae54234a2c238de5b67ddeac64f57ba3210)
BibTeX
@misc{dc354d30f096fa,
  title = {Does a Learned Corrector Beat a Simple Retreat? Evidence from a Frozen VLA — arXiv:2610.06921},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.06921},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_354d30f096fa; sha256 2c1312c9b39f792c715d27d10351aae54234a2c238de5b67ddeac64f57ba3210},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.06921. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Before deploying runtime recovery for a frozen vision-language-action (VLA) policy, one must establish that an intervention improves success beyond ordinary run-to-run variation and that its complexity adds value over a simple action. We evaluate these questions on frozen $\pi_{0.5}$ across four RoboTwin tasks. For each test seed, we pair rollouts with and without correction and include a same-seed base-policy re-run as a placebo. Seed-cluster intervals and prespecified comparison rules assess net gains against stochastic outcome changes. Across 3,888 paired episodes, the full pipeline raises success on beat_allowbreak block_allowbreak hammer by $+13.5$\,pp (95\% interval $[+9.4,+17.7]$), with no detectable gain on the other three tasks at the deployed weight. Among failed base episodes on the responsive task, $43.2\%$ succeed on a plain re-run, compared with $63.5\%$ after correction; many nominal rescues therefore reflect the base policy's own variability. A fixed-time trigger and scripted return to an earlier joint configuration produce a net gain with no detected difference from the learned pipeline across two rounds, although our prespecified equivalence criterion is not met consistently. Pausing and a constant-action control do not yield comparable gains. On this benchmark, the decision to intervene depends strongly on the task, and a paired placebo plus a simple retreat baseline are needed to establish what learned correction contributes.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.RO, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2610.06921 — 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 2026-10-07T12:42:33.484Z dalla pagina https://arxiv.org/list/cs.RO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).