2018-11-19 14:30:00 2018-11-19 15:30:00 America/New_York Research Seminar Series - Bernat Corominas-Murtra Research Associate, Institute for Science & Technology Austria FRNY B124

November 19, 2018

Research Seminar Series - Bernat Corominas-Murtra

Event Date: November 19, 2018
Time: 2:30 - 3:30 PM
Location: FRNY B124
Contact Name: Erin Gough
Contact Phone: 765-496-0606
Contact Email: egough@purdue.edu
Open To: all
Priority: No
School or Program: Industrial Engineering
College Calendar: Show
Research Associate, Institute for Science & Technology Austria

"Hierarchy, levels of organization and the different cores of complex networks"

ABSTRACT

The existence of different levels of organization is a key feature of both natural and artificial complex systems. Disentangling the different layers of organization is, therefore, one of the central targets of contemporary research on the topic. Achieving a sound and robust network decomposition mechanism would provide extraordinary information about the structure of the system. The task is far from obvious, but, to start with, we must acknowledge the role of two main complementary concepts that are central to this framework: Hierarchy and nesting. In this talk we will go in depth into both concepts and how they shed light in the study of complex networks. We will, first, present a deeply grounded formal approach resulting from the convergence of order theory, information theory, theoretical morphology and network theory that allows constructing a 3D morphospace of hierarchies and hence comparing the hierarchical organization of ecological, cellular, technological, and social networks. Interestingly, the space of observed hierarchies is much smaller than the space of possible hierarchies, indicating that strong organizational and evolutionary bottlenecks are at work when shaping the structure of interlinked systems. Second, we will explore the structure of the core of the network, through to so-called nesting algorithms or core-extraction algorithms —among them, the well-known K-core or the recently proposed M-core. These methods enable us to decompose the network in a Russian doll-like fashion, putting the emphasis on various topological quantities, therefore, highlighting different features of the core of the network. Special treatment will be reserved to the “Weak core”, based on a delicate technique of core extraction that identifies the communities that glue the potentially different, highly connected clusters that populate the core of the network. Surprising results arise from the application of the nesting algorithms to real systems, such as virtual societies, indicating, for example, that the communities of individuals acting as connectors —as opposed to the hubs— play a central role in keeping the cohesiveness of the social elites.

 

Photo of Bernat Corominas-MurtraBIO

The emergence of complexity is one of the most fascinating phenomena of nature. From human language to stem cell biology, biological evolution or the development of technology, the emergence of complex order seems to defy the laws of physics. As a researcher I try to disentangle small bits of this formidable problem, following an approach that integrates non equilibrium physics, stochastic processes, complex networks and information theory. I address specific problems, such as the nature of hierarchy, the emergence of scaling or the properties of systems with non-stable phase spaces.

 

Logo of 150 Giant Leaps