MMCGAN: Generative Adversarial Network with Explicit Manifold Prior

Abstract

Generative Adversarial Network(GAN) provides a good generative framework to produce realistic samples, but suffers from two recognized issues as mode collapse and unstable training. In this work, we propose to employ explicit manifold learning as prior to alleviate mode collapse and stabilize training of GAN. Since the basic assumption of conventional manifold learning fails in case of sparse and uneven data distribution, we introduce a new target, Minimum Manifold Coding (MMC), for manifold learning to encourage simple and unfolded manifold. In essence, MMC is the general case of the shortest Hamiltonian Path problem and pursues manifold with minimum Riemann volume. Using the standardized code from MMC as prior, GAN is guaranteed to recover a simple and unfolded manifold covering all the training data. Our experiments on both the toy data and real datasets show the effectiveness of MMCGAN in alleviating mode collapse, stabilizing training, and improving the quality of generated samples.

References in corpus (12)

• Adam: A Method for Stochastic Optimization
• UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Spectral Normalization for Generative Adversarial Networks
• SGDR: Stochastic Gradient Descent with Warm Restarts
• Energy-based Generative Adversarial Network
• Adversarially Learned Inference
• Generalization and Equilibrium in Generative Adversarial Nets (GANs)
• Many Paths to Equilibrium: GANs Do Not Need to Decrease a Divergence At Every Step
• Improving GANs Using Optimal Transport