Online social search brings forth a new way to harness the Internet for answers. However, the personal and often sensitive information is unwittingly exposed to others when a person looks for an expert via the underlying social network. In this paper, we propose a model in which a node's behavior of looking for an expert is adjusted by his awareness of the potential expertise of his contacts. We derive the optimal distribution of nodes' awareness level that minimizes the system's privacy exposure, and prove that it corresponds to the unique Nash equilibrium. Our analysis shows that the optimal distribution over a posed question is inversely proportional to the square root of the corresponding expertise density.
展开▼