Search papers, labs, and topics across Lattice.
DI-ENS
1
0
3
Cut deep learning costs without sacrificing accuracy by swapping exact backpropagation for unbiased, randomized approximations of vector-Jacobian products.
