Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
1
0
3
0
Vision models can learn to denoise and focus on semantic object regions without pixel-level supervision, simply by minimizing a learned energy function over spatial features.