Anna Melkonyan: Contagion processes in weighted visual networks

BCCN Berlin / Technische Universität Berlin

Abstract

 

Coordination in human and social animal groups depends on the underlying network structure of the group. Since social interactions often rely on information received from sensory organs, we decided to use a newly-developed network that imitates the working of a visual system. Most animals, including humans, rely on visual information as a rapid and efficient transmission method. Unlike metric and topological networks, connections in visual network are based on visibility of the neighboring nodes, meaning that the occlusions of the visual field have to be considered, as well as the distance to the neighbor. As a result of that, on one hand in dense networks fewer neighbors are visible due to the occlusion, while in in sparse networks fewer neighbors are recognized due to large distance. The number and strength of these links also affects the information propagation within visual network.

Static properties and information transmission via simple and complex contagion have already been investigated in unweighted visual networks. Hence, here we look into a weighted visual networks, where the strength of the link depends on the distance between nodes. We study how the distance and thus density influence the clustering, in strength, average shortest distance, as well as the transmission of visual information via simple and complex contagion. This additional investigation of visual networks could be of interest fro understanding spread of visual information as well as coordinated movement and behavior in visual animal groups.

 

Additional Information

Master Thesis Defense

 

Organized by

Prof. Dr. Pawel Romanczuk   & Prof. Dr. Richard Kempter   / Lisa Velenosi