Modeling and Optimizing Diffusion of Information in Social Networks
Online Social Networks (OSN) have been shown to have benefits as a medium for fast, widespread information dissemination. They provide fast access to large scale data in little-to-no cost. Local interactions between friends in a network can create global effects of promoting an idea or a product and making them popular in large-scale. Why are some products/ideas vastly more successful than others? Are the reasons repeatable, i.e. are there any properties we can extract to recreate the same successful result for the dissemination of another product/idea? These are questions that have been puzzling many professions for ages. In order to answer those questions 1) We need to have an accurate model of human behavior, information diffusion in OSNs 2) scalable algorithms that can solve optimization problems on these models.
We are working on new models that capture the behavior that is observed in todays OSNs. We study various models that capture the existence of competing campaigns in social networks, existence of users of different roles. We also work on scalable algorithms to solve problems such as limiting the spread of misinformation in OSNs.