Search papers, labs, and topics across Lattice.
3
0
6
8
Planning in latent space, rather than directly in action space, overcomes the exponential decay in feasible trajectories, unlocking efficient long-horizon decision making for embodied agents.
Ditch the clunky architectures: a single diffusion model can now handle vision, language, and robot control to achieve SOTA manipulation performance.
Normalizing Flows can now be used to train robust robotic policies in online reinforcement learning, thanks to a simple stabilization technique that overcomes previous instability issues.
