Search papers, labs, and topics across Lattice.
Served in an advisory role
1
0
2
Pre-trained vision encoders are shockingly bad at distinguishing identity when background context is controlled, but a simple contrastive learning scheme can fix it.
