Metraipolis › archive › dc_cd26c3005ebf
World Models with Predictable Long-Horizon Marginals — arXiv:2609.32657
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.32657
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T20:55:51.236Z
- SHA-256 of the archived text
d4c423dd82eabadeb81e82b61c4e76c750c624ce3a80ebed037712a13162c6ed- Stable URL of this page
https://metraipolis.com/documento/dc_cd26c3005ebf
How to cite
World Models with Predictable Long-Horizon Marginals — arXiv:2609.32657 — arXiv (en), https://arxiv.org/abs/2609.32657 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cd26c3005ebf (Metraipolis archive dc_cd26c3005ebf, sha256 d4c423dd82eabadeb81e82b61c4e76c750c624ce3a80ebed037712a13162c6ed)
BibTeX
@misc{dccd26c3005ebf,
title = {World Models with Predictable Long-Horizon Marginals — arXiv:2609.32657},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.32657},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cd26c3005ebf; sha256 d4c423dd82eabadeb81e82b61c4e76c750c624ce3a80ebed037712a13162c6ed},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32657. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Accurate one-step predictions do not ensure that a world model's rollouts retain the data distribution. We make the model's decoded stationary law explicit by learning a decoder of a fixed Gaussian reference and constraining the behaviour-averaged transition to preserve that reference. For controlled systems, a joint transition uses a conditional action chart to preserve behaviour occupancy without requiring invariance at each fixed action. Joint state--action rotations and parallel Gaussian noise give an exactly preserving transition with a tractable conditional density. We derive an absolute convergence bound from finite initialization banks and control departure from the reference through conditional action-space divergence. Across $216$ fitted pixel checkpoints on twelve control tasks, the occupancy model with a reference mixture retains every evaluated chain at $10^5$ steps in all $36$ task--seed cells, with a rollout-minus-reference energy-statistic difference of $-0.0002\pm0.0003$ (training-seed standard error). Each of the four nonpreserving comparison arms loses chains, although the Gaussian arm is more accurate at ten steps. An offline DreamerV3 reference also achieves better short-horizon accuracy. These results distinguish three properties of a world model: the distribution it approaches, the rate of approach, and the conditional dynamics it learns. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.32657 — 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 2026-10-07T20:29:00.906Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).