[Ml-stat-talks] Fwd: Wilks Statistics Seminar: Po-ling Loh, Today, February 12th @ 12:30pm, Sherrerd Hall 101
Barbara Engelhardt
bee at princeton.edu
Fri Feb 12 10:29:57 EST 2016
Talk of interest today.
---------- Forwarded message ----------
From: Carol Smith <carols at princeton.edu>
Date: Fri, Feb 12, 2016 at 10:11 AM
Subject: Wilks Statistics Seminar: Po-ling Loh, Today, February 12th @
12:30pm, Sherrerd Hall 101
To: wilks-seminar at princeton.edu
*** Wilks Statistics Seminar ***
DATE: Today, February 12, 2016
TIME: 12:30pm
LOCATION: Sherrerd Hall, room 101
SPEAKER: Po-ling Loh, University of Pennsylvania
TITLE: On Centrality in Random Growing Trees: Confidence Sets for Source
Estimators and Persistence
ABSTRACT: We discuss recent work regarding random growing trees. A common
theme is the evolution of the most central node(s) in such trees and how
they behave probabilistically as the tree grows. Our first result concerns
source estimation in a diffusion spreading over a regular tree. We show
that it is possible to construct confidence sets for the diffusion source
with size independent of the number of infected nodes. Our estimators are
motivated by analogous results in the literature concerning identification
of the root node in preferential attachment and uniform attachment trees;
at the core of our proofs is a probabilistic analysis of Polya urns
corresponding to the number of uninfected neighbors in specific subtrees of
the infection tree. We then turn to the problem of persistence, and
demonstrate that the aforementioned confidence sets have the property of
persistence (i.e., settling down after finitely many steps), with
probability 1. Our theory holds for regular diffusion trees, as well
uniform attachment and linear and sublinear preferential attachment trees.
--
Barbara E Engelhardt
Assistant Professor
Department of Computer Science
Center for Statistics and Machine Learning
Princeton University
http://www.cs.princeton.edu/~bee
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cs.princeton.edu/pipermail/ml-stat-talks/attachments/20160212/c402c3fd/attachment.html>
More information about the Ml-stat-talks
mailing list