Mathematics Colloquia and Seminars

Return to Colloquia & Seminar listing

Modeling interaction and information updating in networks

Mathematical Physics & Probability

Speaker: David Aldous, UC Berkeley
Location: 2112 MSB
Start time: Wed, May 11 2011, 4:10PM

The vague phrase "information flow through networks" could be used in many contexts. Envisaging social networks, here is one possible high level abstraction. Each pair of people interact at random times of arbitrary given rate; when they interact, they update their information according to some given rule. This framework encompasses scores of models and thousands of papers, in fields such as statistical physics; epidemic theory; broadcast algorithms on graphs; ad hoc networks; social learning theory. The talk will present snapshots of 8 models. The usual names of these models are dull and uninformative, so for this talk let me invent more evocative names: Hot Potato, Pandemic, Leveller, Pothead, Deference, Fashionista, Gordon Gecko, and Preserving Principia. (Note: the talk is intended to be accessible to anyone from EECS interested in probabilistic aspects of networks).