Adversarial Symmetric Variational Autoencoder
This work addresses the challenge of enhancing generative modeling for image data, representing an incremental advancement in VAE architectures.
The authors tackled the problem of improving variational autoencoders by developing a symmetric VAE that minimizes the symmetric KL divergence between joint distributions of data and codes, while maximizing marginal log-likelihoods, and demonstrated state-of-the-art reconstruction and generation on image datasets.
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.