Categorical Reparameterization with Gumbel-Softmax

by   Eric Jang, et al.
University of Cambridge
Stanford University

Categorical variables are a natural choice for representing discrete structure in the world. However, stochastic neural networks rarely use categorical latent variables due to the inability to backpropagate through samples. In this work, we present an efficient gradient estimator that replaces the non-differentiable sample from a categorical distribution with a differentiable sample from a novel Gumbel-Softmax distribution. This distribution has the essential property that it can be smoothly annealed into a categorical distribution. We show that our Gumbel-Softmax estimator outperforms state-of-the-art gradient estimators on structured output prediction and unsupervised generative modeling tasks with categorical latent variables, and enables large speedups on semi-supervised classification.


page 1

page 2

page 3

page 4


Efficient Marginalization of Discrete and Structured Latent Variables via Sparsity

Training neural network models with discrete (categorical or structured)...

ReCAB-VAE: Gumbel-Softmax Variational Inference Based on Analytic Divergence

The Gumbel-softmax distribution, or Concrete distribution, is often used...

Stochastic gradient descent with gradient estimator for categorical features

Categorical data are present in key areas such as health or supply chain...

Efficient algorithms for modifying and sampling from a categorical distribution

Probabilistic programming languages and other machine learning applicati...

StructureBoost: Efficient Gradient Boosting for Structured Categorical Variables

Gradient boosting methods based on Structured Categorical Decision Trees...

Bridging Discrete and Backpropagation: Straight-Through and Beyond

Backpropagation, the cornerstone of deep learning, is limited to computi...

A t-distribution based operator for enhancing out of distribution robustness of neural network classifiers

Neural Network (NN) classifiers can assign extreme probabilities to samp...

Code Repositories


categorical variational autoencoder using the Gumbel-Softmax estimator

view repo

Please sign up or login with your details

Forgot password? Click here to reset