NCP-VAE: Variational Autoencoders with Noise Contrastive Priors

Jyoti Aneja, Alex Schwing, Jan Kautz, Arash Vahdat

Keywords: Variational Autoencoders, Noise Contrastive Estimation, Sampling

Abstract: Variational autoencoders (VAEs) are one of the powerful likelihood-based generative models with applications in various domains. However, they struggle to generate high-quality images, especially when samples are obtained from the prior without any tempering. One explanation for VAEs’ poor generative quality is the prior hole problem: the prior distribution fails to match the aggregate approximate posterior. Due to this mismatch, there exist areas in the latent space with high density under the prior that do not correspond to any encoded image. Samples from those areas are decoded to corrupted images. To tackle this issue, we propose an energy-based prior defined by the product of a base prior distribution and a reweighting factor, designed to bring the base closer to the aggregate posterior. We train the reweighting factor by noise contrastive estimation, and we generalize it to hierarchical VAEs with many latent variable groups. Our experiments confirm that the proposed noise contrastive priors improve the generative performance of state-of-the-art VAEs by a large margin on the MNIST, CIFAR-10, CelebA 64, and CelebA HQ 256 datasets.

One-sentence Summary: We address the prior hole problem in VAEs using an energy-based prior, trained with noise contrastive estimation.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2010.02917/code)

Reviewed Version (pdf): https://openreview.net/references/pdf?id=d4JryoCmSV