Search papers, labs, and topics across Lattice.
Harvard University
1
0
2
Partial data augmentation can match the statistical benefits of full augmentation, challenging the notion that complete symmetry is necessary for optimal learning.
