Search papers, labs, and topics across Lattice.
1
0
3
Current vision-language-action models choke on dynamic robotic manipulation because they lack spatiotemporal reasoning, but a new dataset and architecture, DOMINO and PUMA, close the gap.
