Adversarial Symmetric Variational Autoencoder

Abstract

A new form of variational autoencoder (VAE) is developed, in which the joint distribution of data and codes is considered in two (symmetric) forms: ($i$) from observed data fed through the encoder to yield codes, and ($ii$) from latent codes drawn from a simple prior and propagated through the decoder to manifest data. Lower bounds are learned for marginal log-likelihood fits observed data and latent codes. When learning with the variational bound, one seeks to minimize the symmetric Kullback-Leibler divergence of joint density functions from ($i$) and ($ii$), while simultaneously seeking to maximize the two marginal log-likelihoods. To facilitate learning, a new form of adversarial training is developed. An extensive set of experiments is performed, in which we demonstrate state-of-the-art data reconstruction and generation on several image benchmark datasets.

Accepted to NIPS 2017

References in corpus (14)

• Auto-Encoding Variational Bayes
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Improved Techniques for Training GANs
• InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
• Energy-based Generative Adversarial Network
• Adversarial Feature Learning
• Adversarially Learned Inference
• Adversarial Autoencoders
• Stein Variational Gradient Descent: A General Purpose Bayesian Inference Algorithm
• Triple Generative Adversarial Nets
• Triangle Generative Adversarial Networks
• Deconvolutional Paragraph Representation Learning
• Symmetric Variational Autoencoder and Connections to Adversarial Learning
• Generative Deep Deconvolutional Learning

Cited by in corpus (4)

• Triangle Generative Adversarial Networks
• Lifelong Teacher-Student Network Learning
• Solving Inverse Problems by Joint Posterior Maximization with Autoencoding Prior
• Learning latent representations across multiple data domains using Lifelong VAEGAN