Activation Maximization Generative Adversarial Nets

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.

Accepted as a conference paper on ICLR 2018

References in corpus (30)

• Conditional Generative Adversarial Nets
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Generative Adversarial Text to Image Synthesis
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Densely Connected Convolutional Networks
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• Improved Training of Wasserstein GANs
• Improved Techniques for Training GANs
• NIPS 2016 Tutorial: Generative Adversarial Networks
• Energy-based Generative Adversarial Network
• BEGAN: Boundary Equilibrium Generative Adversarial Networks
• Adversarially Learned Inference
• Autoencoding beyond pixels using a learned similarity metric
• f-GAN: Training Generative Neural Samplers using Variational Divergence Minimization
• A note on the evaluation of generative models
• Unsupervised and Semi-supervised Learning with Categorical Generative Adversarial Networks
• SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient
• Synthesizing the preferred inputs for neurons in neural networks via deep generator networks
• Neural Photo Editing with Introspective Adversarial Networks
• Mode Regularized Generative Adversarial Networks
• LR-GAN: Layered Recursive Generative Adversarial Networks for Image Generation
• Plug & Play Generative Networks: Conditional Iterative Generation of Images in Latent Space
• Least Squares Generative Adversarial Networks
• Generative Visual Manipulation on the Natural Image Manifold
• Learning to Draw Samples: With Application to Amortized MLE for Generative Adversarial Learning
• Deep Directed Generative Models with Energy-Based Probability Estimation
• Stacked Generative Adversarial Networks
• Class-Splitting Generative Adversarial Networks
• McGan: Mean and Covariance Feature Matching GAN

Cited by in corpus (9)

• Towards the Automatic Anime Characters Creation with Generative Adversarial Networks
• DiffWave: A Versatile Diffusion Model for Audio Synthesis
• The Unusual Effectiveness of Averaging in GAN Training
• TimeMatch: Unsupervised Cross-Region Adaptation by Temporal Shift Estimation
• Understanding the Effectiveness of Lipschitz-Continuity in Generative Adversarial Nets
• The Partial Response Network: a neural network nomogram
• Guiding the One-to-one Mapping in CycleGAN via Optimal Transport
• Quantifying point cloud realism through adversarially learned latent representations
• Diagnostic Visualization for Deep Neural Networks Using Stochastic Gradient Langevin Dynamics