Community Research and Development Information Service - CORDIS

ERC

StrucLim Report Summary

Project ID: 617747
Funded under: FP7-IDEAS-ERC
Country: Hungary

Mid-Term Report Summary - STRUCLIM (Limits of discrete structures)

The goal of the StrucLim research project is the development of a mathematical theory which studies the large scale (approximative) behaviour of very large structures. Examples of such structures include social networks, neural networks, large data sets, fractals and complicated random systems. The results of the StrucLim research group include both theory building and solutions to well known open problems. The group has made significant progress in the limit theory of sparse networks which has been considered to be a major challenge in the area. Using this theory the PI achieved state of the art results in the famous Sidorenko conjecture known from extremal combinatorics. A part of the StrucLim research program was devoted to the study of large random networks in which each node has a fixed number of connections. It has been a classical open problem, related to random matrix theory, to describe the structure of the eigenvectors of such networks. Using limit techniques combined with information theoretic tools the research group found an almost complete answer to this problem. Another component of the StrucLim research plan was the further development of the limit approach to higher order Fourier analysis with a special emphasis on the quadratic case. Numerous research targets were successfully reached in this area.

Contact

Peter Pal Palfy, (Director)
Tel.: +36 1 4838308
Fax: +36 1 4838333
E-mail
Record Number: 191614 / Last updated on: 2016-11-21