Metraipolis › archive › dc_4f2b35cc20de
An Action Is Worth One Patch: Unified World-Action Modeling with PatchWAM — arXiv:2609.25961
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.25961
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T16:43:44.594Z
- SHA-256 of the archived text
84f74a6678eb5b826211ca19d414585f58f37c43a9517d4f894ca02002de61ee- Stable URL of this page
https://metraipolis.com/documento/dc_4f2b35cc20de
How to cite
An Action Is Worth One Patch: Unified World-Action Modeling with PatchWAM — arXiv:2609.25961 — arXiv (en), https://arxiv.org/abs/2609.25961 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4f2b35cc20de (Metraipolis archive dc_4f2b35cc20de, sha256 84f74a6678eb5b826211ca19d414585f58f37c43a9517d4f894ca02002de61ee)
BibTeX
@misc{dc4f2b35cc20de,
title = {An Action Is Worth One Patch: Unified World-Action Modeling with PatchWAM — arXiv:2609.25961},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.25961},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4f2b35cc20de; sha256 84f74a6678eb5b826211ca19d414585f58f37c43a9517d4f894ca02002de61ee},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25961. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Generative visual models offer a foundation for learning representations of physical dynamics, yet their extension to continuous control raises a fundamental question: do visual prediction and action generation require separate computational pathways? Existing approaches usually introduce trainable action heads or separate action experts to bridge low-dimensional states and high-dimensional visual representations. In this work, we explore whether the visual backbone's existing capacity can also support control when actions are expressed in a compatible representation. Thus, we introduce PatchWAM (Patch World-Action Model), which treats continuous actions as another type of patch through a fixed mapping called Action-as-Patch. This allows a single model to predict both how the robot should move and what the scene may look like afterward. Visual prediction and action generation become parts of the same generative process, without a dedicated action head or separate action expert. Experiments with subsampled training windows show gains over a matched dual-expert control, while benchmark evaluations reach 91.8% success rate on LIBERO-Plus and 96.12% on RoboTwin 2.0 in a full-data setting with additional augmented demonstrations. More broadly, the result suggests that capability need not be added where it can be inherited: the constraint on extending a generative backbone is the interface a new signal is written in, not the capacity to model it. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-23 in cs.RO). Categorie arXiv: Robotics (cs.RO). Pagina del lavoro: https://arxiv.org/abs/2609.25961 — 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 2026-10-08T15:56:53.475Z dalla pagina https://arxiv.org/catchup/cs.RO/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).