Metraipolis › archive › dc_70922d55e522

Do Not Cut When Uncertain: Rejectable and Calibrated Decision Heads for VLA Policies in Robotic Harvesting — arXiv:2609.35039

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.35039
License (as archived)
CC0-1.0
Archived
2026-10-08T00:17:31.945Z
SHA-256 of the archived text
ec7ef6bddf428e74659f39b6bbba2c2ac2709e770376b687e57c43d304a5922a
Stable URL of this page
https://metraipolis.com/documento/dc_70922d55e522

How to cite

Do Not Cut When Uncertain: Rejectable and Calibrated Decision Heads for VLA Policies in Robotic Harvesting — arXiv:2609.35039 — arXiv (en), https://arxiv.org/abs/2609.35039 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_70922d55e522 (Metraipolis archive dc_70922d55e522, sha256 ec7ef6bddf428e74659f39b6bbba2c2ac2709e770376b687e57c43d304a5922a)
BibTeX
@misc{dc70922d55e522,
  title = {Do Not Cut When Uncertain: Rejectable and Calibrated Decision Heads for VLA Policies in Robotic Harvesting — arXiv:2609.35039},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35039},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_70922d55e522; sha256 ec7ef6bddf428e74659f39b6bbba2c2ac2709e770376b687e57c43d304a5922a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35039. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-Language-Action (VLA) policies trained with behavior cloning or flow matching are optimized to output an action trajectory, but they cannot express "I don't know" or "I should not act." In robotic harvesting, occlusion makes single-frame decisions fundamentally ambiguous: identical pixels can correspond either to a cuttable stem or to no stem at all. Existing VLAs are forced to commit, leading to high-confidence errors with irreversible consequences. We argue that the failure mode of a VLA is determined not by backbone scale but by its output interface. We propose Rejectable and Calibrated Decision Heads (RCDH), a typed, rejectable, and calibrated output interface that can be attached to a frozen VLA backbone without retraining or new features. RCDH introduces (i) a decision schema with explicit rejection and ordered, conditional decomposition, and (ii) a calibration procedure for risk-aware abstention. We evaluate RCDH on a robotic harvesting platform with controllable leaf occlusion, comparing generative, enumerated, calibrated, and rejectable interfaces. We show that replacing only the output head restores out-of-distribution usability under occlusion while preserving in-distribution performance. We further test whether the ordering of the rejection space is critical. Our results suggest that the right to refuse, rather than a larger model, is the missing interface for reliable manipulation under uncertainty.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-29 in cs.RO).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2609.35039 — 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 2026-10-07T23:52:16.590Z dalla pagina https://arxiv.org/catchup/cs.RO/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).