Multivariate Information Theory: Difficulties and Recent Progress

Document Type

Presentation Abstract

Presentation Date

2-12-2024

Abstract

Information theory has been widely applied within the sciences. Much of the appeal comes from it providing a substraight-agnostic method of quantifying the interactions within a system. In the bivariate case, many measures have reasonably robust interpretations stemming from their status as solutions to operational problems. In the multivariate case, however, most measures are extensions of their bivariate cousin, but lack the operational interpretation. In this talk, we review many multivariate information measures that have been employed in data analysis, their interpretational difficulties, and some recent progress in producing interpretable, robust multivariate measures.

Additional Details

February 12, 2024 at 3:00 p.m. Math 103

This document is currently not available here.

Share

COinS