analiza aiqbench
LaWAM introduces an efficient robot control framework that leverages latent space dynamics to predict scene evolution without the computational overhead of pixel-level video generation. By repurposing forward decoders from pretrained vision models, it achieves state-of-the-art success rates in benchmark tasks while significantly reducing inference latency compared to existing World-Action Models. This advancement addresses a critical bottleneck in deploying dynamic-aware policies for real-world robotic manipulation systems.
LaWAM: Latent World Action Models for Efficient Dynamics-Aware Robot Policies
arXiv:2606.15768v1 Announce Type: cross Abstract: Vision-Language-Action models (VLAs) leverage large-scale vision-language pretraining for semantic robot control, but often lack explicit foresight into how robot actions change the scene. World-Action Models (WAMs) address this limitation by conditioning policies on predicted futures, yet existing approaches typically rely on computationally expensive video generation with substantial pixel-level redundancy. We present LaWAM, a Latent World Action Model that exposes predictive dynamics to robot policies through compact latent visual subgoals instead of reconstructed future video. At the core of LaWAM is a latent-action-conditioned Latent World Model (LaWM). We obtain LaWM by training a latent action model in the latent space of a pretrained vision foundation model and repurposing its forward decoder to predict future observation features for scene evolution. LaWAM then conditions action generation on these predicted latent visual subgoals to enable dynamics-aware robot control. LaWAM achieves state-of-the-art or competitive success rates (SRs) across LIBERO (98.6% SR), RoboTwin (91.22% SR), and real-world manipulation tasks while retaining low-latency inference. LaWAM runs in 187 ms per action-chunk prediction and achieves up to 24x lower wall-clock latency than pixel-space WAMs.
Abstract:Vision-Language-Action models (VLAs) leverage large-scale vision-language pretraining for semantic robot control, but often lack explicit foresight into how robot actions change the scene. World-Action Models (WAMs) address this limitation by conditioning policies on predicted futures, yet existing approaches typically rely on computationally expensive video generation with substantial pixel-level redundancy. We present LaWAM, a Latent World Action Model that exposes predictive dynamics to robot policies through compact latent visual subgoals instead of reconstructed future video. At the core of LaWAM is a latent-action-conditioned Latent World Model (LaWM). We obtain LaWM by training a latent action model in the latent space of a pretrained vision foundation model and repurposing its forward decoder to predict future observation features for scene evolution. LaWAM then conditions action generation on these predicted latent visual subgoals to enable dynamics-aware robot control. LaWAM achieves state-of-the-art or competitive success rates (SRs) across LIBERO (98.6% SR), RoboTwin (91.22% SR), and real-world manipulation tasks while retaining low-latency inference. LaWAM runs in 187 ms per action-chunk prediction and achieves up to 24x lower wall-clock latency than pixel-space WAMs.
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2606.15768 [cs.RO]
(or arXiv:2606.15768v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.15768 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 · 272 cuvinte
Subiecte conexe
DeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imagination
arXiv · 9 sept. 2026
GeoSem-WAM: Geometry- and Semantic-Aware World Action Models
arXiv · 3 iun. 2026
DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation
arXiv · 22 sept. 2026
SLIP-VLA: Single-Step Latent Imagination for Policy Learning in Vision-Language-Action Models
arXiv · 29 sept. 2026