DeepAI AI Chat
Log In Sign Up

GAN Dissection: Visualizing and Understanding Generative Adversarial Networks

by   David Bau, et al.

Generative Adversarial Networks (GANs) have recently achieved impressive results for many real-world applications, and many GAN variants have emerged with improvements in sample quality and training stability. However, visualization and understanding of GANs is largely missing. How does a GAN represent our visual world internally? What causes the artifacts in GAN results? How do architectural choices affect GAN learning? Answering such questions could enable us to develop new insights and better models. In this work, we present an analytic framework to visualize and understand GANs at the unit-, object-, and scene-level. We first identify a group of interpretable units that are closely related to object concepts with a segmentation-based network dissection method. Then, we quantify the causal effect of interpretable units by measuring the ability of interventions to control objects in the output. Finally, we examine the contextual relationship between these units and their surrounding by inserting the discovered object concepts into new images. We show several practical applications enabled by our framework, from comparing internal representations across different layers, models, and datasets, to improving GANs by locating and removing artifact-causing units, to interactively manipulating objects in the scene. We provide open source interpretation tools to help peer researchers and practitioners better understand their GAN models.


page 1

page 3

page 5

page 6

page 11

page 12

page 14

page 16


Visualizing and Understanding Generative Adversarial Networks (Extended Abstract)

Generative Adversarial Networks (GANs) have achieved impressive results ...

Understanding the Role of Individual Units in a Deep Neural Network

Deep neural networks excel at finding hierarchical representations that ...

Seeing What a GAN Cannot Generate

Despite the success of Generative Adversarial Networks (GANs), mode coll...

Generative Adversarial Networks (GANs): What it can generate and What it cannot?

Why are Generative Adversarial Networks (GANs) so popular? What is the p...

Visual Indeterminacy in GAN Art

This paper explores visual indeterminacy as a description for artwork cr...

Scenario generation for market risk models using generative neural networks

In this research, we show how to expand existing approaches of generativ...

Learning to Generate Chairs with Generative Adversarial Nets

Generative adversarial networks (GANs) has gained tremendous popularity ...

Code Repositories


Analytical tools for visualizing and understanding the neurons of a GAN

view repo



view repo