Metraipolis › archive › dc_142e853031d8

No Free Checker: A Survey of Verifiers for Robot Policies — arXiv:2609.09250

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.09250
License (as archived)
CC0-1.0
Archived
2026-10-16T07:25:12.615Z
SHA-256 of the archived text
ab606d60be336799c642abc3d90a3cf1c729603901b96268664ffa5e1e786b55
Stable URL of this page
https://metraipolis.com/documento/dc_142e853031d8

How to cite

No Free Checker: A Survey of Verifiers for Robot Policies — arXiv:2609.09250 — arXiv (en), https://arxiv.org/abs/2609.09250 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_142e853031d8 (Metraipolis archive dc_142e853031d8, sha256 ab606d60be336799c642abc3d90a3cf1c729603901b96268664ffa5e1e786b55)
BibTeX
@misc{dc142e853031d8,
  title = {No Free Checker: A Survey of Verifiers for Robot Policies — arXiv:2609.09250},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.09250},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_142e853031d8; sha256 ab606d60be336799c642abc3d90a3cf1c729603901b96268664ffa5e1e786b55},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.09250. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A verifier for robot policies reads a candidate behavior and returns a score for how well it did, used both to evaluate vision-language-action policies and to train them. Verifiers range from success detectors and reward models to runtime monitors, safety filters, and temporal-logic specifications. We survey roughly 150 verifiers and compare them along two properties. Availability is how much a verdict costs, how early in a rollout the verdict arrives, and how often a verdict can be asked for. Availability rises as verdicts get cheaper, earlier, and denser. Credibility is how much a high score tells us about the task. Credibility falls as the judgment becomes gameable and self-serving. We group the verifiers by who supplies the judgment: human verifiers, rule-based and formal verifiers, learned and pretrained verifiers, and model-intrinsic verifiers. Across the four families, we find that credibility falls as availability rises. Regardless of who supplies the judgment, there is no free checker. We then examine what validates a verifier itself, and how much a high score tells us. Three measures appear in the literature: agreement with human labels, the performance of the policy it trains, and behavior under reward hacking. We close with nine metrics that make a verifier claim checkable, and coordinates for the verifiers still to be built.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-10 in cs.AI).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Systems and Control (eess.SY).
Pagina del lavoro: https://arxiv.org/abs/2609.09250 — 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 2026-10-16T07:00:16.291Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).