Optimizing Neural Networks
Walt discusses the concept of measuring the rate of change of the output of a neural network with respect to the input and how it relates to the Leapstep constraint. He also talks about the evolution of their work when applying it to different datasets and domains, highlighting the challenges faced when scaling from MNIST to CFAR and from CFAR to ImageNet.In this clip
From this podcast
Data Skeptic
Adversarial Explanations
Related Questions