Stochastics Seminar
Spring 2022 Friday 3:004:00 PM
Room for inperson: LCB 215
Zoom information:
Meeting ID: 910 5244 8043
Passcode: Email the organizers
Join the seminar mailing list
Date  Speaker  Title (click for abstract, if available) 

Friday, January 21st 
Xuan Wu
University of Chicago 
In this talk, we will discuss the LUE, focusing on the scaling limits. On the hardedge side, we construct the $\alpha$Bessel line ensemble for all $\alpha \in \mathbb{N}_0$. This novel Gibbsian line ensemble enjoys the $\alpha$squared Bessel Gibbs property. Moreover, all $\alpha$Bessel line ensembles can be naturally coupled together in a Bessel field, which enjoys rich integrable structures. We will also talk about work in progress on the softedge side, where we expect to have the Airy field as the scaling limit. This talk is based on joint works with Lucas Benigni, PeiKen Hung, and Greg Lawler. 
Friday, January 28th 
Evan Sorensen
UWMadison 
In the last 1015 years, Busemann functions have been a key tool for studying semiinfinite geodesics in planar first and lastpassage percolation. We study Busemann functions in the semidiscrete Brownian lastpassage percolation (BLPP) model and use these to derive geometric properties of the full collection of semiinfinite geodesics in BLPP. This includes a characterization of uniqueness and coalescence of semiinfinite geodesics across all asymptotic directions. To deal with the uncountable set of points in BLPP, we develop new methods of proof and uncover new phenomena, compared to discrete models. For example, for each asymptotic direction, there exists a random countable set of initial points out of which there exist two semiinfinite geodesics in that direction. Further, there exists a random set of points, of Hausdorff dimension 1/2, out of which, for some random direction, there are two semiinfinite geodesics that split from the initial point and never come back together. We derive these results by studying variational problems for Brownian motion with drift. Based on joint work with Timo Seppalainen. 
Friday, February 4th 
Zhipeng Liu
University of Kansas 
In the recent twenty years, there has been a huge development in understanding the universal law behind a family of 2d random growth models, the socalled KardarParisiZhang (KPZ) universality class. Especially, limiting distributions of the height functions are identified for a number of models in this class. Different from the height functions, the geodesics of these models are much less understood. There were studies on the qualitative properties of the geodesics in the KPZ universality class very recently, but the precise limiting distributions of the geodesic locations remained unknown. In this talk, we will discuss our recent results on the onepoint distribution of the geodesic of a representative model in the KPZ universality class, the directed last passage percolation with iid exponential weights. We will give the explicit formula of the onepoint distribution of the geodesic location joint with the last passage times, and its limit when the parameters go to infinity under the KPZ scaling. The limiting distribution is believed to be universal for all the models in the KPZ universality class. We will further give some applications of our formulas. 
Friday, February 11th 
Michael Damron
Georgia Tech 
Given edgelengths (t_e) assigned to the edges of Z^d for d \geq 2, each vertex draws a directed edge to its closest neighbor to form the ``nearest neighbor'' graph. NandaNewman studied these graphs when the t_e's are i.i.d. and continuously distributed and showed that the undirected version has only finite components, with size distribution whose tail decays like 1/n!. I will discuss recent work with B. Bock and J. Hanson in which the t_e's are only assumed to be translationinvariant and distinct. Here, we can show that there are no doublyinfinite paths and completely characterize the set of possible graphs. In particular, for d=2, the number of infinite components is either 0,1, or 2, and for d \geq 3, it can be any nonnegative integer. I will also mention relations to both geodesic graphs from firstpassage percolation and the coalescing walk model of ChaikaKrishnan. 
Friday, February 18th 
B. Cooper Boniece
University of Utah 
The diffusion coefficient of a Lévy process — the variance of its Gaussian part — can be viewed as an elementary case of a quantity known as the integrated volatility, whose estimation in the highfrequency sampling regime has been a topic of active research for the past 2+ decades. However, even in the fundamental case of a univariate Lévy process, there remain significant practical challenges to efficient estimation of this quantity when sample paths exhibit extreme jump behavior.

Friday, February 25th 
Ahmed BouRabee
University of Chicago 
Pareto peeling describes a family of algorithms for multidimensional sorting. I will discuss joint work with Peter Morfe in which we show that Pareto peeling of large random point clouds approximates the solution of a HamiltonJacobi equation. The limiting equations are closely related to the longest chain problem. 
Friday, March 4th 
Axel Saenz Rodriguez
Oregon State University 
I will present exact probability formulas, obtained via the Bethe Ansatz, for interacting particle systems on a onedimensional lattice. The Bethe Ansatz, originally introduced in 1931 by Hans Bethe, is a method to diagonalize some linear operators that have certain symmetries; I will make this more precise in the talk. In particular, I will focus on two models: the HeisenbergIsing spin1/2 chain (aka XXZ spin1/2 chain) and the asymmetric simple exclusion process (ASEP). I will present some wellknown results from the literature for the ASEP and some recent results based on joint work with C.A. Tracy for the XXZ. 
Friday, March 18th 
Sayan Das
Columbia University 
The KPZ equation is a fundamental stochastic PDE that can be viewed as the logpartition function of continuum directed random polymer (CDRP). In this talk, we will first focus on the fractal properties of the tall peaks of the KPZ equation. This is based on separate joint works with LiCheng Tsai and Promit Ghosal. In the second part of the talk, we will study the KPZ equation through the lens of polymers. In particular, we will discuss localization aspects of CDRP that will shed light on certain properties of the KPZ equation such as ergodicity and limiting Bessel behaviors around the maximum. This is based on joint work with Weitao Zhu. 
Friday, March 25th 
Yier Lin
University of Chicago 
We show a limit shape result of the KPZ equations under weak noise scaling, when we condition the value of the KPZ equations at one point to be very large. We also show that under such conditioning, the shape of the noise in the KPZ equation is asymptotically given by the optimizer of the L^4 GagliardoNirenbergSobolev inequality. 
Thursday, March 31st Special time: 11:00am 
Zbigniew J. Jurek
University of Wrocław 
The class, ID, of infinitely divisible distributions appears as the class of limiting distributions of uniformly infinitesimal triangular arrays. It is intimately connected with Lévy stochastic processes. On the level of characteristic functions (Fourier transforms) infinitely divisible distributions are characterized by the LévyKhintchine formula. In this lecture, following Kazimierz Urbanik (1972 and 1973), we will present an increasing sequence of subclasses (convolution semigroups) that begins with the class of normal distributions and ends (after taking closure) with the whole class ID. Each subclass will be described, among others, via the random integrals and characteristic functions. [Also we may mention the general conjecture on random integral representations.] 
Friday, April 1st 
Michael Cranston
UC Irvine 
We examine statistical properties of integers when they are sampled using the Riemann zeta distribution and compare to similar properties when they are sampled according to "uniform" or harmonic distributions. For example, as the variable in the Riemann zeta function approaches 1, a central limit theorem and large and moderate deviations for the distinct number of prime factors for the sampled integer can be readily derived. These results can then be deduced for the uniform distribution via a Tauberian Theorem. 
Friday, April 8th  No seminar this week


Friday, April 15th  No seminar this week


Friday, April 22nd 
Atilla Yilmaz
Temple University 
After giving a general and selfcontained introduction to the homogenization of HamiltonJacobi (HJ) equations, including the classical results in the cases where the Hamiltonian is periodic in the spatial variable $x$ or convex in the gradient variable $p$, I will focus on viscous HJ equations in one space dimension with separable Hamiltonians of the form $G(p) + V(x,\omega)$, where $G$ is a nonconvex function and $V$ is a stationary & ergodic random potential that satisfies a valley & hill condition (which holds in many natural examples, but fails in the periodic case). I will present several recent results on this class of HJ equations (by various combinations of A. Davini, E. Kosygina, O. Zeitouni and myself), where homogenization is established by showing that, outside of the intervals where the effective Hamiltonian turns out to be flat (due to the valley & hill condition), there is a unique sublinear corrector (which is a notion I will introduce) with certain properties. In the special case where $G$ is the minimum of two identical parabolas, these sublinear correctors have convenient representations involving Brownian motion in a random potential. More generally, the existence & uniqueness of these sublinear correctors can be proved using ODE methods that bypass the need for explicit representations, which I will demonstrate when $G$ is quasiconvex. 
Friday, April 29th 
Xiaohui Chen
University of Illinois at UrbanaChampaign 
We introduce the diffusion Kmeans clustering method on Riemannian submanifolds, which maximizes the withincluster connectedness based on the diffusion distance. The diffusion Kmeans constructs a random walk on the similarity graph with vertices as data points randomly sampled on the manifolds and edges as similarities given by a kernel that captures the local geometry of manifolds. The diffusion Kmeans is a multiscale clustering tool that is suitable for data with nonlinear and nonEuclidean geometric features in mixed dimensions. Given the number of clusters, we propose a polynomialtime convex relaxation algorithm via the semidefinite programming (SDP) to solve the diffusion Kmeans. In addition, we also propose a nuclear norm regularized SDP that is adaptive to the number of clusters. In both cases, we show that exact recovery of the SDPs for diffusion Kmeans can be achieved under suitable betweencluster separability and withincluster connectedness of the submanifolds, which together quantify the hardness of the manifold clustering problem. We further propose the localized diffusion Kmeans by using the local adaptive bandwidth estimated from the nearest neighbors. We show that exact recovery of the localized diffusion Kmeans is fully adaptive to the local probability density and geometric structures of the underlying submanifolds. Joint work with Yun Yang (UIUC). 
Stochastics Seminar for Spring 2022 is organized at the University of Utah by B. Cooper Boniece, Anna Little, and Xiao Shen.
To receive email announcements please join the seminar mailing list.
This web page is maintained by Tom Alberts.
Past Seminars:
 Fall 2021  Spring 2022
 Fall 2020  Spring 2021
 Fall 2019  Spring 2020
 Fall 2018  Spring 2019
 Fall 2017  Spring 2018
 Fall 2016  Spring 2017
 Fall 2015  Spring 2016
 Fall 2014  Spring 2015
 Fall 2013  Spring 2014
 Fall 2012  Spring 2013
 Fall 2011  Spring 2012
 Fall 2010  Spring 2011
 Fall 2009  Spring 2010
 Fall 2008  Spring 2009
 Fall 2007  Spring 2008
 Fall 2006  Spring 2007
 Fall 2005  Spring 2006
 Fall 2004  Spring 2005
 Fall 2003  Spring 2003
 Fall 2002  Spring 2002
 Fall 2001
 Winter 2000
 Fall 1999
 Spring 1998