Narayana Santhanam (University of Hawaii)
2nd floor interaction area
Slow mixing Markov processes
We consider estimation problems in slow mixing Markov processes. These occur, for example, when aggregating information on a topic with polarized views (Republican vs Democrat in politics, for instance) by browsing through the Internet. We will see why our regular intuition fails in slow mixing setups, and outline our recent results on this topic. We will conclude with some open problems we are thinking about now.