Scientific

Finite range decomposition of Gaussian fields

Speaker: 
Roland Bauerschmidt
Date: 
Tue, Jun 12, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

I will show how to decompose the Gaussian free field on a (weighted) graph into a sum of finite range Gaussian fields, which are smoother than the original field and have spatially localized correlations.

Class: 

Interacting Particle Systems 5

Speaker: 
Omer Angel
Date: 
Mon, Jun 11, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

Particles attempt to follow a simple dynamic (random walk, constant flow, etc) in some space (interval, line, cycle, arbitrary graph). Add a simple interaction between particles, and the behaviour can change completely. The resulting dynamical systems are far more complex than the ingredients suggest. These processes (interchange process, TASEP, sorting networks, etc) have diverse to many topics: growth processes, queuing theory, representation theory, algebraic combinatorics. I will discuss recent progress on and open problems arising from several models of interacting particle systems.

Class: 

Longest convex chains

Speaker: 
Gergely Ambrus
Date: 
Mon, Jun 11, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

A classical problem in probability is to determine the length of the longest increasing subsequence in a random permutation. Geometrically, the question can be formulated as follows: given n independent, uniform random points in the unit square, find the longest increasing chain (polygonal path through the given points) connecting two diagonally opposite corner of the square, where "length" means the number of points on the chain. The variant of the problem I am going to talk about asks for the length of the longest convex chain connecting two vertices. We determine the asymptotic expectation up to a constant factor, and derive strong concentration and limit shape results. We also prove an ergodic result as well as giving a heuristic argument for the exact asymptotics of the expectation. Some of these results are joint with Imre Barany.

Class: 

On Pólya Urn Schemes with Infinitely Many Colors.

Speaker: 
Debleena Thacker
Date: 
Mon, Jun 11, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

In this talk, we extend the mutlicolor P/'olya urn schemes to countably infinitely many colors. We index the colors by \mathbb{Z}. Throughout the talk, we discuss mainly replacement matrices arising out of random walks. We show that the proportion of colors with suitable centering and scaling show central tendencies. Also the centering and scaling are fairly general. This behavior is in sharp contrast with the finite color case, where the asypmtotic behavior of the proportion of colors are determined by the qualitative properties (transience or recurrence) of the Markov chain underlying the replacement matrix. We also extend the infinite color case to fairly general graphs on \mathbb{R}^{d} and show that the proportion of colors show central tendencies similar to that in the case for \mathbb{Z}. Even the centering and scaling remains same.

Class: 

Distributional fixed points and attractors in queueing theory.

Speaker: 
Sergio Ivan Lopez Ortega
Date: 
Mon, Jun 11, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

(Joint work with Pablo Ferrari) In 1956, Burke proved that the departure process of a stationary queue with Poisson arrivals is a Poisson process. In other words, Poisson process is a distributional fixed point for ./M/1 operation. Mountford and Prabhakar's theorem ('95) showed that Poisson process is not only a fixed point but an attractive distribution over some wide class of ergodic point processes on the line. Some extensions to non-markovian servers were made by Mairesse and Prabhakar ('03) for the existence of fixed points, and Prabhakar ('03) for the attractiveness. In 2001 O'Connell and Yor showed some brownian analogues to Burke's theorem. For a Brownian queue, Brownian motion is a distributional fixed point. In this talk we will show some results about the attractiveness of Brownian motion under the Brownian queue operation.

Class: 

Integral representations for the self-avoiding walk

Speaker: 
Gordon Slade
Date: 
Fri, Jun 8, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

The self-avoiding walk is a fundamental model in probability, combinatorics and statistical mechanics, for which many of the basic mathematical problems remain unsolved. Recent and ongoing progress for the four-dimensional self-avoiding walk has been based on a renormalization group analysis. This analysis takes as its starting point an exact representation of the self-avoiding walk problem as an equivalent problem for a perturbation of a Gaussian integral involving anti-commuting variables (fermions). This lecture will give a self-contained introduction to fermionic Gaussian integrals and will explain how they can be used to represent self-avoiding walks.

The talk is mainly based on the paper: D.C. Brydges, J.Z. Imbrie, G. Slade. Functional integral representations for self-avoiding walk. Probability Surveys, 6:34--61, (2009)

Class: 

Random Maps 4

Speaker: 
Gregory Miermont
Date: 
Tue, Jun 12, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

The study of maps, that is of graphs embedded in surfaces, is a popular subject that has implications in many branches of mathematics, the most famous aspects being purely graph-theoretical, such as the four-color theorem. The study of random maps has met an increasing interest in the recent years. This is motivated in particular by problems in theoretical physics, in which random maps serve as discrete models of random continuum surfaces. The probabilistic interpretation of bijective counting methods for maps happen to be particularly fruitful, and relates random maps to other important combinatorial random structures like the continuum random tree and the Brownian snake. This course will survey these aspects and present recent developments in this area.

Class: 

Interacting Particle Systems 4

Speaker: 
Omer Angel
Date: 
Fri, Jun 8, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

Particles attempt to follow a simple dynamic (random walk, constant flow, etc) in some space (interval, line, cycle, arbitrary graph). Add a simple interaction between particles, and the behaviour can change completely. The resulting dynamical systems are far more complex than the ingredients suggest. These processes (interchange process, TASEP, sorting networks, etc) have diverse to many topics: growth processes, queuing theory, representation theory, algebraic combinatorics. I will discuss recent progress on and open problems arising from several models of interacting particle systems.

Class: 

Random Maps 5

Speaker: 
Gregory Miermont
Date: 
Mon, Jun 11, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

The study of maps, that is of graphs embedded in surfaces, is a popular subject that has implications in many branches of mathematics, the most famous aspects being purely graph-theoretical, such as the four-color theorem. The study of random maps has met an increasing interest in the recent years. This is motivated in particular by problems in theoretical physics, in which random maps serve as discrete models of random continuum surfaces. The probabilistic interpretation of bijective counting methods for maps happen to be particularly fruitful, and relates random maps to other important combinatorial random structures like the continuum random tree and the Brownian snake. This course will survey these aspects and present recent developments in this area.

Class: 

Seed bank models with long range dependence

Speaker: 
Adrian Gonzalez-Casanova
Date: 
Thu, Jun 7, 2012
Location: 
PIMS, University of British Columbia
Conference: 
PIMS-MPrime Summer School in Probability
Abstract: 

In this talk I present a new model for seed banks, where individuals may obtain their type from ancestors which have lived in the near as well as the very far past. We discern three parameter regimes of the seed bank age distribution, which lead to substantially different behaviour in terms of genetic variability, in particular with respect to fixation of types and time to the most recent common ancestor. The classical Wright-Fisher model, as well as a seed bank model with bounded age distribution considered by Kaj, Krone and Lascoux (2001) are particular cases of the model. The mathematical methods are based not only on Markov chains, but also on renewal theory as well as on a Gibbsian approach introduced by Hammond and Sheffield (2011) in a different context. This talk is based in a joint work with Jochen Blath, Noemi Kurt, Dario Span`o.

Class: 

Pages