Metraipolis › archive › dc_398f4f366e30
Before We Adapt: The Zero-Shot Generalization Bottleneck in Continual World Models — arXiv:2610.41041
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41041
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T08:11:43.539Z
- SHA-256 of the archived text
5bcb5226c54b19ec36dad83710030f7d0e86e36cec83c7cd30bbf5a7e93ed621- Stable URL of this page
https://metraipolis.com/documento/dc_398f4f366e30
How to cite
Before We Adapt: The Zero-Shot Generalization Bottleneck in Continual World Models — arXiv:2610.41041 — arXiv (en), https://arxiv.org/abs/2610.41041 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_398f4f366e30 (Metraipolis archive dc_398f4f366e30, sha256 5bcb5226c54b19ec36dad83710030f7d0e86e36cec83c7cd30bbf5a7e93ed621)
BibTeX
@misc{dc398f4f366e30,
title = {Before We Adapt: The Zero-Shot Generalization Bottleneck in Continual World Models — arXiv:2610.41041},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41041},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_398f4f366e30; sha256 5bcb5226c54b19ec36dad83710030f7d0e86e36cec83c7cd30bbf5a7e93ed621},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41041. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Continual reinforcement learning (CRL) requires agents to operate under environmental changes while reusing prior knowledge whenever possible. We argue that model-based reinforcement learning provides a promising framework for this setting because learned world models (WMs) can, in principle, support change detection, knowledge reuse, adaptation, and acquisition of new knowledge. However, such benefits depend heavily on whether the learned model can generalize beyond the contexts in which it was trained. In this work, we propose to formulate this problem with a contextual Markov decision process in which context is decomposed into behavioral and physical components. This view organizes continual WM learning into reuse, adaptation, and acquisition of new knowledge. As a first step, we study direct reuse when behavioral rules remain unchanged. Using Dyna-Q and TD-MPC in controlled FourRooms tasks, we find that zero-shot reuse can degrade substantially under physical-context changes and that this behavior depends strongly on the observation representation used by the WM. Additional ablation studies and latent-space diagnostics suggest that learned representations do not automatically provide the invariances needed for reuse. These results identify representation generalization as an important bottleneck for reliable world-model reuse in CRL. 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/2610.41041 — 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 2026-11-03T08:02:38.905Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).