The Generalized Reparameterization Gradient

by   Francisco J. R. Ruiz, et al.

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.


page 15

page 16


Stochastic gradient variational Bayes for gamma approximating distributions

While stochastic variational inference is relatively well known for scal...

Quasi-Monte Carlo Variational Inference

Many machine learning problems involve Monte Carlo gradient estimators. ...

Generalized Transformation-based Gradient

The reparameterization trick has become one of the most useful tools in ...

Implicit Reparameterization Gradients

By providing a simple and efficient way of computing low-variance gradie...

Error bounds for overdetermined and underdetermined generalized centred simplex gradients

Using the Moore–Penrose pseudoinverse, this work generalizes the gradien...

A Spectral Approach to Gradient Estimation for Implicit Distributions

Recently there have been increasing interests in learning and inference ...

Reducing Reparameterization Gradient Variance

Optimization with noisy gradients has become ubiquitous in statistics an...

Please sign up or login with your details

Forgot password? Click here to reset