LGMLMay 22, 2018

Implicit Reparameterization Gradients

arXiv:1805.08498v4263 citations
Originality Incremental advance
AI Analysis

This provides a solution for training latent variable models with distributions previously incompatible with standard gradient techniques, though it is incremental in extending existing methods.

The paper tackles the limitation of the reparameterization trick for continuous distributions like Gamma and Beta by introducing an implicit differentiation method, resulting in faster and more accurate gradient estimators as shown in experiments.

By providing a simple and efficient way of computing low-variance gradients of continuous random variables, the reparameterization trick has become the technique of choice for training a variety of latent variable models. However, it is not applicable to a number of important continuous distributions. We introduce an alternative approach to computing reparameterization gradients based on implicit differentiation and demonstrate its broader applicability by applying it to Gamma, Beta, Dirichlet, and von Mises distributions, which cannot be used with the classic reparameterization trick. Our experiments show that the proposed approach is faster and more accurate than the existing gradient estimators for these distributions.

Code Implementations2 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes