Gaussian comparisons meet convexity: Precise analysis of structured signal recovery

Speaker: 

Christos Thrampoulidis

Institution: 

MIT

Time: 

Tuesday, November 14, 2017 - 11:00am to 11:50am

Host: 

Location: 

RH 306

Gaussian comparison inequalities are classical tools that often lead to simple proofs of powerful results in random matrix theory, convex geometry, etc. Perhaps the most celebrated of these tools is Slepian’s Inequality, which dates back to 1962. The Gaussian Min-max Theorem (GMT) is a non-trivial generalization of Slepian’s result, derived by Gordon in 1988. Here, we prove a tight version of the GMT in the presence of convexity. Based on that, we describe a novel and general framework to precisely evaluate the performance of non-smooth convex optimization methods under certain measurement ensembles (Gaussian, Haar). We discuss applications of the theory to box-relaxation decoders in massive MIMO, 1-bit compressed sensing, and phase-retrieval.

From Spaces to Scales to Ordinals

Speaker: 

Jeffrey Bergfalk

Institution: 

Cornell University

Time: 

Monday, October 16, 2017 - 4:00pm to 5:50pm

Host: 

Location: 

RH 440R

We describe a number of related questions at the interface of set theory and homology theory, centering on (1) the additivity of strong homology, and (2) the cohomology of the ordinals. In the first, the question is, at heart: To how general a category of topological spaces may classical homology theory be continuously extended? And in the tension between various potential senses of continuity lie a number of delicate set-theoretic questions. These questions led to the consideration of the Cech cohomology of the ordinals; the surprise was that this is a meaningful thing to consider at all. It very much is, describing or suggesting at once (i) distinctive combinatorial principles associated to the nth infinite cardinal, for each n, holding in ZFC, (ii) rich connections between cofinality and dimension, and (iii) higher-dimensional extensions of the method of minimal walks.

HEAT KERNEL ESTIMATES FOR TIME FRACTIONAL EQUATIONS

Speaker: 

Panki Kim

Institution: 

Seoul National University

Time: 

Friday, October 20, 2017 - 2:00pm to 3:00pm

Host: 

Location: 

NS2 1201

 In this talk, we first discuss existence and uniqueness of weak solutions to general time fractional equations and give their probabilistic representation. We then talk about sharp two-  sided estimates for fundamental solutions of general time fractional equations in metric measure spaces. This is a joint work with  Zhen-Qing Chen(University of Washington, USA), Takashi Kumagai (RIMS, Kyoto University, Japan) and Jian Wang (Fujian Normal University, China).

Minimizers of the sharp Log entropy on manifolds with non-negative Ricci curvature and flatness

Speaker: 

Qi Zhang

Institution: 

UC Riverside

Time: 

Tuesday, November 21, 2017 - 4:00pm

Host: 

Location: 

RH 306

Consider the scaling invariant, sharp log entropy (functional)
introduced by Weissler on noncompact manifolds with nonnegative Ricci
curvature. It can also be regarded as a sharpened version of
Perelman's W entropy  in the stationary case. We prove that it has a
minimizer if and only if the manifold is isometric to $\R^n$.
Using this result, it is proven that a class of noncompact manifolds
with nonnegative Ricci curvature is isometric to $\R^n$. Comparing
with some well known flatness results in on asymptotically flat
manifolds and asymptotically locally Euclidean (ALE) manifolds, their
decay or integral condition on the curvature tensor is replaced by the
condition that the metric converges to the Euclidean one in C1 sense
at infinity. No second order condition on the metric is needed.

Mean Value Theorems for Riemannian Manifolds via the Obstacle Problem

Speaker: 

Jeremy LeCrone

Institution: 

University of Richmond

Time: 

Tuesday, February 13, 2018 - 3:00pm to 4:00pm

Host: 

Location: 

RH 306

In this talk, I will discuss recent results produced with co-authors Ivan Blank (KSU) and Brian Benson (UCR) regarding a formulation of the Mean Value Theorem for the Laplace-Beltrami operator on smooth Riemannian manifolds. We define the sets upon which mean values of (sub)-harmonic functions are computed via a particular obstacle problem in geodesic balls. I will thus begin by discussing the classical obstacle problem and then an intrinsic formulation on manifolds developed in our recent paper. After demonstrating how the theory of obstacle problems is leveraged to produce our Mean Value Theorem, I will discuss local and global theory for our family of mean value sets and potential connections between the properties of these sets and the geometry of the underlying manifold.

Regularity of time-fractional reaction-diffusion problems and their solution by a graded-mesh finite difference method

Speaker: 

Martin Stynes

Institution: 

Beijing Computational Science Research Center

Time: 

Thursday, December 7, 2017 - 11:00am to 11:50am

Host: 

Location: 

RH340P

 

A reaction-diffusion initial-boundary problem with a Caputo time derivative of order $\alpha\in (0,1)$ is considered. The solution of such a problem is discussed; it is shown that in general the solution has a weak singularity near the initial time $t=0$, and sharp pointwise bounds on the derivatives of this solution are derived. These bounds are used in a new analysis of the standard L1 finite difference method for the time derivative combined with a standard finite difference approximation for the spatial derivative. This analysis encompasses both uniform meshes and meshes that are graded in time, and includes new stability and consistency bounds. The final convergence result shows clearly how the regularity of the solution and the grading of the mesh affect the order of convergence of the difference scheme, so one can choose an optimal mesh grading to solve the problem numerically.

Numerical computations towards Hamilton-Jacobi equations in probability space from mean field games

Speaker: 

Raymond Chow

Institution: 

UCLA

Time: 

Monday, October 9, 2017 - 4:00pm to 5:00pm

Host: 

Location: 

RH306

We explore possibility of computing solutions of a certain type of infinitely dimensional Hamilton-Jacobi equations in probability space that arises in the theory of mean field games. Numerical solution to such HJ-PDE was difficult owing to the high dimension of the PDE after discretization of a function space. We propose to utilize a Hopf formula coming from an optimal control approach. The resulting formula is an optimization problem involving a d dimensional HJ-PDE constraint, i.e. the mean field equations, which can be computed using a standard finite difference scheme. In particular, our method will provide us one possible way to compute proximal maps of Wasserstein metrics. They may be of importance in computing optimization problems involving Wasserstein metrics. Our techniques may have applications in optimal transport, mean field games and optimal control in the space of probability densities.

Localization in the droplet spectrum of the random XXZ quantum spin chain

Speaker: 

Abel Klein

Institution: 

UCI

Time: 

Thursday, October 12, 2017 - 2:00pm

Location: 

RH 340P

We study the  XXZ quantum spin chain in  a random field. This model is particle number preserving, which allows  the reduction to an infinite system of discrete many-body random Schrodinger operators.  We exploit this reduction to prove a form of  Anderson localization in the droplet  spectrum of the XXZ quantum spin chain Hamiltonian. This yields a strong form of dynamical exponential clustering for eigenstates  in the droplet spectrum: For any pair of local observables,  the sum of the associated correlators over these states decays exponentially  in the distance between the  local observables. Moreover,  this exponential clustering persists under the time evolution in the  droplet spectrum.

Pages

Subscribe to UCI Mathematics RSS