Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
3
0
4
Robots can now adapt to any camera setup without retraining, simply by injecting compact scene tokens into existing VLA models.
Unleashing pretrained video diffusion models for autonomous driving yields surprisingly strong planning performance, suggesting a new path beyond vision-language models.
Forget rigid policies: ESCAPE's adaptive execution and spatial memory enable SOTA long-horizon mobile manipulation, even without detailed instructions.