Skip to main content.
Advanced search >
<< Back to previous page Print

<< Thursday, February 21, 2013 >>

Remind me

Tell a friend

Add to my Google calendar (bCal)

Download to my calendar

Bookmark and ShareShare

Information dynamics in online social networks

Seminar: Departmental | February 21 | 3-4 p.m. | Soda Hall, 430/438: Wozniak Lounge

Lada Adamic, Univ. of Michigan

Electrical Engineering and Computer Sciences (EECS)

Social networks, especially online social networks, are driven by information sharing. These exchanges can be surprisingly revealing of the underlying network dynamics. In online question and answer forums, who answers whom reveals not only their relative expertise, but the subjectivity of the topic. Furthermore, using just the sequence and network position of messages exchanged within a community, one can partly infer the diversity and novelty of their information content. How individuals come to share information in the first place has been obscured by a confound between homophily and social influence. In a large-scale Facebook experiment we tease the two factors apart, finding that while strong ties are individually more influential, collectively it is the weak ties that wield more influence and provide more diverse information exposure. What makes diffusion in networks even more interesting is that the information dynamics extends beyond mere spread. As individuals select, modify and replicate transmitted content, the content itself undergoes a process akin to evolution.