GAN Dissection: Visualizing and Understanding Generative Adversarial Networks David Bau1,2, Jun-Yan Zhu1, Hendrik Strobelt2,3, Bolei Zhou4, Joshua B. Tenenbaum1, William T. Freeman1, Antonio Torralba1,2 1MIT CSAIL, 2MIT-IBM Watson AI Lab, 3IBM Research, 4The Chinese University of Hong Kong Historical note: this paper's "GAN Paint" was one of the first image editing tools with GANs, showing for the
![GAN Dissection](https://cdn-ak-scissors.b.st-hatena.com/image/square/fa849204acebe5b2adc67da1ab52b16a9cc270a4/height=288;version=1;width=512/https%3A%2F%2Fgandissect.csail.mit.edu%2Fimg%2Fganpaint_v3.gif)