Metraipolis › archive › dc_80ee15a2a406

World-Coherent Decoding: Self-Verifying Test-Time Planning for World Action Models — arXiv:2609.02159

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.02159
License (as archived)
CC0-1.0
Archived
2026-10-17T05:49:53.818Z
SHA-256 of the archived text
87dfeed198c0eaaedb1c770fe7b931e7b4d5c20f80508179d4bf0dd1c777ff12
Stable URL of this page
https://metraipolis.com/documento/dc_80ee15a2a406

How to cite

World-Coherent Decoding: Self-Verifying Test-Time Planning for World Action Models — arXiv:2609.02159 — arXiv (en), https://arxiv.org/abs/2609.02159 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_80ee15a2a406 (Metraipolis archive dc_80ee15a2a406, sha256 87dfeed198c0eaaedb1c770fe7b931e7b4d5c20f80508179d4bf0dd1c777ff12)
BibTeX
@misc{dc80ee15a2a406,
  title = {World-Coherent Decoding: Self-Verifying Test-Time Planning for World Action Models — arXiv:2609.02159},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.02159},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_80ee15a2a406; sha256 87dfeed198c0eaaedb1c770fe7b931e7b4d5c20f80508179d4bf0dd1c777ff12},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.02159. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

World Action Models (WAMs) aim to control robots by stochastically generating visual futures and then decoding actions, but empirical observations indicate that the results can strongly depend on which future is selected. We propose World-Coherent-Decoding (WCD), a self-verifying test-time planning framework that treats WAM rollouts as falsifiable future--action hypotheses. At each decision step, WCD samples multiple candidates from a frozen WAM and ranks them using internal generative signals: flow-based video surprisal for visual plausibility and action path effort for action-generation stability. After execution, the realized observation audits the selected imagination, yielding an imagination--reality mismatch that trains a lightweight online predictor for future candidate selection. Thus, WCD converts delayed self-verification into pre-execution reliability estimation without updating the backbone model. On RoboTwin 2.0, WCD improves Hard success under limited randomized-scene supervision from $55.80\%$ to $60.90\%$, with a $+16.43$ gains on Horizon-3 tasks, and shows qualitative robustness on real Franka visual-shift tests. These results highlight a simple principle: test-time scaling for WAMs depends less on sampling more futures than on selecting reliable ones.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-03 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.02159 — 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 2026-10-17T05:25:58.071Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-03?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).