VQ-DRAW: A Sequential Discrete VAE
arXiv:2003.01599
Abstract
In this paper, I present VQ-DRAW, an algorithm for learning compact discrete representations of data. VQ-DRAW leverages a vector quantization effect to adapt the sequential generation scheme of DRAW to discrete latent variables. I show that VQ-DRAW can effectively learn to compress images from a variety of common datasets, as well as generate realistic samples from these datasets with no help from an autoregressive prior.
References in corpus (12)
- Adam: A Method for Stochastic Optimization
- Categorical Reparameterization with Gumbel-Softmax
- Neural Discrete Representation Learning
- Progressive Growing of GANs for Improved Quality, Stability, and Variation
- DRAW: A Recurrent Neural Network For Image Generation
- Training Deep Nets with Sublinear Memory Cost
- Neural Variational Inference and Learning in Belief Networks
- Variational inference for Monte Carlo objectives
- Generating Diverse High-Fidelity Images with VQ-VAE-2
- Towards Conceptual Compression
- Memory-Efficient Backpropagation Through Time
- Black Box Variational Inference