Search papers, labs, and topics across Lattice.
University of California, Santa Cruz
1
0
3
6
Forget local semantic alignment: CAST unlocks temporally coherent video retrieval and generation by explicitly modeling visual state transitions.
