Metraipolis › archive › dc_01cccb4351d6
Overcoming Statistical Bias in Action-Controllable World Models — arXiv:2608.04653
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.04653
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T11:48:03.674Z
- SHA-256 of the archived text
c401a1be416e1202e6675e26aaa4ae1e47bbddbc096936dd9e2afb114072e744- Stable URL of this page
https://metraipolis.com/documento/dc_01cccb4351d6
How to cite
Overcoming Statistical Bias in Action-Controllable World Models — arXiv:2608.04653 — arXiv (en), https://arxiv.org/abs/2608.04653 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_01cccb4351d6 (Metraipolis archive dc_01cccb4351d6, sha256 c401a1be416e1202e6675e26aaa4ae1e47bbddbc096936dd9e2afb114072e744)
BibTeX
@misc{dc01cccb4351d6,
title = {Overcoming Statistical Bias in Action-Controllable World Models — arXiv:2608.04653},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.04653},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_01cccb4351d6; sha256 c401a1be416e1202e6675e26aaa4ae1e47bbddbc096936dd9e2afb114072e744},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.04653. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Action-conditioned world models aim to predict how visual environments evolve under an agent's actions. Yet future frames are often highly predictable from visual inertia and recurring motion patterns alone. This creates a shortcut: models can fit the data by exploiting statistical biases without making their visible dynamics meaningfully depend on the action. As a result, different actions may produce similar futures, while motion may persist even under zero action. The key question is how to reduce reliance on statistical shortcuts from dominating action-conditioned prediction. We argue that action control requires more than injecting action features; it requires enforcing consistency under counterfactual changes to actions and observations. Based on this insight, we introduce CoCo, a Counterfactual Consistency framework to enhance action controllability through two complementary constraints. Multi-step counterfactual consistency constrains reference, inverse-action, and zero-action rollouts, while action-spatial counterfactual consistency enforces consistent predictions under mirrored scenes and transformed actions. Together, they reduce reliance on statistical shortcuts from substituting for action-dependent dynamics. We further introduce Action Response Consistency (ARC) and Drift Energy (DE) to assess action controllability, together with Mini-SSMB for same-state, multi-action counterfactual evaluation. On Mini-SSMB, our full model achieved ARC_inv of 0.412 and ARC_ref of 0.483, while reducing DE by 17.07% relative to the baseline. On VP2 visual planning, it achieves the highest average success rate among SOTA models, at 73.1%. Experiments on BAIR and RoboNet further show that these gains preserve video prediction quality and transfer across model settings. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-06 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Robotics (cs.RO). Pagina del lavoro: https://arxiv.org/abs/2608.04653 — 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 2026-10-31T11:37:40.698Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-08-06?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).