Symmetric Variational Autoencoder and Connections to Adversarial Learning
This work addresses the unification of VAE and adversarial learning methods for machine learning researchers, but it appears incremental as it builds on existing techniques.
The authors proposed a symmetric variational autoencoder (sVAE) based on symmetric Kullback-Leibler divergence, which connects to adversarial learning and unifies VAE and adversarial techniques, with experiments validating its utility.
A new form of the variational autoencoder (VAE) is proposed, based on the symmetric Kullback-Leibler divergence. It is demonstrated that learning of the resulting symmetric VAE (sVAE) has close connections to previously developed adversarial-learning methods. This relationship helps unify the previously distinct techniques of VAE and adversarially learning, and provides insights that allow us to ameliorate shortcomings with some previously developed adversarial methods. In addition to an analysis that motivates and explains the sVAE, an extensive set of experiments validate the utility of the approach.