**Seminars and Colloquiums**

for the week of October 8, 2018

for the week of October 8, 2018

*SPEAKERS*

**Monday**

Nikolay Brodskiy, University of Tennessee

** Tuesday**

Theodora Bourni, University of Tennessee

**Wednesday**

Remus Nicoara, University of Tennessee

**Thursday**

Stephen A. Collins-Elliott, University of Tennessee

Daniel Shankman, Purdue University

Ioakeim Ampatzoglou, University of Texas at Austin

**Friday**

Suzanne Lenhart, University of Tennessee

Nicholas Zabaras, University of Notre Dame

**TEA TIME**

3:00 PM – 3:30 PM

Monday, Tuesday, & Wednesday

Rooms: Ayres 401

Hosted by: Pawel Grzegrzolka & Delong Li

Topics: How to transition from coursework to research; how to transition from a graduate student to a faculty member; how to transition from a graduate student to an industrial position.

**Monday, 10/8**

**TOPOLOGY/ GEOMETRY SEMINAR
**TITLE: Dimension of tree-graded spaces II

SPEAKER: Nikolay Brodskiy, University of Tennessee

TIME: 3:35 PM-4:25 PM

ROOM: Ayres 406

The concept of tree-graded space was introduced by C. Drutu and M. Sapirwhen they proved (jointly with D. Osin) that a finitely generated group G is relatively hyperbolic with respect to finitely generated subgroups H1, . . . , Hn if and only if every asymptotic cone of G is tree-graded with respect to the limits of sequences of cosets of the subgroups Hi. We will explore how various dimension-like properties of a tree-graded space can be derived from the corresponding properties of the building pieces of the space.

**Tuesday, 10/9**

**MINIMAL SURFACES SEMINAR
**TITLE: Colding-Minicozzi Paper 2--last part and Paper 1--part 1

SPEAKER: Theodora Bourni, University of Tennessee

TIME: 4:00 PM-5:30 PM

ROOM: Ayres 121

Blow-up techniques and Foliations by catenoids.

**Wednesday, 10/10**

**ANALYSIS SEMINAR
**TITLE: Spin model commuting squares and intermediate subfactors.

SPEAKER: Remus Nicoara, University of Tennessee

TIME: 2:30 PM-3:20 PM

ROOM: Ayres 113

We discuss a class of spin model commuting squares which yields subfactors (inclusions of von Neumann algebras) with intermediate subfactors.

**Thursday, 10/11**

**MATHEMATICAL DATA SCIENCE SEMINAR
**TITLE: Interval Estimation through Subsampling of Random Losses: Quantifying Coin Use in a Roman Rural Economy

SPEAKER: Stephen A. Collins-Elliott, University of Tennessee

TIME: 2:10 PM-3:10 PM

ROOM: Ayres 405

Quantifying archaeological artifacts is essential for understanding the scale of past socioeconomic activities and is straightforward enough when collecting finds, whether on the surface or from excavated strata. However, estimating how much of a material was in use in given location at a particular moment in time from those artifacts is quite a different problem. Numerous issues must be confronted, such as: an unknown population, a collected sample whose composition is beyond the control of the investigator, and high levels of uncertainty around the dates of use. In this talk, I discuss the use of simulation and resampling methods, like the bootstrap and random subsampling, to obtain summary statistics and hence an interval estimate of the frequency of coin use at rural Roman sites from the late first century BCE to the early fourth century CE.

**JR. COLLOQUIUM
**TITLE: Introduction to L-functions

SPEAKER: Daniel Shankman, Purdue University

TIME: 3:40 PM-4:35 PM

ROOM: Ayres 405

L-functions are mysterious complex analytic functions whose zeroes and poles are connected to the behavior of prime numbers. The most famous L-function is the Riemann zeta function. First, I will prove some things about the Riemann zeta function in order to investigate the convergence of the infinite sum 1/2 + 1/3 + 1/5 + 1/7 + etc. of the reciprocals of the prime numbers. Then I will define a Dirichlet L-function and use it to prove that there are infinitely primes of the form 4n+1. There is a much easier way to prove this, without using any analysis, but the method which I will present, using L-functions, can be generalized to show that there are infinitely many prime numbers of the form an + b, where a and b are integers without any prime factors in common. This generalization is a theorem originally proved by Dirichlet in 1837, and there is no known proof of this which does not use analysis.

**GEOMETRIC ANALYSIS SEMINAR
**TITLE: Α rigorous derivation of a ternary Boltzmann equation for a classical system of particles

SPEAKER: Ioakeim Ampatzoglou, University of Texas at Austin

TIME: 4:00 PM-5:00 PM

ROOM: Ayres 121

In this paper we present a rigorous derivation of a ternary Boltzmann equation describing the motion of a classical system of particles with three particles instantaneous interactions. The equation serves as a a kinetic model for a dense gas in non-equilibrium, and is for the first time derived from laws of instantaneous three particle interactions, preserving momentum and energy.

**Friday, 10/12**

**MATH BIOLOGY SEMINAR
**TOPIC: Top Down Control

LEADER: Suzanne Lenhart, University of Tennessee

TIME: 10:10 AM-11:00 AM

ROOM: Ayres 401

**COLLOQUIUM
**TITLE: Bayesian Deep Learning for Predicting Modeling in Science and Engineering

SPEAKER: Nicholas Zabaras, University of Notre Dame

TIME: 3:35 PM-4:35 PM

ROOM: Ayres 405

We will briefly review recent advances in the solution of stochastic PDEs using Bayesian deep encoder-decoder networks. These models have been shown to work remarkably well for uncertainty quantification tasks in very-high dimensions. In this talk through examples in computational physics and chemistry, we will address their potential impact for modeling dynamic multiphase flow problems, accounting for model form uncertainty in coarse grained RANS simulations and providing the means to coarse graining in atomistic models. Emphasis will be given to the small data domain using Bayesian variational approaches. The training of the network is performed using Stein variational gradient descent. We will show both the predictive nature of these models as well as their ability to capture output uncertainties induced by the random input, limited data and model error. In closing, we will outline the integration of these surrogate models with generative adversarial networks for the solution of inverse problems.

*If you are interested in giving or arranging a talk
for one of our seminars or colloquiums, please review our
calendar. *

*If you have questions, or a date you would like to confirm,
please contact mlangfo5
AT
utk DOT edu *

**Past notices:**