Improving Variational Auto-Encoders using convex combination linear Inverse Autoregressive Flow

Abstract

In this paper, we propose a new volume-preserving flow and show that it performs similarly to the linear general normalizing flow. The idea is to enrich a linear Inverse Autoregressive Flow by introducing multiple lower-triangular matrices with ones on the diagonal and combining them using a convex combination. In the experimental studies on MNIST and Histopathology data we show that the proposed approach outperforms other volume-preserving flows and is competitive with current state-of-the-art linear normalizing flow.

Published at Benelearn 2017 (Eindhoven, the Netherlands)

References in corpus (3)

• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• NICE: Non-linear Independent Components Estimation
• Predicting distributions with Linearizing Belief Networks

Cited by in corpus (6)

• An Introduction to Variational Autoencoders
• Normalizing Flows: An Introduction and Review of Current Methods
• Improving Variational Autoencoders for New Physics Detection at the LHC with Normalizing Flows
• Unbiased Gradient Estimation for Variational Auto-Encoders using Coupled Markov Chains
• Structured Dropout Variational Inference for Bayesian Neural Networks
• General Invertible Transformations for Flow-based Generative Modeling