Tuesday, April 7, 2020 at 12:00 pm ET
Seminar will be remote via Zoom
Registration required at:
Nonidentifiability andnonparametric random graphhypothesis testing
Joshua Agterberg(JHU – MINDS Fellow)
Abstract: Hypothesis testing for randomgraphs is a relatively new field, and existing methods have focused mostly onspecific random graph models, such as the stochastic block model and itsvariants. However, there are not currently many consistent hypothesis tests forgeneral low-rank random graphs, which suffer from a uniqueform ofnonidentifiability. In this talk I will be discussing two types ofnonidentifiability that arise when using spectral methods to study randomgraphs, and I will show how limiting results for one type of nonidentifiabilitycan be used to find a consistent nonparametric hypothesis test for equality ofdistribution for general low-rank random graphs.
Bio: Joshua Agterberg is a PhD student inApplied Mathematics and Statistics at Johns Hopkins University where he isadvised by Carey Priebe. His research interests include statistical inferencefor random graphs, kernel methods, spectral perturbation theory,high-dimensional statistics, and nonparametric statistics, with an emphasis onthe interplay between optimization, statistics, and matrix analysis. This year,he is a MINDS fellow, a Counselman fellow, and an AMS apprentice teachingfellow.