|
|
|
|
Home
Search |
Upcoming Seminars
MONDAY, March 31, 2008
GRADUATE STUDENT RESEARCH SEMINAR PSA 103 12:00 p.m.
Trent Lalonde, Department of Mathematics and Statistics
"Overdispersion in Generalized Linear Models"
ABSTRACT: Normal linear model theory is very useful for
situations with continuous response data. When response
variables are discrete, normal models may not be appropriate.
In this case it is common to apply generalized linear modeling
techniques instead. These models for categorical data often
come with an assumed relationship between the mean and the
variance of the response, called the mean-variance
relationship. When the actual variation in the response exceeds
this assumed variance as a function of the mean, the data are
overdispersed.
This talk will motivate overdispersion in generalized linear
models, explain sources of overdispersion, and describe some
techniques commonly used to correct a model with overdispersed
response data. Limitations of current approaches will be
discussed, as well as possible extensions of overdispersion
methods to hierarchical models.
Bagels and juice will be served in PSA 103 at 11:50 a.m.
COLLOQUIUM PSA 206 3:40 p.m.
Saul Abarbanel, Tel Aviv University
"Nonlinear PML Equations for Time Dependent Electromagnetics
in Three Dimensions"
ABSTRACT: In this talk we present a new set of nonlinear PML
equations for the multi-dimensional Maxwell's equations and
show that they are strongly well posed and temporally stable.
They are presented in vector form and so may be directly
written in any set of orthogonal coordinates. Another virtue of
this set of equations is that there is no need for nonphysical
variables in the absorbing layer. This greatly reduces the
number of equations that one solves numerically (from 12 to 6
in the three dimensional case).
Refreshments will be served in PSA 206 at 3:15 p.m.
TUESDAY, April 1, 2008
COMPUTATIONAL AND APPLIED MATHEMATICS PROSEMINAR
PSA 206 3:15 p.m.
Saul Abarbanel, Tel Aviv University
"Long-Time Performance of Unsplit PMLs using Explicit Second
Order Schemes"
ABSTRACT: A gradual long-time growth of the solution in
perfectly matched layers (PMLs) has been previously reported in
the literature. This undesirable phenomenon may hamper the
performance of the layer, which is designed to truncate the
computational domain for unsteady wave propagation problems.
For unsplit PMLs, prior studies have attributed the growth to
the presence of multiple eigenvalues in the amplification
matrix of the governing system of differential equations. In
this talk we analyze the temporal behavior of the unsplit PMLs
for some commonly used second order explicit finite-difference
schemes that approximate Maxwell's equations. Our conclusion is
that in addition to having the PML as a potential source of
long-time growth, the type of layer and the choice of the
scheme play a major role in how rapidly this growth may
manifest itself and whether or not it manifests itself at all.
APPLIED ANALYSIS AND PDE READING SEMINAR PSA 546 3:00 p.m.
For more information, contact Svetlana Roudenko.
WEDNESDAY, April 2, 2008
MATLAB WORKSHOP PSA 206 1:30 p.m.
Brett Shoelson, The MathWorks
"Bioinformatics with MATLAB"
AGENDA: This free technical session will demonstrate how you
can use MATLAB as a flexible platform for technical computing
and application development in engineering, math, and science
curricula and research.
1:30 p.m. Registration Sign-in (walk-ins welcome)
2:00 - 4:30 p.m. Workshop
* Introduction to MATLAB: Visualizing gene expression
*Algorithm development
*Function generation
*Process automation
*Report generation
*GUI building
*Deployment
* Introduction to the Bioinformatics Toolbox
*Phylogenetic tree example: comparing human and
chimpanzee DNA
*Data Mining (accessing and searching NCBI databases)
*Microarray data analysis
* Clustering tools
* Interconnectivity of tools: using neural networks to
classify pixels in an image
* Automation and visualization: Atomic force microscopy and
the notion of serendipity
Target audience: Professors, researchers, and students.
Registration at www.mathworks.com/seminars/asuapr08
For more information contact Tim Mathieu, Account Manager in
Education Department at the MathWorks:
Tim.Mathieu@mathworks.com, phone 508-647-7016
NUMBER THEORY SEMINAR PSA 308 1:40 p.m.
Daniel Corral, Department of Mathematics and Statistics
"A Connected Topology for the Integers, II"
ABSTRACT: We resume our exposition of the paper "A Connected
Topology for the Integers" by Solomon W. Golomb. As in the
previous talk, let D denote the Golomb topology generated by
the empty set and a certain type of arithmetic progression in
the integers. In this talk we will discuss a theorem of Golomb
which states that the Dirchlet theorem on arithmetic
progressions is equivalent to the density of the primes in
(\mathbf{Z}, D). It will also be shown that the set of primes
has empty interior in (\mathbf{Z}, D). The final result in the
Golomb paper is that the set of positive integers n, where 6n-1
and 6n+1 are a prime pairing, is closed in a stronger topology
than D. If time permits we will begin a discussion of a paper
of A. M Kirch or present the related H. Furstenberg proof of
the infinitude of primes.
THURSDAY, April 3, 2008 MATLAB WORKSHOP ECA 225 9:00 a.m.
AGENDA: This free technical session will demonstrate how you
can use MATLAB as a flexible platform for technical computing
and application development in engineering, math, and science
curricula and research.
9:00 - 10:30 a.m. Parallel Computing, Distributed Engine
10:40 - 12:00 p.m. Image Processing (with examples from
Bioinformatics)
Target audience: Professors, researchers, and students.
Registration at www.mathworks.com/seminars/asuapr08
For more information contact Tim Mathieu, Account Manager in
Education Department at the MathWorks:
Tim.Mathieu@mathworks.com, phone 508-647-7016
COLLOQUIUM PSA 206 1:40 p.m.
V. Sverak, McKnight University Professor,
School of Mathematics, University of Minnesota
"PDE Aspects of the Navier-Stokes Equations"
ABSTRACT: The solutions of the Navier-Stokes equations can
exhibit a very rich variety of behaviors, from simple but
illuminating text-book type examples, to complicated phenomena
which remain mathematically unexplained. In this talk I will
try to give some illustrations of this.
Refreshments will be served in PSA 206 at 1:10 p.m.
MATLAB WORKSHOP ECA 221 2:00 p.m.
George Thiers, The MathWorks
"Advanced MATLAB"
AGENDA: This free technical session will demonstrate how you
can use MATLAB as a flexible platform for technical computing
and application development in engineering, math, and science
curricula and research.
2:00 p.m. Registration Sign-in (walk-ins welcome)
2:30 - 4:30 p.m. Workshop
* MATLAB and memory
* Functions of all types
* What's new in Release 2008a
Target audience: Professors, researchers, and students.
Registration at www.mathworks.com/seminars/asuapr08
For more information contact Tim Mathieu, Account Manager in
Education Department at the MathWorks:
Tim.Mathieu@mathworks.com, phone 508-647-7016
NATIONAL MATHEMATICS AWARENESS MONTH
BARBEQUE SOCIAL
Oasis Cantina Patio (south side of PSA Bldg) 5:00 p.m.
Tickets can be purchased in PSA 211:
$3 - advance tickets purchased before 12:00 p.m. on
Wednesday, April 2;
$4 - after 12:00 pm. on April 2 or at the event.
Menu: Choice of Beef, Pork or Chicken barbeque sandwich with
chips and soda.
NATIONAL MATHEMATICS AWARENESS MONTH
PUBLIC LECTURE MUR 201 7:30 p.m.
Alan Taylor, Union College
"The Mathematics of Voting"
ABSTRACT: We will give a quick survey illustrating the kinds of
mathematical questions and answers that arise from real-world
voting systems. Many of these results assert that certain
election-theoretic desiderata are impossible to attain.
Examples include: a "simple" description of the US federal
system, an equally good alternative to majority rule, a fair
method of apportionment on which to base the electoral college,
and voting systems for three or more alternatives in which
honesty is the best policy. Along the way we'll see that
sometimes (i) having a vote is just like not having a vote,
(ii) a candidate can lose to an opponent that everyone likes
less, and (iii) gaining a vote can lead to a loss.
Refreshments will be served after the lecture.
FRIDAY, April 4, 2008
NATIONAL MATHEMATICS AWARENESS MONTH
SEMINAR PSA 206 1:40 p.m.
Alan Taylor, Union College
"A Peculiar Connection between the Axiom of Choice and
Predicting the Future"
ABSTRACT: In the February issue of the Monthly, Chris Hardin
and I had an article with the above title. There we showed that
the axiom of choice implies that there is a strategy for
predicting the values of an arbitrary function defined on the
reals ~V based on its previous values ~V that is almost always
correct. Specifically, given the values of f on {x: x < t} the
strategy produces a guess for f(t), and for all t ~V except
those in a countable, nowhere dense set ~V it is correct. This
suggests the more general question: To what extent is a
function's value at a point x of a topological space determined
by its values in an arbitrarily small (deleted) neighborhood
of x? For continuous functions, the answer is typically
"always" and the method of prediction of f(x) is just the
limit operator. With C. Hardin, we've generalized this to the
case of an arbitrary function mapping a topological space to an
arbitrary set. We show that the best one can ever hope to do is
to predict correctly except on a scattered set. Moreover, we
obtain a predictor whose error set is always scattered.
MATH BIOLOGY SEMINAR ECG 237 3:40 p.m.
Jean-Marc Fellous, University of Arizona
"Is It Time for Spikes? Precision and Reliability of the
Neural Code"
|