Computing the Unique Information
Pradeep Kumar Banerjee, Johannes Rauh, and Guido Montúfar
Contact the author: Please use for correspondence this email.
Submission date: 09. Nov. 2017
published in: IEEE international symposium on information theory (ISIT) from June 17 to 22, 2018 at the Talisa Hotel in Vail, Colorado, USA
Piscataway, NY : IEEE, 2018. - P. 141 - 145
DOI number (of the published article): 10.1109/ISIT.2018.8437757
Keywords and phrases: Positive information decomposition, mutual information, alternating divergence minimization
Download full preprint: PDF (268 kB)
Link to arXiv: See the arXiv entry of this preprint.
Given a set 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, complementary, and shared 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 we evaluate its performance against other techniques.