Activation Maximization Generative Adversarial Nets

TL;DR

This paper introduces AM-GAN, a novel GAN variant leveraging class label information through activation maximization, providing improved sample quality and diversity, validated by new metrics and state-of-the-art results on CIFAR-10.

Contribution

We propose AM-GAN, a new GAN framework that enhances sample quality by integrating activation maximization and class-aware gradients, along with a novel evaluation metric called AM Score.

Findings

AM-GAN achieves state-of-the-art Inception Score of 8.91 on CIFAR-10.

AM Score correlates better with true sample quality than Inception Score.

AM-GAN outperforms baseline methods in both Inception Score and AM Score.

Abstract

Class labels have been empirically shown useful in improving the sample quality of generative adversarial nets (GANs). In this paper, we mathematically study the properties of the current variants of GANs that make use of class label information. With class aware gradient and cross-entropy decomposition, we reveal how class labels and associated losses influence GAN's training. Based on that, we propose Activation Maximization Generative Adversarial Networks (AM-GAN) as an advanced solution. Comprehensive experiments have been conducted to validate our analysis and evaluate the effectiveness of our solution, where AM-GAN outperforms other strong baselines and achieves state-of-the-art Inception Score (8.91) on CIFAR-10. In addition, we demonstrate that, with the Inception ImageNet classifier, Inception Score mainly tracks the diversity of the generator, and there is, however, no reliable evidence that it can reflect the true sample quality. We thus propose a new metric, called AM Score, to provide a more accurate estimation of the sample quality. Our proposed model also outperforms the baseline methods in the new metric.