Quantifying high-order correlations via multivariate extensions of the mutual information
Event details
| Date | 14.11.2018 |
| Hour | 14:00 › 15:00 |
| Speaker | Fernando Rosas Fernando Rosas received the B.A. degree in music composition and philosophy, the B.Sc. degree in mathematics, and the M.S. and Ph.D. degrees in engineering sciences from the Pontifícia Universidad Católica de Chile. He is currently a Marie Sklodowska-Curie Research Fellow in the Department of Mathematics and the Department of Electronic Engineering at Imperial College London. Previously, he worked as a Postdoctoral Researcher at the Department of Electrical Engineering of KU Leuven, and as Research Fellow at the Department of Electrical Engineering of National Taiwan University. His research interests lie in the interface between communication and information theory, complexity science and computational neuroscience. |
| Location | |
| Category | Conferences - Seminars |
The interdependencies that can exist between three or more variables are often nontrivial, poorly understood and, yet, are paramount for advances in many domains of science. In this talk we present an introduction into high-order correlations: statistical dependencies that exist between groups of variables that cannot be reduced to pair-wise interactions. For this, we discuss the most prominent frameworks for computing these properties from data, which provide alternative multivariate generalizations of Shannon's mutual information. Moreover, we introduce a new framework that provides an unified perspective over these previously unrelated approaches, and develop methods for developing a taxonomy of interdependency structures between three or more variables. We illustrate these ideas on two case studies: cellular automata and Baroque music. Finally, we explore implications of these methods to computational neuroscience, and discuss future work.
References:
[1] F. Rosas, P.A. Mediano, M. Ugarte and H.J. Jensen, ``An information-theoretic approach to self-organisation: Emergence of complex interdependencies in coupled dynamical systems'', in Entropy, vol. 20 no. 10: 793, pp.1-25, Sept. 2018.
https://www.mdpi.com/1099-4300/20/10/793
[2] F. Rosas, V. Ntranos, C. J. Ellison, S. Pollin and M. Verhelst, ``Understanding Interdependency Through Complex Information Sharing'', in Entropy, vol. 18 no. 2: 38, pp.1-27, Jan. 2016.
https://www.mdpi.com/1099-4300/18/2/38
Practical information
- Informed public
- Free
Organizer
- IPG Seminar