Search papers, labs, and topics across Lattice.
1
0
2
Existing vision-language models fall flat when it comes to spotting time-dependent robot errors, but TIMID nails it with weak supervision and a clever VAD architecture.
