no code implementations • ICLR 2022 • Tiago Salvador, Stephanie Cairns, Vikram Voleti, Noah Marshall, Adam Oberman
However, they still have drawbacks: they reduce accuracy (AGENDA, PASS, FTC), or require retuning for different false positive rates (FSN).
1 code implementation • 21 Nov 2019 • Gordon MacDonald, Andrew Godbout, Bryn Gillcash, Stephanie Cairns
We construct a new architecture for deep neural networks where all layers (except the output layer) of the network are a combination of rotation, permutation, diagonal, and activation sublayers which are all volume preserving.