The Department of Mathematics and Statistics offers a weekly colloquium series on Thursdays. Most colloquia begin between 3:00 and 3:30. Seminars alternate between those aimed at Undergraduate (type U) and Graduate (type G) audiences.
Type 
Date 
Title 
Speaker 
U 
3/5/15  Sara Fett, Clinical Statistics LeighGomezDahl Cancer Center Statistics, Mayo Clinic Department of Health Sciences Research 

U 
3/3/15  Petr Kovar VSB  Technical University of Ostrava and UMD 

G 
2/26/15  Marcin Krzywkowski University of Johannesburg, South Africa, Gdansk University of Technology, Poland, and Polish Academy of Sciences 

U 
2/17/15  Kristine Snyder University of Michigan 

U 
2/12/15  Isabelle KemajouBrown University of Minnesota  Twin Cities 

U 
2/9/15  Bernard Chan San Diego State University 

U 
2/5/15  Xue Gong Ohio University 

U 
2/3/15  Sylwia CichaczPrzenioslo  
U 
1/30/15  Bethany Kubik United States Military Academy 

U 
1/26/15  Karl Wimmer Duquesne University 

U 
1/23/15  Sarah Anderson Clemson University 

G 
1/16/15  Katie Borchert UMD M.S. Candidate 

G 
12/18/14  Brittany Fanning UMD M.S. Candidate 

G 
12/11/14  John Greene, Mathematics & Statistics, UMD  
U 
12/4/14  William Krossner, Ph.D  
G 
11/20/14  Dalibor Froncek, Department of Mathematics & Statistics, UMD  
U 
10/02/14  Tracy A. Bibelnieks, Ph.D  
U 
9/25/14  Zhuangyi Liu, Department Head, Mathematics & Statistics, UMD  
U
 9/19/14  Greg Ostergren 

G 
8/18/14  Xuran Yan UMD M.S. Candidate 

G 
7/25/14  Shinjini Kar UMD M.S. Candidate 

G 
7/18/14  Xiaowen Fang UMD M.S. Candidate 

G 
7/8/14  Ondrej Zjevik UMD M.S. Candidate 

G 
6/16/14  Michal Hrabia UMD M.S. Candidate 

G 
6/12/14  Xiao Li UMD M.S. Candidate 

G 
6/9/14  Long Chen UMD M.S. Candidate 

G 
5/21/14  Wenchuan Guo UMD M.S. Candidate 

G 
5/20/14  Michael Lillegard UMD M.S. Candidate 

G 
5/19/14  Zhaobin Kuang UMD M.S. Candidate 

G 
5/14/14  Zhengfei Rui UMD M.S. Candidate 

U 
5/8/14  Jasmine Helgeson, Chenxiao Hu and Jing Lv, UMD Mathematics & Statistics Undergraduate Students  
G 
5/1/14  Greg Bard, University of Wisconsin  Stout  
U 
4/24/14  Philip Bauer, Jordan Maiers and Adam Swinney, UMD Mathematics & Statistics Undergraduate Students  
G 
4/17/14  Qingzhao Wang, Ph.D. Canidate in Mechanical Engineering, Virginia Polytechnic Institute  
U 
4/10/14  Jasmine Helgeson, Penghuan Ni, Tim Schoenheider, Philip Bauer, Jesse Schmieg and Joe Tonianato, UMD Mathematics & Statistics Undergraduate Students  
G 
4/3/14  William Keith, Assistant Professor, Mathematical Sciences, Michigan Tech University  
U 
3/27/14  Ross Garberich, Minneapolis Heart Institute Foundation, UMD Alumni  
G 
3/13/14  Thomas Cameron, UMD Alumni, Math Graduate Student  
U 
3/6/14  Karl Wimmer, Assistant Professor, Department of Mathematics, Duquesne University (UMD Graduate)  
U 
2/20/14  Xuan Li, Assistant Professor, Statistics, UMD  
U 
2/6/14 
Uwe Leck, Associate Professor, Department of Mathematics, UWS  
G 
1/30/14 *rescheduled to 2/13/14 
Steven Chiou, Assistant Professor, Statistics, UMD  
U 
12/12/13  Kyle Krueger and Brett Bozyk, UMD Graduates  
U 
11/21/13  Lucas Gloege, Inne Singgih and Ondrej Zjevik, Graduate Students, UMD  
U 
11/15/13  Ken Stanley, Director, Dubois Project of Oberlin College  
G 
11/14/13  Steve Butler, Department of Mathematics, Iowa State University  
U 
11/7/13  Richard Green, Department of Mathematics, UMD  
G 
10/31/13  Brian Hinderliter, Department of Mechanical and Industrial Engineering, UMD  
U 
10/24/13  Ron Regal, Department of Mathematics, UMD; William Salmon, Department of Linguistics, UMD 

G 
10/17/13  David Clark, Department of Mathematics, UMTC  
U 
10/10/13  Math Majors and Information Security in a Digital Age: Any Connection? 
Dr. William Krossner 
G 
10/3/13  Families of Infinite Series With Interesting Limiting Structure 
John Greene, Department of Mathematics, UMD 
U 
9/20/13  Joseph Gallian, Department of Mathematics, UMD  
G 
9/19/13  Michael Dorff, Department of Mathematics, Brigham Young University  
U 
9/19/13  Michael Dorff, Department of Mathematics, Brigham Young University 
 Archived 201213 Colloquia
 Archived 201112 Colloquia
 Archived 201011 Colloquia
 Archived 200910 Colloquia
 Archived 20089 Colloquia
 Archived 20078 Colloquia
 Archived 20067 Colloquia
 Archived 20056 Colloquia
Archives
Statistics in Medical Research
by
Sara Fett, Clinical Statistics
LeighGomezDahl Cancer Center Statistics
Mayo Clinic
Department of Health Sciences Research
Abstract:
We will share an overview of Mayo Clinic and Department of Health Sciences Research (HSR). Our department’s vision is to enhance the understanding of health and human disease through collaborative research teams. To help articulate what this looks like we will provide an overview of the divisions within HSR and the variety of projects we work on along with the differing positions (or job titles) we use to support these projects. We both manage analytical jobs within the Division of Biomedical Statistics and Informatics and as such we will delve more deeply into the roles and responsibilities held by our PhD, MS and BS level statistical and programming support. We will share more about our recruitment efforts and what skills work best for our teambased approach to medical research.
Time: Thursday, March 5, 2015; 3:004:00PM
Location: Chem 150
The Mathematics of the games Dobble and Sprouts
by
Petr Kovar
VSB  Technical University of Ostrava and UMD
Abstract:
In this talk we use mathematics to examine the games of Dobble and Sprouts. We identify some mistakes in the instruction manual for Dobble. We also describe how to create a game similar to Dobble. Using graph theory we determine lower and upper bounds on the number of moves in the game Sprouts. Moreover, we show that a modified version of Sprouts, which appears to be a fair game, is not. Some open problems related to both games will be mentioned.
Time: Tuesday, March 3, 2015; 3:004:00PM
Location: Chem 150
by
Marcin Krzywkowoski
University of Johannesburg, South Africa
Gdansk University of Technology, Poland, and Polish Academy of Sciences
Abstract:
The topic is the hat problem in which each of “n” players is randomly fitted with a blue or red hat. Then everybody can try to guess simultaneously his own hat color by looking at the hat colors of the other players. The team wins if at least one player guesses his hat color correctly, and no one guesses his hat color wrong; otherwise the team loses. The aim is to maximize the probability of winning. In this version every player can see everybody excluding himself. We consider such a problem on a graph, where vertices correspond to players, and a player can see each player to whom he is connected by an edge. First we prove some general theorems about the hat problem on a graph. Next we solve the problem for trees and cycles. Then we consider the hat problem on graphs such that the only know information are the degrees of vertices. Next we give an upper bound on the maximum chance of success for graphs with a neighborhooddominated vertex. Then we consider the problem on disconnected graphs and on graphs with a universal vertex.
Time: Thursday, February 26, 2015; 3:004:00PM
Location: SCC 130
Exploring Brain and Body Dynamics in Human Locomotion
by
Kristine Snyder
University of Michigan
Abstract:
While walking and running are everyday activities for most of us, researchers still have a rather limited understanding of how the brain and body interact to produce human locomotion. In this talk, I will offer some examples of how mathematics can be used to help explain the neural processes involved in walking and running. These include identifying the processes involved in minimizing energy expenditure, analyzing the effect of mechanical artifact on mathematical analyses of electroencephalography during walking, and using mathematical modeling to examine dynamic causality between neural sources. I will then discuss how mathematical models, experimental analysis, and the development of appropriate mathematical measures can interact to help us answer some remaining fundamental questions about neural activity during gait.
Time: Tuesday, February 17, 2015; 2:003:00PM
Location: MWAH 191
Fibonacci Sequence, Golden Ratio and Nature
by
Isabelle KemajouBrown
University of Minnesota  Twin Cities
Abstract:
In 1202, the Italian Mathematician Leonardo Bonacci known as Fibonacci, investigated how fast rabbits could breed in ideal circumstances. We shall use this example to introduce the Fibonacci sequence, describe more examples in real life where Fibonacci numbers and Golden ration occur naturally. We shall further show how to use simple mathematical tools to derive an explicit expression for Fibonacci numbers.
Time: Thursday, February 12, 2015; 4:005:00PM
Location: MonH 70
Analysis and Dynamics of Networked Systems
by
Bernard Chan
San Diego State University
Abstract:
Systems in various areas of science and engineering often have internal structures connecting numerous interacting components. Interactions between the subsystems may produce dynamics that are not predicted based on analyses of the individual parts. In application, such unexpected dynamics may produce undesired behavior. Therefore, it is important to understand the possible behavior by analyzing the system as a whole.
In this presentation, we will discuss various mathematical techniques that are used to analyze network models. Instead of statistical examinations of such systems, we will focus on the possible dynamics. Modeling, representation, bifurcation, synchrony patterns and other issues in coupled networks are discussed. We will also use an example from engineering to show that these techniques can be used to improve the performance of a system.
Time: Monday, February 9, 2015; 3:004:00PM
Location: MonH 80
Circular statistical methods in dispersion in of clusters in a cellcycle coupling model
by
Xue Gong
Ohio University
Abstract:
In this talk, the speaker will give an introduction on basic circular statistical methods to deal with angular data. Angular data are used in biology, meteorology, geography, and in many other areas; for examples, ocean current directions, orientation of fracture planes, and departure directions of animals. Because angular data have no true zero value, and the designation of high and low values is arbitrary, this type of data cannot be analyzed with commonly used statistical techniques. The speaker will use circular statistical methods to study the effects of random perturbations on collective dynamics of a large ensemble of interacting yeast cells. Two biologically motivated mechanisms of dispersion will be considered and compared with additive Gaussian white noise perturbations. The results can be used to predict the strength of coupling among the cells from experimental data.
Time: Thursday, February 5, 2015; 3:004:00PM
Location: BohH 90
by
Sylwia CichaczPrzenioslo
UMD
Abstract:
Martin Gardner described the {map coloring game} in Scientific American, which was first posed by Steven Brams in 1980 and independently by Hans Bodlaender in 1991. The map coloring game is a twoperson game played by Alice and Bob, who alternate moves. Alice begins by coloring a region of a map and then Bob colors a different region. They alternate coloring regions that have not been previously colored until all regions are colored. Alice tries to minimize the number of colors that need to be used, and Bob wants to maximize this number, such that no adjacent regions have the same color and no new color can be introduced unless it is forced because of adjacency.
We present some results on this topic.
Time: Tuesday, February 3, 2015; 3:004:00PM
Location: MWAH 195
by
Bethany Kubik
United States Military Academy
Abstract:
We study two different coloring problems through origami. First, we produce a flat foldable crane and explore coloring of the crease pattern, that is, the pattern made by the creases on the paper. Next, we explore colorability through threedimensional objects created with multiple sheets of folded paper.
*Please bring a writing utensil. Origami folding paper will be provided!
Time: Friday, January 30, 2015; 3:004:00PM
Location: SCC 120
How hard are old school Nintendo games?
by
Karl Wimmer
Duquesne University
Abstract:
In this talk, we will discuss the mathematics behind what it means for a problem to be "computationally hard." We relate this notion to the P=NP problem, which is one of the most important open questions in mathematics and computer science. We illustrate these ideas by showing that largescale versions of several games from Nintendo's largest franchises (Mario, Donkey Kong, Legend of Zelda, Metroid, and Pokémon) are computationally hard.
Time: Monday, January 26, 2015; 3:004:00PM
Location: ABAH 225
by
Sarah Anderson
Clemson University
Abstract:
How do QR codes such as the one shown below store information while incorporating logos or other cool designs? What makes them scan properly even if they are smudged or torn or the material on which they are written is damaged? This is just one example of error correction at work. Error control is essential in all digital storage and communications and is realized via encoding schemes from error correcting codes. ReedSolomon codes are algebraic geometry codes, a large family of error correcting codes used in a variety of applications and a topic of current research. In this talk, we explore the mathematics behind codes used for error correction.
Time: Friday, January 23, 2015; 3:004:00PM
Location: SCC 120
A Study of Longitudinal Analysis With Application
by
Katie Borchert
UMD M.S. Candidate
Abstract:
A longitudinal study is a study in which the researcher makes observations of the same variable on the same individual. This creates a problem of independence since the observations of one variable on one individual clearly are not independent. Therefore, we must adjust the statistical model in order to correct for this. The correction uses multivariate statistical techniques to create a variation of multiple linear regression. To demonstrate how this correction works we will provide an example using data on children’s development. The goal of the analysis is to determine which of a few factors, if any, have an effect on a child’s academic performance. We predicted that poverty level would have the greatest negative effect and breastfeeding would have the greatest positive effect. After conducting our analysis we found that over time poverty affects a child’s ability to do well in math by a coefficient of 8.54 (p = 0.0000). Breastfeeding increases their ability over time by 4.11 (p = 0.0409). Also, Reading ability turned out to have an effect over time as well at a rate of 0.3308 (p = 0.0000). Time spent with family and sleeping did not appear to have any effect.
Time: Friday, January 16, 2015; 2:004:00PM
Location: SCC 130
Periodic Behavior in a Class of Second Order Recurrence Relations Over the Integers
by
Brittany Fanning
UMD M.S. Candidate
Abstract:
Suppose we start a sequence with and , and define the rest of the sequence using the relation
The resulting sequence is 4, 13, 50, 22, 4, 140, 4, 13, … When a sequence repeats after terms, we call the solution periodic with period . Thus the sequence above is periodic with period six.
In general, given the system
where is a rational number, and are integers, we are interested in finding when periodic solutions occur. In other words, we want values and initial conditions for specific values of and which lead to a periodic solution. Using a common linear algebra method for solving recurrence relations, we developed a search method. The method and results will be outlined in this talk.
Time: Thursday, December 18, 2014, 2:003:00PM
Location: SCC 130
by
John Greene, Mathematics & Statistics, UMD
Abstract:
Time: Thursday, December 11, 2014, 3:004:00PM
Location: SCC 130
Results of an Experiment in Teaching Mathematics Comprehension to Young Students
by
William Krossner, Ph.D.
Abstract:
Could many (some?) (any?) 6^{th} grade U.S. students be taught to use, and even better, learn to be able to prove, one or two theorems in number theory hitherto presented, if at all, to upperlevel college mathematics majors? Come to this colloquium and find out. And, if the answer to any part of the first sentence is “Yes,” who cares, and what difference does it make?
Equivalently, are there any implications for the nature of the current math curriculum in the lower grades, or how U.S. mathematicians of the future might be discovered and nurtured? Would U.S. citizens in general reduce their high levels of “innumeracy” and “math phobia” if some modest number of activities such as the above can be inserted into their school experience? If so, might more people enter STEM vocations?
The point is not the theorems, but in developing the understanding by earlygrade students of the true nature of mathematics, as compared with the drill, drill, drill, calculate, calculate, calculate, learn next technique, learn next technique, learn next technique that constitutes the bulk of math education from kindergarten to middle school to high school to college majors such as engineering, you name it.
Relevant fact to consider: a 2014 recipient of the Fields Medal (for mathematicians, the equivalent of the Nobel Prize) was a 1995 member of UMD Professor Joseph Gallian’s annual summer math experience held at UMD. Could this relatively short exposure to advanced mathematics have helped to influence his decision as to a career choice?
Time: Thursday, December 4, 2014, 3:004:00PM
Location: Chem 150
From Soccer Tournament to Graph Theory and Back
by
Dalibor Froncek
Department of Mathematics & Statistics, UMD
Abstract:
In 2002, I was asked to schedule a rather unusual soccer tournament with 7 teams. After doing that, I started to wonder (as every mathematician would do) about a general solution for the existence of such tournaments. In 2004, along with my thenstudents Petr Kovar and Tereza Kovarova I wrote my first article about this topic. Since then, about 40 articles were published in this area.
Ten years and ten articles later, I am still interested in this exciting topic. Jerimi Walker wrote her Master's Project in this area and published her results in a journal. Michael Lillegard wrote his Master's Thesis on a related topic, and Aaron Shepanik has made a good progress in his Master's Thesis already. Sylwia CichaczPrzenioslo became interested in it when she was visiting our department in 2007 as Fulbright Junior Research Fellow, and by now is the leading author in this area with about 12 papers.
I will show how sports tournaments are related to distance magic graph labelings and present some techniques for constructing such tournaments, including those most recently obtained by Aaron. I will also mention some open problems that may be good topics for further student research.
Time: Thursday, November 20, 2014, 3:004:00PM
Location: SCC 130
Where’s the Math? BIG Mathematical Research Problems
By
Tracy A. Bibelnieks, Ph.D.
Abstract:
Amazing research opportunities for mathematicians can be found in every sector of business, industry and government (BIG). This colloquium will present a selection of current BIG mathematical research problems across a variety of topic areas within mathematics. The presentation and introduction to the mathematical techniques used in researching solution strategies for the problems will be accessible to all undergraduates at some point in their major or minor! Students will be invited to engage in the session at various points to brainstorm, innovate and unleash their mathematical creativity. As a further note, this colloquium is also intended to interest and invite students to consider diving deeper into BIG research during spring semester in MATH 4095.
On the wave equation with local KelvinVoigt dampin
By
Zhuangyi Liu
Abstract:
In this talk, we will investigate the properties of the solution to the wave equation with local KelvinVoigt damping. The properties of its solution turned out to be surprisingly different from the wave equation with viscous damping. The wellknown Geometric Optics condition does not apply in this case. We will show that the properties rely on the smoothness and the growth rate of the damping coefficient function at the interface.
Gaining Success Early in Your Career  Opportunities for Math/Stat Majors
by
Greg Ostergren
UMD Mathematics Graduate 1977
Abstract:
Greg Ostergren, Chairman, President and CEO of several insurance companies with over a billion dollars in annual revenue, will share why he believes that the opportunities for math/stat majors have never been brighter. These opportunities are expanding exponentially and globally, crossing into every industry and field of study. Reflecting upon both his personal experiences throughout his career and his leadership responsibilities as CEO and as a Board member of numerous organizations, Mr. Ostergren will provide his insight into how math/stat majors should position themselves for a highly successful and lucrative career. He will also discuss principles and ideas for career advancement, especially in the early stages of your career.
Multivariate Bayesian Logistic Regression
by
Xuran Yan
UMD M.S. Candidate
Abstract:
When the outcome of a study only has two values, people usually apply logistic regression model to analyze the data. The estimation of the parameters in logistic model involves NewtonRaphson Method, which also generates the variance of the parameters at the same time. However, sometimes there are additional prior information we want to take into consideration, then the Bayesian approach is to be employed. To get the posterior distribution of the parameters in Bayesian approach, Gibbs Sampling, a special case of MCMC algorithm is needed in the calculation. Applications to the drug safety assessing data of both approaches will be given.
Time: Monday, August 18, 2014, 1:003:00PM
Location: SCC 130
Probability Based Premium Calculation for LTCI
by
Shinjini Kar
UMD M.S. Candidate
Abstract:
Long Term Care (LTC) Insurance is a recent field of study. In this project we discuss the probabilistic model depicting rates of disability in the elderly population and how to use that model and the corresponding probability observations to calculate insurance premiums for covers that assure LTC coverage in the event of disability. The two key aspects of this study come from an Italian paper (Levantesi et al.) which discusses the mathematical model and gives us a set of data with which to work, and a German paper (Helms et al.) which provides us with a method of premium calculation. Some trends and reasonable values are observed from our calculations.
Time: Tuesday, July 25, 2014, 3:005:00PM
Location: SCC 130
A Study of Recommender Systems With Applications
by
Xiaowen Fang
UMD M.S. Candidate
Abstract:
In this project, we studied KNearest Neighbor (KNN), asymmetric Singular Value Decomposition (SVD) and Restricted Boltzmann Machine (RBM) methods in recommender systems. We applied these methods to three datasets: movielens100k, AmazonMeta and R3.Yahoo!Music. By comparing their results, we found that the recommender methods are very much data oriented. Since AmazonMeta was collected based on friend recommendation settings, we obtained the lowest root mean square error (RMSE) among the three datasets by using the KNN algorithm. The users in the AmazonMeta dataset are correlated, RBM has large RMSE compared to the KNN and asymmetric SVD methods. For R3.Yahoo!Music all three methods give moderate RMSE, but none of the methods give RMSE less than 1. For the movieLens100k data, asymmetric SVD shows the best prediction among the three methods.
Time: Friday, July 18, 2014, 2:004:00PM
Location: SCC 130
Symmetric Chain Decompositions of Partially Ordered Sets
by
Ondrej Zjevik
UMD M.S. Candidate
Abstract:
A partially ordered set, or poset, is a set of elements and a binary relation which determines an order within elements. Various combinatorial properties of finite and ordered posets have been extensively studied during the last 4 decades. The Sperner property states that the size of the largest subset of pairwise incomparable elements does not exceed the size of the largest level set in an ordered poset. Since a symmetric chain decomposition is a sufficient condition for the Sperner property, we may prove the Sperner property by finding a symmetric chain decomposition for a poset.
In this paper we focus on three types of posets: the Boolean algebra, Inversion poset and the Young’s lattice. An explicit construction for a symmetric chain decomposition is known only for Boolean algebras. No explicit construction has been found for Inversion posets and Young’s lattices, a symmetric chain decomposition was found only for a small subset of these posets. Using a maximal flow, we introduce an algorithm for finding this decomposition. We present our results and discuss two implementations of this algorithm.
Time: Tuesday, July 8, 2014, 3:005:00PM
Location: SCC 130
Estimation of Large Covariance Matrices using the POET Estimator
by
Michal Hrabia
UMD M.S. Candidate
Abstract:
Covariance matrix estimation is a standard problem of multivariate statistics. Nevertheless, in highdimensional case, the classical sample covariance estimator S is not consistent anymore, since its maximum eigenvalue is different from the maximum eigenvalue of the population covariance matrix. Usually, to deal with this issue, we assume sparsity of the covariance matrix. There are cases, however, when the sparsity assumption is inappropriate due to strong relationship between the covariates and when some other estimation methods need to be used.
In this project, we present an estimator that can be used in the case when covariance matrix is not sparse, the Principal Orthogonal Component Thresholding (POET) estimator, introduced by Fan, Liao and Mincheva in 2013. To form this estimator, we use the spectral decomposition of the sample covariance matrix. We keep the part of S formed by the first K principal components and perform thresholding on the remaining part. The final estimator is the sum of these two parts. We derive the POET estimator using two different approaches and present the theorem about the convergence of POET, by Fan, Liao and Mincheva. Finally, we show how using POET can improve stock portfolio allocation.
Time: Monday, June 16, 2014, 3:005:00PM
Location: SCC 130
S&P Effect and Investment Classification
by
Xiao Li
UMD M.S. Candidate
Abstract:
S&P effect refers to the turbulence of price movement exhibited by the stocks that have been newly added to S&P 500 index. In previous paper, positive intermediateperiod return is documented, while few touch the price reversal in the sixday period following the effective day.
In this paper, we first demonstrate the price reversal being significant which sheds light on lucrative short selling profit. Second, based on empirical evidence, we propose 8 explaining features and later we knock down to two best ones. In the end, we utilise SVM (support vector machine) to nonparametrically construct classification region.
Time: Thursday, June 12, 2014, 2:004:00PM
Location: SCC 130
by
Long Chen
UMD M.S. Candidate
Abstract:
Extreme value theory can be used to predict the occurrence of rare events, such as extreme food, large insurance losses, stock market crash, or human life expectancy.
In this project, we apply the extreme value theory to the athletic events.
For athletic events, we mainly focus on the estimation of best athletic performance in near future using extreme value theory. Two types of estimation methods will be used, namely, moment method and maximum likelihood method (MLE). We will give estimation of future athletic record using both methods and compare the results.
Time: Monday, June 9, 2014, 3:005:00PM
Location: SCC 130
Central Limit Theorem for Testing the Equality of Covariance Matrices
by
Wenchuan Guo
UMD M.S. Candidate
Abstract:
In this project, we are interested in testing for the equality of k covariance matrices of p dimensional multivariate normal distributions where the likelihood ratio test is used. The asymptotic distribution of the test statistics under different conditions is needed. In traditional multivariate analysis, the dimension p is considered to be a fixed constant that is unchanged as n approaches infinity. However, in practice, the dimension may be proportional to sample size n. In this project, we treat the dimension p as a function of sample size n and studied the hypothesis test under the general conditions that min ni/p ≥ c > 1. We also assume the number of populations k to be a variable that changes with sample size ni under the condition k/ni → ∞. The limit distribution of the test statistics when sample size n goes to infinity is a chisquare distribution which is invalid when dimension p and number of populations k are large. We derive the central limit theorem when p, k are variables. Numerical simulations for the two different approximations including histograms of simulated values and estimation of power and size are presented at the end of this project.
Time: Wednesday, May 21, 2014, 3:005:00PM
Location: SCC 130
Magic Boxes and Related Topics
by
Michael Lillegard
UMD M.S. Candidate
Abstract:
Magic boxes are a 3dimensional generalization of magic rectangles, which in turn are a classical generalization of the magic square. In this talk, two new generalizations of the magic box are introduced: the magic box set and the magic hollow box. Several necessary and sufficient conditions for the existence of these structures are examined, as well as conditions which preclude the existence of these structures.
Time: Tuesday, May 20, 2014, 3:005:00PM
Location: SCC 130
The Asymptotic Expansion of Eigenvalues for an Abstract System of Coupled Evolution Equations
by
Zhaobin Kuang
UMD M.S. Candidate
Abstract:
This project develops the methods and techniques to estimate and compute the asymptotic expansion of the roots of the following quartic equation about :
where, as , and the parameters are , , and . This quartic equation is the characteristic equation of the following abstract system of coupled evolution equations:
where is a selfadjoint, positive definite operator on a complex Hilbert space .
We will show that our work is an essential step towards a complete stability and regularity analysis of the abstract system. For root estimation, we first solve different quartic equations in the parameter space numerically. A piecewise linear regression algorithm is then employed to determine the pattern of the leading term of the roots in different regions of the parameter space as . An algorithm based on modified Taylor expansion is proposed to asymptotically expand the roots of the quartic equation. We also provide a simple proof for determining the leading term of the roots. Based on the leading terms of these eigenvalues sequences, we will provide reasonable conjectures on the stability and regularity of the solution to the abstract system. Order of polynomial stability as well as order of Gevrey class will also be conjectured. We will also prove the optimality of the order given that the conjectures proposed are true.
Time: Monday, May 19, 2014, 2:004:00PM
Location: SCC 130
Use of Chebyshev Polynomials to Construct and Hilbert Space Inner Product to Solve Eigenvalue Problems With High Accuracy
by
Zhengfei Rui
UMD M.S. Candidate
Abstract:
Based on Chebyshev polynomials, and Hilbert Space Inner Product functions, which satisfy the given boundary conditions, can be constructed to solve a eigenvalue problem. With such two basis functions, it is easy to discretize the given eigenvalue functions and solve the corresponding eigenvalues by exploiting the properties of Chebyshev polynomials. Moreover these two techniques lead to higher accuracy results than the Tau method.
Time: Thursday, May 14, 2014, 1:003:00PM
Location: Chem 153
Three Undergraduate Research Projects
by
Jasmine Helgeson, Chenxiao Hu and Jing Lv
Abstract:
In this colloquium three UMD undergraduate students will describe their research projects.
Jasmine Helgeson examined the extent to which policies have created differences in the amount of electrical power generated from wind turbines across the US.
Chenxiao Hu used the community college GPA of transfer students enrolled in UMD General Biology to create a model for predicting class grades. She will explain particularly interesting effects of SCSE major and gender.
Jing Lv studied how volume and the price/earnings ratio affects stock price based on company size.
Time: Thursday, May 8, 2014, 3:004:00PM
Location: MoH 203
Breaking Codes by Solving Polynomials: Algebraic Cryptoanalysis
by
Gregory Bard
University of Wisconsin  Stout
Abstract:
Algebraic cryptanalysis is the process of turning a cipher into a system of polynomial equations, and then solving the equations to obtain the secret key of the cipher. First, it is interesting to see how alldigital circuits, but especially hardware ciphers, can be encoded as polynomial systems of equations mod 2. Second, it also exciting to see a specific example, and so I will discuss my dissertation work on breaking the cipher Keeloq, which is used in nearly all automobiles with remote keyless entry. I will briefly touch on work done upon some other ciphers, and other noncryptographic applications.
Time: Thursday, May 1, 2014, 3:00PM
Location: SCC 130
OutofClass Mathematics Experiences
by
Philip Bauer, Jordan Maiers and Adam Swinney
Abstract:
In this colloquium two UMD students will describe their honors projects and a third will discuss his internship. Philip Bauer explored some properties of continued fractions; Jordan Maiers investigated the behavior of the Mandelbrot set; and Adam Swinney was a Travelers Insurance, oil and gas actuarial summer student.
Time: Thursday, April 24, 2014, 3:004:00PM
Location: MoH 203
Computational Investigations of Boundary Effects on CFD Simulations of Thermoacoustic Instabilities
by
Qingzhao Wang
Abstract:
The increasing awareness of environmental issues has been challenging gas turbine designers since early 90s. Lean premixed combustion is one of the strategies to achieve low NOx emission, and has been widely applied in gas turbines. However, near the leancombustion limit, the gas turbines are more susceptible to thermoacoustic instabilities, which may cause loud noise, violent vibration, structural destruction, and time and economic cost. Thus, it is desirable and essential to predict and control the occurrence of thermoacoustic instabilities. Many studies have been focusing on the Computational Fluid Dynamics (CFD) simulations to explore this phenomenon, but they either made simplifications and ignore some complex but important mechanisms, or required impractically long time for computations. In this talk, I will first present the key issues in simulating the coupling process of heat release rate and pressure fluctuations involved in thermoacoustic instability phenomenon. Then I will talk about the potential sensitivity analysis approaches that are applicable to investigations of boundary condition effects on thermoacoustic instabilities. The focus is on the formulation of Continuous Sensitivity Equation (CSE) method applied to the Direct Numerical Simulation (DNS) of thermoacoustic instability problems. This proposed sensitivity analysis approach only requires a single run of the CFD simulation. Moreover, the sensitivities of field variables, pressure, velocity and temperature to boundarycondition parameters are directly obtained from the solution to sensitivity equations. Thermoacoustic instability is predicted by the Rayleigh criterion and indicated by Rayleigh index. The sensitivity of Rayleigh index is computed utilizing the sensitivities of field variables. This approach is validated through the 1D thermally induced acoustics problem.
Time: Thursday, April 17, 2014, 3:00PM
Location: SCC 130
Mathematical Contest in Modeling 2014
by
Jasmine Helgeson, Penghuan Ni, Tim Schoenheider, Philip Bauer, Jesse Schmieg and Joe Toninato
UMD Mathematics & Statistics Undergraduate Students
Abstract:
Each February, a nationwide international Mathematical Contest in Modeling (MCM) is held. Contestants have 96 hours to select from one of two problems and submit a solution. This year, two teams represented UMD. One team selected PROBLEM A: The KeepRightExceptToPass Rule, to develop a model to determine whether this is an effective rule in promoting better traffic flow, or whether some other rule might be better. The other team selected PROBLEM B: College Coaching Legends, to build a mathematical model to choose the best college coach or coaches from among either male or female coaches in such sports as college hockey, field hockey, football, baseball, basketball, or soccer.
The teams will discuss the contest problems, their proposed solutions, and their overall experience with the competition.
Time: Thursday, April 10,2014, 3:004:00PM
Location: MoH 203
mregular partitions and etafunction symmetries
by
William Keith
Abstract:
A famous result in partition theory is Ramanujan's congruences, that the number of partitions of 5n+4 is divisible by 5, those of 7n+5 divisible by 7, and those of 11n+6 divisible by 11. These are now understood as members of an infinite family of such congruences, unified by the symmetries of modular forms. More recent work has been devoted to finding congruences for the mregular partitions, those in which parts may not be divisible by m. These are now numerous, but we do not yet have a similar unifying structure. This talk will outline each of these ideas, demonstrate that it is now fairly easy to prove many conjectured congruences with current techniques, and lay out a few ideas, tentative as yet, for constructing such general theorems.
Time: Thursday, April 3,2014, 3:00PM
Location: SCC 130
The Role of Statisticians in Medical Reserach
by
Ross Garberich
Abstract:
Statisticians play a critical role in medical research. They help design studies, validate and analyze data and write abstracts, manuscripts and reports for publication. Additionally, as hospitals and healthcare systems face continued pressure to reduce costs and improve patient outcomes, the roles statisticians’ play continues to expand, Aided by the implementation of electronic health records, statisticians are beginning to play key roles in reporting realtime analytics and predictive modeling for patients who are still admitted to the hospital.
The presentation gives a brief overview of the varying roles that statisticians play in the field of medicine, the different types of business that hire statisticians, and some examples of different projects on which I have worked.
Time: Thursday, March 27 ,2014, 3:004:00PM
Location: MoH 203
NonLinear Eigenvalue Problems
by
Thomas Cameron
Abstract:
The standard linear eigenvalue problem Av = λ v and the generalized linear eigenvalue problem Av = λBv are introduced in the latter part of undergraduate school and become well known in graduate school.
Moreover, we have successful numerical algorithms for finding both the eigenvalue λ and the eigenvector v, specifically the QR, QZ, and Francis algorithm to name only a few.
However, when it comes to nonlinear eigenvalue problems, for example
(Aλ^{2} + Bλ + C)v = 0
it is rare for a student to come across this problem in their course studies, even at the graduate level. Truthfully, the best numerical algorithms of the day, perhaps in defeat, simply linearize the problem and refer to our old friends, QR, QZ, and Francis.
We will take a look at some general properties of nonlinear eigenvalue problems, a brief history of the numerical methods for solving the linear eigenvalue problem, some nonlinear methods, and a Gersgorin like theorem for the bounds of the eigenvalues with some practical results.
Time: Thursday, March 13 ,2014, 3:004:00PM
Location: SCC 130
What is Theoretical Computer Science
by
Karl Wimmer
Abstract:
In this talk, we will give a survey of various areas of theoretical computer science and the role that mathematics plays in these areas. We will discuss some applications such as machine learning and voting theory, as well as some theoretical topics. We will also discuss some realworld problems (that we would really like to be able to solve), which appear to be computationally hard not only to solve exactly, but also to solve approximately.
The material presented in the talk will be heavily influenced by the problems studied during the ``Real Analysis in Computer Science'' program (which the speaker attended as a Research Fellow) held during Fall 2013 at the Simons Institute for the Theory of Computing at the University of CaliforniaBerkeley.
Time: Thursday, March 6,2014, 3:004:00PM
Location: MoH 203
Xuan Li
Abstract:
Regression is a statistical approach to investigating the relationship between variables. Regression techniques have been widely used in almost every field.
This talk will be given by STAT 5511 Regression Analysis students (2013 Fall semester) based on their class projects:
 Estimation of Risk Rates using Arbitrage Pricing Theory (Michal Hrabia)
 Lake Superior  clear and cold, almost good enough to drink (Tim Cyr, Yvette Ibrahim, Dave Ongaro, Kelly Peterson, Andrea Samuelson, Miranda Steinmetz)
 Effects of Minimum Wage on Inflation (Katherine Borchert, Shinjini Kar, Cole Mathson)
 Cases of Infectious Diseases: Determinants of the Number of Deaths due to Tuberculosis across Nations (Matthew Arthur, Jasmine Helgeson, Xiao Li, Penghuan Ni, Kyle Vezina, Zichao Wang)
Time: Thursday, February 20,2014, 3:004:00PM
Location: MoH 203
Combinatorial Problems Motivated by Databases
by
Uwe Leck
Abstract:
Some combinatorial problems and results will be discussed that arise in the context of restoration of lost information in distributed databases. Consider a set T of lattice points in a k x k grid and call it a configuration. You can think of the points in T as faulty nodes that need to be repaired or decoded. Performing a step of decoding means transforming T into a new configuration T’ by removing all points belonging to some horizontal or vertical line L, under the constraint that only t points of T are on L (where t is some given number). T is decodable if it can be transformed into the empty set by an appropriate sequence of decoding steps. Examples of interesting questions in this context are: What is the largest size of a decodable configuration? Among all decodable configurations with the same given number of points, which are the hardest to decode (i.e., which require the most decoding steps)? What are the smallest decodable configurations that require some given number of decoding steps? How does all this generalize to higher dimensional grids?
Time: Thursday, February 6,2014, 3:004:00PM
Location: MoH 203
Recent Advances in Rankbased Inference for Accelerated Failure Time
by
Steven Chiou
Abstract:
Semiparametric accelerated failure time (AFT) models have not been used as frequently as Cox relative risk models in such settings due to lack of efficient and reliable computing routines for inferences. The challenge roots in the nonsmoothness of the rankbased estimating functions. The recently proposed induced smoothing approach, which provides fast and accurate rankbased inferences for AFT models. The induced smoothing approach is generalized to incorporate weights to accommodate different rank weights and additional weights from missing data and various sampling schemes. The variances are estimated with an efficient resampling approach that avoids solving estimating equations repeatedly. I will demonstrate the methods in a numerical study. As expected, the proposed estimators were obtained much faster without loosing accuracy in comparison to those from nonsmooth estimating equations, and the variance estimators provided good approximation of the variation in estimation.
Time: Thursday, January, 30 2014, 3:004:00PM
Location: 130 SCC
Kyle Krueger and Brett Bozyk
Abstract:
The world of retail is evolving quickly. Amazon.com has forced the brickandmortar chains to evolve rapidly, and the guiding hand that steers that evolution is analytics. A retail analyst can use massive data sets within tools like SAS and Excel to answer questions about what should go on the shelves, where it should go, how much space should be allocated, and more. Metrics like exclusivity, penetration, and lift can better inform decisions made by the business on a daily basis, and ensure that a company stays relevant to its consumer base.
Time: Thursday, December 12, 2013, 3:004:00PM
Location: 150 Chemistry
Mathematics courses in various parts of the world and study abroad recommendations for UMD mathematics and statistics students
by
Lucas Gloege, Inne Singgih and Ondrej Zjevik
Abstract:
Spending a term abroad can be very expensive and it can delay a student’s progress toward graduation. However, for many students these costs are more than balanced by eyeopening and life changing experiences they have while studying in a foreign country. Planning ahead is important for reducing the negatives and increasing the positives of a semester spent abroad. This includes, in part, determining what courses to take at UMD and what courses to take abroad. Since university mathematics and statistics education is not uniform across the world, foreign course syllabi and even UMD course equivalences provide only partial information.
In this colloquium three UMD graduate students, each of whom majored in mathematics as an undergraduate, will talk about their educational experiences in three different parts of the world. Inne Singgih will talk about her bachelor’s degree program at the University of Indonesia in West Java Province and her work as a mathematics educator in Indonesia. Ondrej Zjevik will talk about his bachelor’s degree program at VSB – Technical University of Ostrava, Czech Republic and how it compares with the mathematics programs at UMD. Lucas Gloege spent a semester abroad at Waikato University in New Zealand, while earning his BS from UMD. He will talk about his preparations for studying abroad, the courses he took at Waikato U. and his experiences in New Zealand. He will also give studyabroad recommendations for UMD mathematics and statistics students.
Time: Thursday, November 21, 2013, 3:004:00PM
Location: 150 Chemistry
Making math fun for mathphobic children
by
Ken Stanley, Director, Dubois Project of Oberlin College
Abstract:
The United States lags many developing countries in educational performance, particularly in our education of children of underrepresented minorities and children from families of lower socioeconomic status. Local funding for schools is part of the problem as students from lowincome families disproportionately attend poorly funded schools, however, local funding does not explain why these children perform poorly in all schools. In Oberlin, Ohio children from all backgrounds attend the same K12 schools, yet the achievement gap persists. Motivation, especially intrinsic motivation, drives educational achievement. The Du Bois Project helps underrepresented minorities and children from low socioeconomic families achieve and maintain excellence in the Oberlin Public Schools by helping them learn to enjoy math. We will explain how we use running, music, simple psychology and motivated Oberlin College students to make math fun.
Time: Thursday, November 15, 2013, 3:004:00PM
Location: 70 Montague Hall
Hypercube orientations with degree restrictions, or what color is my hat?
by
Steve Butler, Department of Mathematics, Iowa State University
Abstract:
There are many variations of hat guessing games, which generically consist of a group of players forming a strategy about how each player will go about guessing the color of their own hat by using information gleaned from looking at the hats of other players (but not their own). Strategies in this game can be rephrased in terms of hypercube orientations with given degree restrictions. We will look at several problems in this direction and give some open problems.
Time: Thursday, November 14, 2013, 3:004:00PM
Location: 130 Solon Campus Center
Why is Tuition So High at UMD?
by
Richard Green, Department of Mathematics, UMD
Abstract:
Tuition at UMD is now 55 times as high as it was in 1960, when I was a student here. Adjusted for inflation, tuition is seven times as high as it was in 1960. In this talk I will demonstrate the fact of tuition increase and I will discuss some of its causes and consequences. I will give three reasons why tuition at UMD should be no more than onethird of what it is now. I will also discuss some recent literature on the management and funding of American higher education, including the radical, democratic idea that public colleges and universities should be tuitionfree.
Time: Thursday, November 7, 2013, 3:004:00PM
Location: 150 Chemistry
Independent Variable Rescale to Solve Parabolic PDEs with Periodic Boundary Conditions Containing State Dependent Coefficients
by
Brian Hinderliter, Department of Mechanical and Industrial Engineering, UMD
Abstract:
Developing a test procedure that can predict the corrosion protection of a coating designed to last for 20 years is challenging, particularly accurately predicting behavior over decades with tests that last months. Fick’s second law of diffusion (parabolic PDE) is applied to coating surfaces exposed to varying surface water concentrations and temperatures (boundary conditions) at cyclical intervals. The diffusion coefficient is temperature dependent, in this example, of an Arrhenius form. The governing equations are solved using Laplace transforms after time rescaling to account for the variation in temperature, which impacts the diffusion coefficient between the wet and dry portions of several accelerated weathering protocols. This serves to predict the asymptotic water concentration in the coating as well as at the coatingmetal interface based on physical constants of the coating and parameters of the accelerated weathering protocol of interest. The analytic solution to Fick’s second law also allows the water concentration at the substratecoating interface to be predicted based on Arrhenius parameters of the diffusion process. This resultant average time of wetness and water concentration variance at the coatingsubstrate interface can be used to more directly compare corrosion and adhesion loss between various accelerated weathering protocols and natural weathering conditions.
Time: Thursday, October 31, 2013, 3:004:00PM
Location: 130 Solon Campus Center
Attitudes and prestige systems among Varieties of Kriol Language in Belize
by
Ron Regal, Department of Mathematics, UMD
William Salmon, Department of Linguistics, UMD
Abstract:
Belize, formerly British Honduras, gained full independence from Great Britain in 1981. While there is a strong influence of English as a result of its colonial history, Belize is linguistically diverse. In an ongoing project, Jennifer Gomez Menjivar, Foreign Languages UMD, and William Salmon, Linguistics, UMD, are examining the attitude and prestige system in place among varieties of Kriol, an Englishbased creole. As part of their study, Drs. Menjivar and Salmon had respondents rate speakers from Belize City and Punta Gorda, two different cities in Belize, on 16 personality attributes. Will Salmon will describe their study, and Ron Regal will tell some about considerations statistical data analysis and an analysis in this study in particular.
The talk will give an introduction to one aspect of linguistics and will be particularly of interest to students potentially interested in double majors in mathematics/statistics and linguistics. Study reports on language attitudes toward varieties of Belizean Kriol in coastal Belize. We used a verbalguise test with 82 participants, collecting both quantitative and qualitative data in Punta Gorda and Belize City, and we found that the variety of Kriol spoken in Belize City is perceived along several dimensions as being of greater prestige than the variety spoken in Punta Gorda. Derivative of these findings is the potentially more interesting fact that there is more than one variety of Kriol spoken in Belize in the first place—a fact which has not been previously reported in the literature. This research is part of an ongoing project investigating the overt and covert linguistic prestige system in place with respect to Belizean Kriol. The larger project will break down attitude factors among the individual ethnic groups in Belize in a series of articles.
Time: Thursday, October 24, 2013, 3:004:00PM
Location: 150 Chemistry
David Clark, Department of Mathematics, UMTC
Abstract: The seven siblings at the Clark farm have a massive list of chores to do. How can they divide them up fairly? Can they divide them fairly? Luckily, designs can help! Designs are a combinatorial object, which encode fairness, balance, and geometric structure. They are useful in a remarkable variety of situations, from dividing chores fairly through designing statistical experiments. One of the key modern uses of designs is to provide flexible and efficient errorcorrecting codes, which protect data transmissions from errors. We will define designs, give many examples, and study how their representation as (0,1)matrices provides a surprising link to cutting edge communications.
Time: Thursday, October 17, 2013, 3:004:00PM
Location: 130 Solon Campus Center
Math Majors and Information Security in a Digital Age: Any Connection?
by
Dr. William Krossner
Abstract: We consider the usual model of a sender of information, a channel of communication, and the intended receiver of this information. In the digital age, what role does mathematics play in rendering the transmitted information secure from the eyes of unintended or malicious receiving entities (such as spies, governments, corporations, cybercriminals, and the like)? Of course, for centuries various cryptographic methods (codes and ciphers) have been invented and used to try to keep information secret.
In the 20th century, the design of cryptographic methods was taken over by mathematics, so much so that today there are many job openings for math majors and graduates in both governmental and commercial entities around the world to deal with information security issues. The speaker will briefly outline one of the earliest allmathematical cipher methods (from 1929), but spend most of the time demonstrating the verylittleknown fact that within each computer operating system (Windows, Mac or Linux) there is a cryptographic laboratory available for use, powerful enough to produce cipher methods difficult for even supercomputers to solve. Any math major is smart enough to use this laboratory, and everything connected with it is free.
Audience members who bring their own USB drives may copy all of the speaker's apps for their own use at the end of the talk.
Time: Thursday, October 10, 2013, 3:004:00PM
Location: 150 Chemistry
Families of Infinite Series With Interesting Limiting Structure
by
John Greene, Department of Mathematics, UMD
Time: Thursday, October 3, 2013, 3:004:00PM
Location: 130 Solon Campus Center
Mathematics: The Science of Patterns
by
Joseph Gallian
Professor of Mathematics, UMD
Abstract: I will discuss three research projects involving patterns I have worked on that are accessible to a broad interdisciplinary audience. One was to figure out the algorithms used by computers to check the validity of product identification numbers on bar codes, credit cards, and books. Another entails deciphering the secret algorithm used by Minnesota and other states to code driver's license numbers. The third was to devise algorithms to create intricate symmetry patterns.
Although intellectual curiosity is the motivation for all my research, the results sometime yield unanticipated applications and unexpected media attention.
These will be described in the presentation.
Time: Friday, September 20, 2013, 3:004:00PM
Location: 70 Montague Hall
Steiner problems, soap films, minimal surfaces, and complex analysis
by
Michael Dorff
Professor of Mathematics, Brigham Young University
Abstract: We will start with a Steiner problem that ask the question what is the shortest path between several points. Then we will transform Steiner problems up one dimension to minimal surfaces that can be modeled by soap films demonstrating a few handson examples. As we do this we will discuss some differential geometry as background for minimal surfaces, diverge into complexvalued harmonic mappings, and discuss the connection between minimal surfaces and complexvalued harmonic mappings. Finally we will discuss some new research using complexvalued harmonic mappings to construct harmonic mappings.
Time: Thursday, September 19, 2013, 4:30PM
Location: 130 Solon Campus Center
How Math is Changing the World
by
Michael Dorff
Professor of Mathematics, Brigham Young University
Abstract: In Oct 2010, an article called "How much math do we really need?" was published in the Washington Post. The author, a mathematician, wrote, "Unlike literature, history, politics and music, math has little relevance to everyday life" and "All the mathematics one needs in real life can be learned in early years without much fuss." Is this true? Have you ever been asked, "What can you do with a degree in math?" Except for teaching, many people are clueless on what you can do with strong math skills. In this talk, we will discuss some of the exciting things mathematicians in business, industry, and government are doing in their careers and how these things are changing the world. And we will reveal the three things that recruiters say every math student should do to get a job.
Time: Thursday, September 19, 2013, 3:004:00PM
Location: 150 Chemistry