Generative Moment Matching Networks

TL;DR

This paper introduces a new deep generative modeling approach using maximum mean discrepancy (MMD) instead of adversarial training, enabling simpler and effective sample generation, especially when combined with auto-encoders, demonstrated on MNIST and face data.

Contribution

The paper proposes a novel generative model using MMD for training, avoiding complex minimax optimization of GANs, and enhances performance by integrating auto-encoders.

Findings

Achieves high-quality sample generation on MNIST and face datasets.

Simplifies training process compared to GANs by using MMD.

Combining auto-encoders with MMD improves sample quality.

Abstract

We consider the problem of learning deep generative models from data. We formulate a method that generates an independent sample via a single feedforward pass through a multilayer perceptron, as in the recently proposed generative adversarial networks (Goodfellow et al., 2014). Training a generative adversarial network, however, requires careful optimization of a difficult minimax program. Instead, we utilize a technique from statistical hypothesis testing known as maximum mean discrepancy (MMD), which leads to a simple objective that can be interpreted as matching all orders of statistics between a dataset and samples from the model, and can be trained by backpropagation. We further boost the performance of this approach by combining our generative network with an auto-encoder network, using MMD to learn to generate codes that can then be decoded to produce samples. We show that the combination of these techniques yields excellent generative models compared to baseline approaches as measured on MNIST and the Toronto Face Database.