analiza aiqbench
The paper presents a lightweight latent world model (257M parameters) that sits between a pretrained vision-language-action policy and its action expert, trained on ground-truth robot actions without pixel reconstruction or large-scale video pretraining. The reported gains are benchmark results on RoboTwin 2.0 and LIBERO variants plus a real-robot improvement for π0.5, all from the authors and not yet independently replicated. For AI readers, the practical question is whether predicting latent futures cheaply improves embodied policies beyond what video-based world models achieve.
WAVE-0: Bridging VLA Policies and WAMs with Lightweight Latent Future Prediction
arXiv:2610.18568v1 Announce Type: new Abstract: Vision-language-action (VLA) policies excel at generalist robotic manipulation but generate actions without explicitly modeling their environmental consequences. World action models (WAMs) incorporate future prediction, yet most rely on costly video generation, while latent alternatives either use predicted futures only as auxiliary targets or infer latent actions from suboptimal visual transitions through costly large-scale video pretraining. We introduce WAVE-0 (World-model-Augmented VLA Execution), a framework that bridges VLAs and WAMs through a lightweight latent world model. By connecting the semantic understanding of pretrained VLAs with explicit future prediction, WAVE-0 allows anticipated action consequences to directly guide action generation. Built on a pretrained VLA, its 257M world model translates VLM-generated latent action plans into predicted future representations that condition the action expert, unifying latent planning, future prediction, and action generation. The world model is trained directly on ground-truth robot actions with a counterfactual ranking objective, requiring neither pixel reconstruction nor large-scale video pretraining. WAVE-0 achieves success rates of 91.47% on RoboTwin 2.0 and 98.7% on LIBERO, along with 88.0%, 61.4%, and 75.0% on LIBERO-Plus, LIBERO-Pro, and LIBERO-Para, respectively. On real-world tasks, it improves the success rate of $\pi_{0.5}$ from 38.3% to 73.3%. These results demonstrate strong manipulation performance and generalization across diverse distribution shifts.
Abstract:Vision-language-action (VLA) policies excel at generalist robotic manipulation but generate actions without explicitly modeling their environmental consequences. World action models (WAMs) incorporate future prediction, yet most rely on costly video generation, while latent alternatives either use predicted futures only as auxiliary targets or infer latent actions from suboptimal visual transitions through costly large-scale video pretraining. We introduce WAVE-0 (World-model-Augmented VLA Execution), a framework that bridges VLAs and WAMs through a lightweight latent world model. By connecting the semantic understanding of pretrained VLAs with explicit future prediction, WAVE-0 allows anticipated action consequences to directly guide action generation. Built on a pretrained VLA, its 257M world model translates VLM-generated latent action plans into predicted future representations that condition the action expert, unifying latent planning, future prediction, and action generation. The world model is trained directly on ground-truth robot actions with a counterfactual ranking objective, requiring neither pixel reconstruction nor large-scale video pretraining. WAVE-0 achieves success rates of 91.47% on RoboTwin 2.0 and 98.7% on LIBERO, along with 88.0%, 61.4%, and 75.0% on LIBERO-Plus, LIBERO-Pro, and LIBERO-Para, respectively. On real-world tasks, it improves the success rate of $\pi_{0.5}$ from 38.3% to 73.3%. These results demonstrate strong manipulation performance and generalization across diverse distribution shifts.
Robotics (cs.RO)
arXiv:2610.18568 [cs.RO]
(or arXiv:2610.18568v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.18568 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 294 cuvinte
Subiecte conexe
Foresight Without Seeing: Latent Futures for World Action Models
arXiv · 2 oct. 2026
TS-WAM: Temporal Latent Actions and World-Action Coupling for Vision-Language Robot Control
arXiv · 16 oct. 2026
Robust-WAM: Bridging Generative Pretraining and Semantic Foresight in World-Action Models
arXiv · 7 aug. 2026
Inferring Action from Future Latent State for Robotic Manipulation
arXiv · 25 aug. 2026