BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//UIC
BEGIN:VEVENT
UID:2019092109525520190916T15000020190916T1550005d869bb75abca@uic.edu
CATEGORIES:MEETING
STATUS:TENTATIVE
DTSTAMP:20190920T075056
DTSTART:20190916T150000
DTEND:20190916T155000
SUMMARY:Combinatorics and Probability Seminar: Multivariate Limit Theorems and Large Variance from Zero Free Regions, by Marcus Michelen
DESCRIPTION:Marcus Michelen (UIC): Multivariate Limit Theorems and Large Variance from Zero Free Regions We prove a multivariate central limit theorem for random variables $(X_1,\ldots,X_d)$ that are \emph{strong Rayleigh}, a higher-dimensional analogue of random variables with real-rooted generating function. The strong Rayleigh property can be thought of as a notion of negative dependence and this limit theorem was conjectured by Ghosh, Liggett and Pemantle. I will present a short history of the strong Rayleigh property and demonstrate how this multivariate limit theorem can be reduced to a univariate limit theorem discussed in the previous week. Back in the univariate setting, let $X \in \{0,\ldots,n\}$ be a random variable and $f_X$ be its probability generating function. We show that if the roots of $f_X$ have argument bounded away from $0$, then the variance of $X$ must be remarkably large. This takes the form of a lower bound that is sharp up to constants, and is proved using similar tools used to prove the limit theorems of the previous week. This talk is based on joint work with Julian Sahasrabudhe. Please click here to make changes to, or delete, this seminar announcement.
LOCATION:612 SEO Chicago IL
CLASS:PRIVATE
END:VEVENT
END:VCALENDAR