Track the frontier, from condensed matter to cosmology.
Search papers, labs, and topics across Lattice.
This paper introduces OneWM, a unified causal training framework that jointly learns world dynamics and action policies within a shared video-action architecture. By treating actions as conditions for world prediction and targets for policy learning, it investigates their reciprocal benefits in embodied settings. The study finds that learning action-conditioned world dynamics enhances policy performance and generalization, while learning to generate actions improves multi-view world prediction fidelity.
Learning to predict the visual consequences of actions significantly boosts policy performance and generalization, while simultaneously learning to generate actions sharpens world models.
World and action modeling are closely related, yet their reciprocal benefits under unified causal training remain unclear. We introduce OneWM, a simple framework for unified causal training that treats actions as both conditions for world prediction and targets for policy learning within a shared video-action architecture. Using this framework, we investigate whether learning forward dynamics conditioned on specified actions improves action learning, and whether learning to generate actions in turn improves action-conditioned world modeling. We use LIBERO, LIBERO-Plus, and RoboCasa365 as embodied testbeds to evaluate both policy learning and world modeling, complemented by real-world policy experiments. Learning the visual consequences of specified actions improves policy performance and generalization across changing scenes. In turn, action learning improves multi-view world-prediction fidelity under the same architecture. Together, these results answer both questions positively, supporting reciprocal benefits between world modeling and action generation under unified causal training.