# Colloquia

## Professor Zhengchang Su 11/14/2012

Large scale annotation of cis-regulatory sequences in prokaryotic

genomes

Date: 11/14/2012

Time: 2:30-3:30 PM

Place: 315 Armstrong Hall

Abstract: Although we now can gain a fairly good understanding of coding

sequences or genes in any newly sequenced prokaryotic genomes thanks to

the development of accurate and efficient gene-finding tools, we know very

little about cis-regulatory sequences or transcription factor binding

sites in the vast majority of sequenced genomes owing to the lack of an

accurate and efficient computational method for their predictions. To

achieve the goal of computational annotation of cis-regulatory binding

sites in all sequenced prokaryotic genomes, we are developing algorithms

and tools for genome-wide de novo prediction of cis-regulatory binding

sites in a large scale through comparative genomics analysis. In my talk,

I will introduce our recent development of computational algorithms and

tools for the simultaneous prediction of cis-regulatory binding sites in a

group of prokaryotic genomes.

## Professor Anthony Hilton 11/9/2012

Bounds on the simple graph and multigraph (r,s,a,t)-threshold numbers

Date: 11/9/2012

Time: 3:30-4:30 PM

Place: 315 Armstrong Hall

Abstract can be downloaded as a PDF

## Professor Bolian Liu 11/8/2012

The sum of Laplacian

eigenvalues

Date: 11/8/2012

Time: 3:30-4:30 PM

Place: 315 Armstrong Hall

Abstract:

Let $G$ be a simple graph with $n$ vertices and $e(G)$ edges. A.E.

Brouwer et al. conjectured that the sum of the $k$ largest Laplacian

eigenvalues of $G$ is at most $e(G)+{k+1 \choose 2}$, where $1\leq

k\leq n$. In this talk , we survey the works for the proof of the conjecture.

## Professor Marcus Wunsch 10/31/2012

The Hunter-Saxton system and

the geodesics on (pseudo-)spheres

Date: 10/31/2012

Time: 3:30-4:30 PM

Place: 315 Armstrong Hall

Abstract can be downloaded as a PDF

## Professor Stacey Levine 10/24/2012

Image Fusion using

Gaussian Mixture Models

Date: 10/24/2012

Time: 3:30-4:30 PM

Place: 315 Armstrong Hall

Abstract:

A number of recent works have demonstrated that using

patches, in lieu of pixels, as image features can more effectively

perform various techniques such as denoising, deblurring, inpainting

and super-resolution. This if often carried out by sparsely

representing the images patches in appropriately defined, possibly

redundant, dictionaries. Yu, Sapiro, and Mallat showed that a related

but more stable solution can be found by estimating the patches using

Gaussian Mixture Models (GMMs), particularly when solving ill-posed

inverse problems such as deblurring and super-resolution. In this talk

we discuss how this GMM approach can be can be used for fusing images

of the same field of view, suffering from any or all of the

above-mentioned degradations. The fusion model retains many of the

nice properties of the single image GMM model such as its equivalence

to finding an optimal sparse representation in a PCA dictionary, and

can be simply modified to handle spatially varying features, including

geometric features (e.g. edges, smooth regions, and textures) as well

as spatially varying noise levels. We will also discuss how some of

these results fair with respect to comparable variational approaches.

## Professor Keith Weber 10/04/2012

Reading and comprehending

mathematical proofs

Date: 10/04/2012

Time: 4:00-5:00 PM

Place: 315 Armstrong Hall

Abstract:

In advanced undergraduate mathematics classes, students spend a

substantial amount of time studying mathematical proofs. Yet it is generally

accepted that most students have difficulty understanding the proofs that

they read. In this talk, I will discuss (a) what it means for a student to

understand a mathematical proof and how this understanding can be assessed,

(b) unproductive beliefs about proof held by students that inhibit

understanding, (c) strategies that students can use when reading a proof

that will increase understanding, and (d) lessons learned from experiments

in which we attempted to teach students to use these strategies.

## Professor Charis Tsikkou 9/24/2012

Conservation Laws with no Classical Riemann Solutions: Existence of

Dafermos profiles for singular shocks.

Date: 9/24/2012

Time: 3:30 PM

Place: 315 Armstrong Hall

Abstract:

The basic tool in the construction of solutions to the Cauchy

problem for conservation laws with smooth initial data is the Riemann

problem. In this talk I will review the results obtained for the solutions to the

Riemann problem and present a system of two equations derived from

isentropic gas dynamics with no classical solution. I will then use the

blowing-up approach to geometric singular perturbation problems to show

that the system exhibits unbounded solutions (singular shocks) with

Dafermos profiles.

## Professor Richard Price 9/20/2012

Binary Black Hole Inspiral: Legends

of the Fall

Date: 9/20/2012

Time: 3:30 PM

Place: G09 White Hall

*Refreshments will be served at 3PM.

This is a talk cosponsored by the Department of Mathematics and the Department of Physics. Any faculty or students interested in meeting with Dr. Price should contact me (ef@math.wvu.edu) or Joann (mayhew@math.wvu.edu) to reserve some time on Friday, 9/21. We will have a lunch with him at 12PM Friday (please join us) and there will be time for meetings after that (roughly 1-4p) to meet individually. If you have some interest in this area I encourage you to schedule a time.

Professor Price's flyer for the colloquium is here

## Professor Moseley 9/6/2012

Date: 9/6/2012

Time: 4:30 PM

Place: 315 Armstrong Hall

I will be giving an undergraduate colloquium on Thursday

September 6, 2012 in Armstrong 315 at 4:30 to help everyone to

understand the "Linear Operator Theory" approach to Math 251 and Math

261. See attached. Not all in Math 251 and Math 261 is Linear Theory,

but enough is to make this approach reasonable, especially for the

engineers. About 90-99% of the students in Math 251 and Math 261 are

engineers. They need to know what a Linear System is from a mathematical

perspective. This begins with the definition of a vector space as an

abstract algebraic structure. Students (engineering and others) find

this hard, but math avoidance is not the answer. In my classes I am

trying to keep them from getting stuck in 3 space. (Please help me.

I'm stuck in 3 space and I can't get out.) The sooner they are exposed

to the Euclidean view of mathematics, the better. This does not

necessarily mean a lot of proofs. I do very few proofs. You and all

of your students are invited. I have enough overheads for 2-3 hours. However, I have cut

it down to a 10-20 minute talk plus questions. Hope to see you there.

PDF for colloquium can be downloaded here

## Mr. Yezhou Wu 7/10/2012

On the inversion of the Vandermonde matrix

Date: 7/10/2012

Time: 2:00 PM

Place: 112 Armstrong Hall

The inversion of the Vandermonde matrix plays important role in the

solution of many problems. For a general n by n matrix it costs

O(n^3) time to calculate the inversion. In this talk we will give an

algorithms in O(n^2) for Vandermonde matrix and show some

applications of linear codes.

## Pages