Improved Techniques for Training GANs (2016-06-10T00:00:00.000000Z)

TL;DR

This work focuses on two applications of GANs: semi-supervised learning, and the generation of images that humans find visually realistic, and presents ImageNet samples with unprecedented resolution and shows that the methods enable the model to learn recognizable features of ImageNet classes.

Abstract

We present a variety of new architectural features and training procedures that we apply to the generative adversarial networks (GANs) framework. We focus on two applications of GANs: semi-supervised learning, and the generation of images that humans find visually realistic. Unlike most work on generative models, our primary goal is not to train a model that assigns high likelihood to test data, nor do we require the model to be able to learn well without using any labels. Using our new techniques, we achieve state-of-the-art results in semi-supervised classification on MNIST, CIFAR-10 and SVHN. The generated images are of high quality as confirmed by a visual Turing test: our model generates MNIST samples that humans cannot distinguish from real data, and CIFAR-10 samples that yield a human error rate of 21.3%. We also present ImageNet samples with unprecedented resolution and show that our methods enable the model to learn recognizable features of ImageNet classes.

Authors

I. Goodfellow

19 papers

Alec Radford

12 papers

Vicki Cheung

3 papers

TL;DR

Abstract

Authors

References28 items

Semi-Supervised Learning with Generative Adversarial Networks

Pixel-Level Domain Transfer

TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems

Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks

Auxiliary Deep Generative Models

Generating images with recurrent adversarial networks

Rethinking the Inception Architecture for Computer Vision

Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

Unsupervised and Semi-supervised Learning with Categorical Generative Adversarial Networks

Towards Principled Unsupervised Learning

Semi-supervised Learning with Ladder Networks

Distributional Smoothing by Virtual Adversarial Examples

Distributional Smoothing with Virtual Adversarial Training

Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks

Stacked What-Where Auto-encoders

Training generative neural networks via Maximum Mean Discrepancy optimization

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Generative Moment Matching Networks

On distinguishability criteria for estimating generative models

Semi-supervised Learning with Deep Generative Models

Intriguing properties of neural networks

Kernel Measures of Conditional Dependence

A Hilbert Space Embedding for Distributions

Measuring Statistical Dependence with Hilbert-Schmidt Norms

GENERATIVE ADVERSARIAL NETS

Deep Learning

ACTIVITY ANALYSIS OF PRODUCTION AND ALLOCATION

Iterative solution of games by ﬁctitious play

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names