Spring 2016

A Framework for Imperfectly Observed Networks

Monday, May 2, 2016 10:30 am11:15 am PDT

Add to Calendar


Calvin Lab

Model a network as an edge-weighted graph, where the (unknown) weight $w_e$ of edge $e$ indicates the frequency of observed interactions, that is over time $t$ we observe a Poisson($t w_e$) number of interactions across edges $e$.  How should we estimate some given statistic of the underlying network?  This leads to wide-ranging and challenging problems.