The Generalized Reparameterization Gradient

Abstract

The reparameterization gradient has become a widely used method to obtain Monte Carlo gradients to optimize the variational objective. However, this technique does not easily apply to commonly used distributions such as beta or gamma without further approximations, and most practical applications of the reparameterization gradient fit Gaussian distributions. In this paper, we introduce the generalized reparameterization gradient, a method that extends the reparameterization gradient to a wider class of variational distributions. Generalized reparameterizations use invertible transformations of the latent variables which lead to transformed distributions that weakly depend on the variational parameters. This results in new Monte Carlo gradients that combine reparameterization gradients and score function gradients. We demonstrate our approach on variational inference for two complex probabilistic models. The generalized reparameterization is effective: even a single sample from the variational distribution is enough to obtain a low-variance gradient.

16 pages, 15 figures, NIPS version

References in corpus (7)

• Auto-Encoding Variational Bayes
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Automatic Differentiation Variational Inference
• Automated Variational Inference in Probabilistic Programming
• Automatic Variational Inference in Stan
• Stochastic gradient variational Bayes for gamma approximating distributions
• Black-Box Policy Search with Probabilistic Programs

Cited by in corpus (15)

• Stochastic Gradient Descent as Approximate Bayesian Inference
• Variational Sequential Monte Carlo
• Tighter risk certificates for neural networks
• A Tutorial on Deep Latent Variable Models of Natural Language
• Semi-Implicit Variational Inference
• Adversarial Message Passing For Graphical Models
• Dirichlet Variational Autoencoder for Text Modeling
• Structured Black Box Variational Inference for Latent Time Series Models
• Reparameterization Gradient for Non-differentiable Models
• Fast Inference for Intractable Likelihood Problems using Variational Bayes
• Joint Emotion Label Space Modelling for Affect Lexica
• Perturbative Black Box Variational Inference
• Variational Inference In Pachinko Allocation Machines
• Simplified Stochastic Feedforward Neural Networks
• Importance Sampled Stochastic Optimization for Variational Inference