Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
5
0
9
5
Learning to predict the visual consequences of actions significantly boosts policy performance and generalization, while simultaneously learning to generate actions sharpens world models.
Decoupling denoising and context writing parameters in autoregressive video models unlocks unprecedented long-horizon generation capabilities, extending continuous output to 24 hours from just 5-second training rollouts.
Video generative priors only translate into robust physical control when paired with explicit world-to-action information routing and synchronized joint denoising rather than standard monolithic fine-tuning.
RoboDojo reveals that integrating simulation and real-world tasks can significantly enhance the evaluation of robot manipulation policies, bridging the gap between theoretical performance and practical deployment.