Visualizing and Understanding Generative Adversarial Networks (Extended Abstract)

by   David Bau, et al.
The Chinese University of Hong Kong

Generative Adversarial Networks (GANs) have achieved impressive results for many real-world applications. As an active research topic, 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 concepts with a segmentation-based network dissection method. 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 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 will open source our interactive tools to help researchers and practitioners better understand their models.


page 1

page 2

page 3

page 4


GAN Dissection: Visualizing and Understanding Generative Adversarial Networks

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

GAN You Do the GAN GAN?

Generative Adversarial Networks (GANs) have become a dominant class of g...

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...

Towards Co-Creative Generative Adversarial Networks for Fashion Designers

Originating from the premise that Generative Adversarial Networks (GANs)...

GAN Lab: Understanding Complex Deep Generative Models using Interactive Visual Experimentation

Recent success in deep learning has generated immense interest among pra...

Simulation of an Elevator Group Control Using Generative Adversarial Networks and Related AI Tools

Testing new, innovative technologies is a crucial task for safety and ac...

Please sign up or login with your details

Forgot password? Click here to reset