Colloquium: Epidemic Spreading in Social Networks by Dr Kitsak/December 2nd 2010

The colloquium topic was information diffusion in networks. Dr. Kitsak mentioned about diseases that spread very fast and killed lots of people in the past and recent years also. The most important point here is to understand how information (or disease) spreads based on network structure and what strategies can be applied to overcome the diseases very fast.There are many things that can influence the spread of information within a network. The three main factors in diffusion are network structure (degree distribution, clustering, connected components, etc.), strength of ties (frequency of communication and strength of influence), and spreading agents. A strong tie can be defined by frequent contact, affinity, or having many mutual contacts.

Dr. Kitsak also mentioned three key models for infection. These are SI, SIS and SIR. SIR stands for Susceptible -> Infected -> Removed/Recovered. Each node in the network can be classified by one of these terms in the SIR model. If a node is susceptible, then it can be infected. An infected node is self-explanatory. A removed node means it was infected and removed from the network. Or, a node can recover and build a natural immunity. The SI model is simply Susceptible -> Infected. The definitions remain the same, but a node is unable to recover and therefore remains infected. The SIS model is Susceptible ->Infected -> Susceptible. One of the important network structure that disease can be spread is worldwide airport network. Dr. Kitsak also mentioned about network immunization. One vaccine strategy may be instead of vaccinating all people vaccinate a threshold part so the community can be divided into clusters and virus can't jump to clusters. Dr. Kitsak also told that our community is scale free network.

So who are the most influential spreader in the network?K-shell is the most robust spreading efficiency indicator. So if we want a disease spread quickly we should find different K-shells at different places in the world.As a conclusion Dr. Kitsak pointed out these factors:

1-Almost no epidemic threshold in Scale-Free Networks.
2- Efficient Immunization Strategy: Immunize at least critical fraction fc of nodes so that only isolated clusters of susceptible individuals remain.
3- Immunization strategy is not reciprocal to spreading strategy.
4-Influential spreaders (not necessarily hubs) occupy the innermost k-cores.