Stop Thinking, Just Do!

Sungsoo Kim's Blog

Exploring And Attacking Neural Networks With Activation Atlases

tagsTags

4 May 2019


Article Source


Exploring And Attacking Neural Networks With Activation Atlases

Abstract

By using feature inversion to visualize millions of activations from an image classification network, we create an explorable activation atlas of features the network has learned which can reveal how the network typically represents some concepts.


comments powered by Disqus