Section: Application Domains

Online Social Networks

Our high-level goal here is to help increase the relevance of content accessed by users, through the elaboration of contact and content recommendation mechanisms, as well as incentive mechanisms. The scientific context in which we phrase this goal is that of:

• modeling information propagation in OSN;

• statistical inference problems raised by the search for improved information propagation. In particular these include community detection for contact and content recommendation, and bandit-like algorithms for active learning of given content type at limited “spamming” cost;

• the mechanism design branch of economic theory, which can be leveraged to conceive reward mechanisms meant to incentivize efficient collaborative content filtering by OSN users.