Metraipolis › archive › dc_e82478a6760c
CausalDreamer: Learning Predictive World Models with Latent Disentanglement — arXiv:2610.12016
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.12016
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T06:23:36.533Z
- SHA-256 of the archived text
4d03695e53acc33f8078c7007e6cb5f75400db5206abf6cb7c6ce910503dbc36- Stable URL of this page
https://metraipolis.com/documento/dc_e82478a6760c
How to cite
CausalDreamer: Learning Predictive World Models with Latent Disentanglement — arXiv:2610.12016 — arXiv (en), https://arxiv.org/abs/2610.12016 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e82478a6760c (Metraipolis archive dc_e82478a6760c, sha256 4d03695e53acc33f8078c7007e6cb5f75400db5206abf6cb7c6ce910503dbc36)
BibTeX
@misc{dce82478a6760c,
title = {CausalDreamer: Learning Predictive World Models with Latent Disentanglement — arXiv:2610.12016},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.12016},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e82478a6760c; sha256 4d03695e53acc33f8078c7007e6cb5f75400db5206abf6cb7c6ce910503dbc36},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.12016. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
World models for control must capture which aspects of the environment respond to the agent's actions and which are relevant to reward. Generative world models such as Dreamer 4 consist of a video tokenizer, which encodes each frame into a latent, and a dynamics model, which is pretrained to predict future latents from past latents and actions. Yet the tokenizer is trained with a reconstruction objective, without action or reward supervision, so its latent provides no explicit mechanism to separate controllable, uncontrollable, reward-relevant, and reward-irrelevant information. We propose \textit{CausalDreamer}, which keeps the tokenizer frozen and re-encodes its latent into a factored representation of four groups along two axes: controllability, where only the two controllable groups receive the action, and reward relevance, learned by predicting the reward from the two reward-relevant groups. The pretrained dynamics model is then fine-tuned to predict the factored representation. We evaluate \textit{CausalDreamer} and the pretrained world model it starts from with model-predictive planning on 20 MMBench2 tasks: 10 clean tasks seen during training and 10 unseen tasks, of which 6 are manipulated variants of clean tasks with a changed background, object, or maze layout, and 4 are new environments. We normalize returns so that a policy taking uniformly random actions scores 0 and an expert scores 1. \textit{CausalDreamer} achieves a 14\% higher normalized score than the pretrained world model on the clean tasks (0.199 vs.\ 0.175) and a 25\% higher score on the manipulated variants (0.307 vs.\ 0.246), while neither model scores meaningfully above the random policy in the new environments. Additionally, our analysis shows that the factored representation separates reward-irrelevant changes, such as a changed background, from its reward-relevant groups.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.12016 — 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 2026-10-09T05:50:31.700Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).