Computing the Unique Information

Abstract

Given a pair of predictor variables and a response variable, how much information do the predictors have about the response, and how is this information distributed between unique, redundant, and synergistic components? Recent work has proposed to quantify the unique component of the decomposition as the minimum value of the conditional mutual information over a constrained set of information channels. We present an efficient iterative divergence minimization algorithm to solve this optimization problem with convergence guarantees and evaluate its performance against other techniques.

To appear in 2018 IEEE International Symposium on Information Theory (ISIT); 18 pages; 4 figures, 1 Table; Github link to source code: https://github.com/infodeco/computeUI

References in corpus (1)

• Reconsidering unique information: Towards a multivariate information decomposition

Cited by in corpus (8)

• A Novel Approach to the Partial Information Decomposition
• A comparison of partial information decompositions using data from real and simulated layer 5b pyramidal cells
• BROJA-2PID: A robust estimator for bivariate partial information decomposition
• Unique Informations and Deficiencies
• Partial Information Decomposition via Deficiency for Multivariate Gaussians
• Unique Information and Secret Key Decompositions
• MAXENT3D_PID: An Estimator for the Maximum-entropy Trivariate Partial Information Decomposition
• Explicit Formula for Partial Information Decomposition