# Adversarial Symmetric Variational Autoencoder

arXiv: https://arxiv.org/abs/1711.04915 · HTML: https://mlsanity.com/paper/1711.04915/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Nov 2017.

## Built on by (most cited first, 11 of 11)

- [Enhancing Diffusion-Based Image Synthesis with Robust Classifier Guidance](https://mlsanity.com/paper/2208.08664/) (2022)
- [Solving Inverse Problems by Joint Posterior Maximization with Autoencoding Prior](https://mlsanity.com/paper/2103.01648/) (2021)
- [Lifelong Teacher-Student Network Learning](https://mlsanity.com/paper/2107.04689/) (2021)
- [Lifelong Mixture of Variational Autoencoders](https://mlsanity.com/paper/2107.04694/) (2021)
- [Symmetric Equilibrium Learning of VAEs](https://mlsanity.com/paper/2307.09883/) (2023)
- [A Unified f-divergence Framework Generalizing VAE and GAN](https://mlsanity.com/paper/2205.05214/) (2022)
- [Symmetric Wasserstein Autoencoders](https://mlsanity.com/paper/2106.13024/) (2021)
- [RENs: Relevance Encoding Networks](https://mlsanity.com/paper/2205.13061/) (2022)
- [Stabilizing Subject Transfer in EEG Classification with Divergence Estimation](https://mlsanity.com/paper/2310.08762/) (2023)
- [Likelihood-Free Variational Autoencoders](https://mlsanity.com/paper/2504.17622/) (2025)
- [Breaking the Illusion: Consensus-Based Generative Mitigation of Adversarial Illusions in Multi-Modal Embeddings](https://mlsanity.com/paper/2511.21893/) (2025)

## Abstract (arXiv)

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.