Universality for the Toda algorithm to compute the eigenvalues of a random matrix.

Speaker: 

Tom Trogdon

Institution: 

UCI

Time: 

Tuesday, October 11, 2016 - 11:00am to 11:50am

Host: 

Location: 

RH 3066

 

 

The Toda lattice, beyond being a completely integrable dynamical system, has many important properties.  Classically, the Toda flow is seen as acting on a specific class of bi-infinite Jacobi matrices.  Depending on the boundary conditions imposed for finite matrices, it is well known that the flow can be used as an eigenvalue algorithm. It was noticed by P. Deift, G. Menon and C. Pfrang that the fluctuations in the time it takes to compute eigenvalues of a random symmetric matrix with the Toda, QR and matrix sign algorithms are universal. In this talk, I will present a proof of such universality for the Toda algorithm.  I will also discuss empirical and rigorous results for other algorithms from numerical analysis.  This is joint work with P. Deift.

A two scale proof of the Eyring-Kramers formula (joint work with Andre Schlichting)

Speaker: 

Georg Menz

Institution: 

UCLA

Time: 

Tuesday, October 4, 2016 - 11:00pm to 11:50pm

Host: 

Location: 

RH 306

We consider a drift-diffusion process on a smooth potential landscape

with small noise. We give a new proof of the Eyring-Kramers formula

which asymptotically characterizes the spectral gap of the generator of

the diffusion. The proof is based on a refinement of the two-scale

approach introduced by Grunewald, Otto, Villani, and Westdickenberg and

of the mean-difference estimate introduced by Chafai and Malrieu. The

new proof exploits the idea that the process has two natural

time-scales: a fast time-scale resulting from the fast convergence to a

metastable state, and a slow time-scale resulting from exponentially

long waiting times of jumps between metastable states. A nice feature

of the argument is that it can be used to deduce an asymptotic formula

for the log-Sobolev constant, which was previously unknown.

Obliquely reflected Brownian motion

Speaker: 

Zhen-qing Chen

Institution: 

University of Washington

Time: 

Tuesday, May 17, 2016 - 1:00pm to 1:50pm

Host: 

Location: 

RH 306

Boundary theory for one-dimensional diffusions is now well understood. Boundary theory for multi-dimensional diffusions is much richer and remains to be better understood. In this talk, we will be concerned with the construction and characterization of obliquely reflected Brownian motions in all bounded simply connected planar domains, including non-smooth domains,
with general reflection vector field on the boundary.  
We show that the family of all obliquely reflected Brownian motions in a given domain can be characterized in two different ways, either by the field of angles of oblique reflection on the boundary or by the stationary distribution and the rate of rotation of the process about a reference point in the domain. We further show that Brownian motion with darning and excursion reflected Brownian motion can be obtained as a limit of obliquely reflected Brownian motions.

Based on joint work with K. Burdzy, D. Marshall and K. Ramanan.

On the % of zeros of Riemann zeta-Function on the critical line.

Speaker: 

Nicolas Martinez Robles

Institution: 

Univ. Illinois

Time: 

Tuesday, April 19, 2016 - 11:00am to 11:50am

Host: 

Location: 

RH 306

Abstract: We will review the techniques used to prove that a positive proportion of the zeros of the Riemann zeta-Function lie on the critical line Re(s)=1/2. The famous Riemann hypothesis states that all the zeros lie there. We will then discuss the mollifiers that allow us to show that > 41% of zeros are critical. This is joint work with A. Roy and A. Zaharescu.

Coupling for Brownian Motion with Redistribution

Speaker: 

Iddo Ben-Ari

Institution: 

University of Connecticut

Time: 

Tuesday, April 5, 2016 - 11:00am to 12:00pm

Location: 

RH 306

We consider a model of Brownian motion on a bounded interval which upon exiting the interval is being redistributed back  into the interval according to a probability measure depending on the exit point, then starting afresh, repeating the above mechanism indefinitely.  It is not hard to show that the process is exponentially ergodic, although characterizing the rate of convergence is non-trivial. In this talk, after providing a general overview of the probabilistic method of coupling and its applications,  I’ll show how to study the ergodicity for the model through coupling, how it leads to an  intuitive and geometric explanation for  the rates of convergence previously obtained analytically, other insights, and more questions. The talk will be accessible to general mathematical audience. 

Reminiscing on partition function zeros and the Lee-Yang circle theorem.

Speaker: 

Marek Biskup

Institution: 

UCLA

Time: 

Tuesday, March 29, 2016 - 11:00am to 12:00pm

Host: 

Location: 

RH 306

I will review some (actually quite old) results by C. Borgs, J.T. Chayes, R. Kotecky and myself concerning the partition function zeros of the Ising model. The focus will be on the fact that, for specific boundary conditions, the zeros lie (in a suitable representation) on the unit circle. I will explain (1) the classic proof of the Lee-Yang circle theorem and (2) how one can nail the positions of the zeros up to exponentially small errors in the system size for the periodic boundary conditions. I may find time to explain how one uses this result to prove the so called Griffiths singularities in site-diluted Ising model.

A Boundedness Trichotomy for the Stochastic Heat Equation

Speaker: 

Davar Khoshnevisan

Institution: 

University of Utah

Time: 

Tuesday, January 26, 2016 - 11:00am to 12:00pm

Host: 

Location: 

RH 306

  We consider the stochastic heat equation with a multiplicative space-time white noise forcing term under standard "intermitency conditions.” The main byproduct of this talk is that, under mild regularity hypotheses, the a.s.-boundedness of the solution$x\mapsto u(t\,,x)$ can be characterized generically by the decay rate, at $\pm\infty$, of the initial function $u_0$. More specifically, we prove that there are 3 generic boundedness regimes, depending on the numerical value of $\Lambda:=\lim_{|x|\to\infty} \vert\log u_0(x)\vert/(\log|x|)^{2/3}$.

The spectral gap for random regular graphs

Speaker: 

Tobias Johnson

Institution: 

USC

Time: 

Tuesday, January 12, 2016 - 11:00am to 12:00pm

Host: 

Location: 

RH 306

Expander graphs are useful across mathematics, all the way from number theory to applied computer science. The smaller the second eigenvalue of a regular graph, the better expander it is. Since this connection was discovered in the 1980s, researchers have tried to pinpoint the second eigenvalue of random regular graphs. The most prominent work in this direction was Joel Friedman's proof of Noga Alon's conjecture from 1985 that for a random d-regular graph on n vertices, the second eigenvalue is almost as small as possible, with high probability as n tends to infinity with d held fixed.

We consider the case of denser graphs, where d and n are both growing. Here, the best result (Broder, Frieze, Suen, Upfal 1999) holds only if d = o(n^(1/2)). We extend this to d = O(n^(2/3)). Our result relies on new concentration inequalities for statistics of random regular graphs based on the theory of size biased couplings, an offshoot of Stein's method. The theory we develop should be useful for proving concentration inequalities in a broad range of settings. This is joint with Nicholas Cook and Larry Goldstein.

Diffusive limits for stochastic kinetic equtions

Speaker: 

Arnaud Debussche

Institution: 

Univ. Rennes

Time: 

Tuesday, November 10, 2015 - 11:00am to 12:00pm

Host: 

Location: 

RH 306

In this talk, we consider kinetic equations containing random

terms. The kinetic models contain a small parameter and it is well

known that, after scaling, when this parameter goes to zero the limit

problem is a diffusion equation in the PDE sense, ie a parabolic equation

of second order. A smooth noise is added, accounting for external perturbation.

It scales also with the small parameter. It is expected that the limit

equation is then a stochastic parabolic equation where the noise is in

Stratonovitch form.

Our aim is to justify in this way several SPDEs commonly used.

We first treat linear equations with multiplicative noise. Then show how

to extend the methods to nonlinear equations or to the more physical

case of a random forcing term.

The results have been obtained jointly with S. De Moor and J. Vovelle.

Low Correlation Noise Stability of Euclidean Sets

Speaker: 

Steve Heilman

Institution: 

UCLA

Time: 

Tuesday, November 24, 2015 - 11:00am to 12:00pm

Host: 

Location: 

RH 306

The noise stability of a Euclidean set is a well-studied quantity.  This quantity uses the Ornstein-Uhlenbeck semigroup to generalize the Gaussian perimeter of a set.  The noise stability of a set is large if two correlated Gaussian random vectors have a large probability of both being in the set.  We will first survey old and new results for maximizing the noise stability of a set of fixed Gaussian measure.  We will then discuss some recent results for maximizing the low-correlation noise stability of three sets of fixed Gaussian measures which partition Euclidean space.  Finally, we discuss more recent results for maximizing the low-correlation noise stability of symmetric subsets of Euclidean space of fixed Gaussian measure.  All of these problems are motivated by applications to theoretical computer science.

Pages

Subscribe to RSS - Combinatorics and Probability