The method of nonlocal maximum principle was initiated by Kiselev et al (Inve. Math. 167 (2007), 445-453), and later developed by Kiselev (Adv. Math. 227 no. 5 (2011), 1806-1826) and other works. The general idea is to show that the evolution of considered equation preserves a suitable modulus of continuity, so that one gets the uniform-in-time control on the solution. In this talk, by using the method of nonlocal maximum principle and introducing some new moduli of continuity, we consider a class of drift-diffusion equations with nonlocal Levy-type diffusion, and we prove the eventual regularity result in the supercritical type cases, where the eventual regularity time can be evaluated small as the supercritical index approaching to the critical index for fixed initial data. We also show the global regularity of the vanishing viscosity solution in the logarithmically supercritical case. The talk is based on joint work with Changxing Miao from IAPCM, China.

In this talk, I will introduce a new geometric inequality:  the Sphere Covering Inequality. The inequality  states that   the  total area  of two {\it distinct}  surfaces with Gaussian curvature  less than 1,   which  are also conformal to  the Euclidean unit disk  with the same conformal factor on the boundary,  must be at least $4 \pi$.  In other words,  the areas of these surfaces must cover the whole unit sphere after a proper rearrangement. We apply the Sphere Covering Inequality to show the best constant of a Moser-Trudinger type inequality conjectured by A. Chang and P. Yang.    Other applications of this inequality  include the classification of certain Onsager vortices  on the sphere,  the radially symmetry of solutions to Gaussian curvature equation on the plane, classification of solutions for mean field equations on flat tori and  the standard sphere, etc.   The resolution of several open problems in these areas will  be presented.  The talk is based on joint work with Amir Moradifam from UC Riverside.

Many real-world networks -- social, technological, biological -- have wonderful structures. Some structures may be apparent (such as trees) while others may be hidden (such as communities). How can we discover hidden structures? Known approaches to "structure mining" in networks come from a variety of disciplines, including probability,  statistics, combinatorics, physics, optimization, theoretical computer science, signal processing and information theory. We will focus on new probabilistic approaches to structure mining. They bring together insights from random matrices, random graphs and semidefinite programming.

This is a joint applied math and probability seminar.

Previous works have shown that arctic circle phenomenons and limiting behaviors of some integrable discrete systems can be explained by a variational principle. In this talk we will present the first results of the same type for a non-integrable discrete system: graph homomorphisms form Z^d to a regular tree. We will also explain how the technique used could be applied to other non-integrable models.

Abstract: Szegö's theorem and the Kac-Murdoch-Szegö theorems are
classical asymptotic results about the distribution of the eigenvalues
of structured matrices. I will explain how these are useful in a
variety of applications (in particular analysis on Heisenberg groups)
_and_ show how they are equivalent to lovely theorems in random matrix
theory.