MINE: Mutual Information Neural Estimation

TL;DR

MINE introduces a neural network-based method for efficiently estimating mutual information in high-dimensional data, enabling improved applications in generative modeling and supervised classification.

Contribution

The paper presents a scalable, trainable neural estimator for mutual information that is consistent and applicable to various machine learning tasks.

Findings

MINE is linearly scalable with data dimensionality and sample size.

Applying MINE improves the performance of adversarially trained generative models.

Using MINE enhances the effectiveness of the Information Bottleneck in classification tasks.

Abstract

We argue that the estimation of mutual information between high dimensional continuous random variables can be achieved by gradient descent over neural networks. We present a Mutual Information Neural Estimator (MINE) that is linearly scalable in dimensionality as well as in sample size, trainable through back-prop, and strongly consistent. We present a handful of applications on which MINE can be used to minimize or maximize mutual information. We apply MINE to improve adversarially trained generative models. We also use MINE to implement Information Bottleneck, applying it to supervised classification; our results demonstrate substantial improvement in flexibility and performance in these settings.