Search papers, labs, and topics across Lattice.
5
0
5
13
This work introduces Adaptive Agents via Latent Topologies (AALT), which requests demonstrations that maximize expected gains in start-goal connectivity and shows that this objective is formally tied to information gain about task reachability.
A decentralized approach using neural HJR learning achieves superior safety and efficiency in multi-arm motion planning, outperforming traditional methods.
Neural motion planners can now navigate complex, multi-room environments with significantly improved robustness by incorporating sparse roadmap guidance as weak supervision.
Coordinating multi-robot teams to complete manipulation tasks just got easier: GoC-MPC handles dynamic task assignments and disturbances without training data or environment models.
Steer your robot's diffusion policy away from failure modes at inference time with a lightweight performance predictor trained via self-supervised attention.
