Echo chambers describe situations where one is exposed only to opinions that agree with their own. In this talk we will discuss the phenomenon of political echo chambers in social media. We identify the different components in the phenomenon and characterize users based on their behavior with respect to content production and consumption. Among other findings, we observe that users who try to bridge the echo chambers have to pay a "price of bipartisanship." We then discuss ideas for combating echo chambers. We first present a model for learning ideological-leaning factors, of social-media users and media sources, in a joint latent space. The model space can be used to develop exploratory and interactive interfaces that can help users to diffuse their information filter bubble. Second we present an influence-based approach for balancing the information exposure of users in the social network.

Invited talk by Aristides Gionis.