Metraipolis › archive › dc_ffbd0095abcf

How Much Planning Is Enough? Reducing Search and Computation in World-Model Planning — arXiv:2610.08350

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.08350
License (as archived): CC0-1.0
Archived: 2026-10-07T11:21:39.692Z
SHA-256 of the archived text: 5576515276d04ccd89dbd6ccafbb87c9fd526e915ccb038bb6301657be6ac6fb
Stable URL of this page: https://metraipolis.com/documento/dc_ffbd0095abcf

How to cite

How Much Planning Is Enough? Reducing Search and Computation in World-Model Planning — arXiv:2610.08350 — arXiv (en), https://arxiv.org/abs/2610.08350 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ffbd0095abcf (Metraipolis archive dc_ffbd0095abcf, sha256 5576515276d04ccd89dbd6ccafbb87c9fd526e915ccb038bb6301657be6ac6fb)

BibTeX

@misc{dcffbd0095abcf,
  title = {How Much Planning Is Enough? Reducing Search and Computation in World-Model Planning — arXiv:2610.08350},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.08350},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ffbd0095abcf; sha256 5576515276d04ccd89dbd6ccafbb87c9fd526e915ccb038bb6301657be6ac6fb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.08350. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Visual world models enable goal-directed control through decision-time action search, but their deployment efficiency is often limited by conservatively large planning budgets. We show that competitive task performance can be achieved without agreement with the Full-budget action, that sufficient budgets vary across model--task pairs, and that iterative planners repeatedly encode solve-invariant context. To address these inefficiencies, we propose {SufficientPlan}, a simple deployment framework that requires no modification to pretrained world models or planner updates. Its {Paired Sequential Budget Certification (PSBC)} component uses paired closed-loop evidence to search for and certify a reduced model--task-specific budget within a predefined Full-performance tolerance. Its {Static-Context Reuse (SCR)} component caches observation and goal representations across search iterations while preserving candidate-dependent planning and selected actions. Experiments across multiple world-model backbones and visual-control tasks show that SufficientPlan substantially reduces search budgets and planning latency while maintaining competitive control performance.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.08350 — 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 2026-10-07T11:15:24.815Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).