DVAE++: Discrete Variational Autoencoders with Overlapping Transformations
Abstract
Training of discrete latent variable models remains challenging because passing gradient information through discrete units is difficult. We propose a new class of smoothing transformations based on a mixture of two overlapping distributions, and show that the proposed transformation can be used for training binary latent models with either directed or undirected priors. We derive a new variational bound to efficiently train with Boltzmann machine priors. Using this bound, we develop DVAE++, a generative model with a global discrete prior and a hierarchy of convolutional continuous variables. Experiments on several benchmarks show that overlapping transformations outperform other recent continuous relaxations of discrete latent variables including Gumbel-Softmax (Maddison et al., 2016; Jang et al., 2016), and discrete variational autoencoders (Rolfe 2016).
Cite
@article{arxiv.1802.04920,
title = {DVAE++: Discrete Variational Autoencoders with Overlapping Transformations},
author = {Arash Vahdat and William G. Macready and Zhengbing Bian and Amir Khoshaman and Evgeny Andriyash},
journal= {arXiv preprint arXiv:1802.04920},
year = {2018}
}
Comments
Published as a conference paper at International Conference on Machine Learning (ICML), 2018