Search papers, labs, and topics across Lattice.
1
0
3
9
Autoencoders can reconstruct rare image features more faithfully by explicitly focusing on statistically uncommon spatial locations during training.