Search papers, labs, and topics across Lattice.
1
0
3
2
Extracting interaction cues from a frozen video model enables robots to achieve up to 90.6% success in manipulation tasks without costly rollout processes.
