Metraipolis › archive › dc_92f980df52bf

DriveReferee: Geometric Safety Verdicts Need Not Be Learned for Driving World-Action Models — arXiv:2609.22762

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.22762
License (as archived): CC0-1.0
Archived: 2026-10-09T21:59:16.467Z
SHA-256 of the archived text: 2ad996bef1ca315cca25192b897910b47f7b5215f51b3f0d0490b973b99a8d54
Stable URL of this page: https://metraipolis.com/documento/dc_92f980df52bf

How to cite

DriveReferee: Geometric Safety Verdicts Need Not Be Learned for Driving World-Action Models — arXiv:2609.22762 — arXiv (en), https://arxiv.org/abs/2609.22762 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_92f980df52bf (Metraipolis archive dc_92f980df52bf, sha256 2ad996bef1ca315cca25192b897910b47f7b5215f51b3f0d0490b973b99a8d54)

BibTeX

@misc{dc92f980df52bf,
  title = {DriveReferee: Geometric Safety Verdicts Need Not Be Learned for Driving World-Action Models — arXiv:2609.22762},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.22762},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_92f980df52bf; sha256 2ad996bef1ca315cca25192b897910b47f7b5215f51b3f0d0490b973b99a8d54},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.22762. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Generative world-action models (WAMs) jointly generate future video and vehicle actions, while their action branches remain primarily optimized by expert imitation. Yet imitation provides no explicit closed-loop geometric verdict for generated trajectories, making verification important during both training and deployment. Closed-loop evaluators can check collision and drivable-area violations, but require privileged scene state unavailable at deployment. Existing approaches often close this gap by learning a verifier from sensor features. For these geometric checks, the rule itself is explicit. For example, collision is determined by whether the rolled-out ego footprint overlaps occupied vehicle space. What is unavailable at deployment is the scene state needed to apply the rule. We introduce DriveReferee, which uses a learned geometry readout to predict the scene representation from camera observations and executes the geometric safety rule directly rather than learning it. The resulting analytic referee evaluates collision and drivable-area safety from a scene state and candidate trajectory. During training, it scores self-sampled trajectories on ground-truth state and distills the resulting preferences into the WAM policy. At deployment, the same referee evaluates generated trajectories on this predicted state and selects a safer alternative when needed. The analytic referee requires no verdict-specific training, and its decisions follow an explicit geometric rule. Under matched candidates and inference budgets, it matches or outperforms all learned-verifier and heuristic baselines. Given the same predicted state and trajectory, learning the verdict provides no measurable downstream gain despite requiring tens of thousands of evaluator-labeled training examples. On the full NAVSIM navtest, DriveReferee reaches 92.02 PDMS with single-camera visual input and no external training data.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.22762 — 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 2026-10-09T21:45:20.811Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).