seminar: Topological properties of self-assembling networks based on information sharing
speaker: Ayse Erzan (Istanbul Technical University)abstract: The establishment of interactions between different components of a system presupposes the satisfaction of a number of constraints, be they physical, chemical, or of some other nature. This calls for sharing a certain amount of information. Even for complex systems with components inhabiting a finite phase space, there is the non-zero probability that certain amount of information is shared, i.e., certain constraints are satisfied, purely at random. We have constructed a string-matching model which is a concrete realisation of this information-theoretic approach, and which can be applied to biological networks among others. A "hidden-variable" version of the model is completely soluble and yields insights into the emergence of complex scaling behaviour from a superposition of classical Erdos-Renyi networks. A reformulation in terms of k-satisfiability problems can be tackled using statistical physics methods developed for stochastic systems.
timetable:
Wed 19 Sep, 10:00 - 11:30, Aula Dini
<< Go back