Understanding synergistic correlations and complex information sharing

Fernando Rosas

KU Leuven University, Belgium

Abstract:

The interactions between three or more variables are frequently nontrivial, poorly understood, and yet, are paramount for future advances in fields such as multiuser information theory, neuroscience and genetics. In this talk, we will study the correlations that can exist between a group of random variables by introducing a framework that characterizes the ways in which information can be shared. The framework is based on the novel notion of information synergy, which measures statistical structures that are present in the whole set of variables but not in the any of them individually. Finally, the framework will be applied to several communication scenarios, providing a more intuitive understanding of their fundamental limits.

Biography:

Fernando Rosas studied Music Composition (BSc 2002), Philosophy (Minor 2002), Mathematics (BSc Hons. 2006) and Electric Engineering (MSc and PhD 2008-2012) in the Pontificia Universidad Catolica de Chile. Now he working as a Postdoctoral Researcher in the group of microelectronics -MICAS- in KU Leuven University, Belgium. His researches ranges from the design of energy-efficient wireless acoustic sensor networks to the development of new theoretical notions of shared information. He is also interested in using Information Theory and Thermodynamics to understand the structure of self-organized networks.