Search papers, labs, and topics across Lattice.
1
0
2
0
The results suggest that shifting part of the modeling burden from network parameters to a structured input representation can support compact multi-task vision systems with reduced data, memory, and compute requirements.