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.

- Archived 2012-13 Colloquia
- Archived 2011-12 Colloquia
- Archived 2010-11 Colloquia
- Archived 2009-10 Colloquia
- Archived 2008-9 Colloquia
- Archived 2007-8 Colloquia
- Archived 2006-7 Colloquia
- Archived 2005-6 Colloquia

Type | Date | Title | Speaker |

G | 9/4/09 | Spatially Correlated Mixture Models with Application in Genomic Hypotheses Testing | Jie Ren, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Kang James, Department of Mathematics and Statistics, UMD |

G | 8/17/09 | Active Control of Impulsive Noise | Ronghua Zhu, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Zhuangyi Liu, Xun Yu, Department of Mathematics and Statistics, UMD |

G | 8/15/09 | Numerical Analysis of Cooling Effects of a Cylinder Head Water Jacket | Qingzhao Wang, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Steven Trogdon, Department of Mathematics and Statistics, UMD |

G | 6/26/09 | Penalized Model-Based Clustering | Wei Zhang, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Kang James, Department of Mathematics and Statistics, UMD |

G | 6/24/09 | Creating Custom Matrices for Organisms with an A+T Rich Genome | Seth Slettedahl, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Marshall Hampton, Department of Mathematics and Statistics, UMD |

G | 6/18/09 | Effects of Monetary Rewards on Reliability of Health Assessment Surveys | Lucas Streng, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Ronald Regal, Department of Mathematics and Statistics, UMD |

G | 6/16/09 | Optimum Mixed Volume for Polynomial Sets | Annette Katich, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Marshall Hampton, Department of Mathematics and Statistics, UMD |

U | 5/7/09 | DEVELOPING A STORM WATER UTILITY | Derek Carlson, Undergraduate Student, UMD |

G | 4/30/09 | SCALING LAWS AND SIMILARITY CONCEPTS APPLIED TO THE ANALYSIS OF STABILITY OF SURFACE AND UNDERGROUND EXCAVATIONS IN ROCK MASSES | Dr. Carlos Carranza-Torres, Associate Professor of Geotechnical Engineering, Department of Civil Engineering, UMD |

U | 4/28/09 | Analysis of Traffic Circles | Jerod Lass,UMD Math/CS Major |

U | 4/23/09 | UNDERGRADUATE HONORS COLLOQUIUM | Jeff Anderson: "What Makes a Player 'Most Valuable': A Statistical Analysis?" Amy Schmidt: "Development of Mathematica Software Tools for Analysis of Differential Equations Systems" Xi Xi: "Does the Capital Asset Pricing Model Work In The Current Financial Turmoil?" Jerod Lass: "Analysis of Traffic Circles" (to be presented in a Computer Science Colloquium, Tuesday, April 28, 3:00 p.m. in SCC 21) |

G | 4/16/09 | NTERACTIVE DRAWINGS OF ALGEBRAIC SURFACES | Professor Joel Roberts, School of Mathematics, University of Minnesota, Minneapolis, Minnesota |

G | 4/10/09 | PENALIZED REGRESSION WITH NETWORKED PREDICTORS AND ITS APPLICATION TO eQTL ANALYSIS | Professor Wei Pan, School of Public Health, University of Minnesota, Minneapolis, Minnesota |

U | 4/9/09 | Math and the Markets | Joseph Artim, Instructor, Department of Finance and Management Information Sciences, UMD |

U | 4/2/09 | MATHEMATICAL CONTEST IN MODELING 2009 | Brian Christner, Matt Holst, Aaron Potvien, Jeff Anderson, Carl Green, Jerod Lass, Undergraduate Students, Department of Mathematics & Statistics, UMD |

U | 3/26/09 | Behind the Scenes of Mobius Transformations Revealed | Jonathan Rogness, Assistant Professor, Department of Mathematics, University of Minnesota Twin Cities |

G | 3/12/09 | MODELING OF THE ROBOTIC "PRESSING AND SLICING" CUT OF BIO-MATERIALS | Dr. Debao Zhou, Department of Mechanical and Industrial Engineering, UMD |

U | 3/5/09 | Some Math Behind M.C. Escher's "Circle Limit" Patterns | Doug Dunham, UMD Department of Computer Science |

G | 2/26/09 | CLASSICAL MARKOV CHAINS: A UNIFYING FRAMEWORK FOR UNDERSTANDING AVIAN REPRODUCTIVE SUCCESS | Mathew Etteson, Ecologist, U.S. Environmental Protection Agency, Duluth, Minnesota |

U | 2/19/09 | Benford's law, or the first digit phenomenon | Dick Green, UMD Department of Mathematics and Statistics |

U | 2/5/09 | Using Mathematics to Create Symmetry Patterns | Joe Gallian, University of Minnesota Duluth |

G | 12/18/08 | GENERALIZED ADDITIVE MODEL (GAM) | Huimin Liu, MS Candidate: Program in Applied and Computational Mathematics, Advisor: Kang James, Department of Mathematics and Statistics, UMD |

G | 12/12/08 | GLOBAL EXISTENCE AND STABILITY OF A QUASILINEAR WAVE EQUATION | Professor Qiong Zhang, Department of Mathematics, Beijing Institute of Technology |

U | 11/20/08 | The daily life of an actuary | Brad Diaz, Actuary at Ingenix/Redden and Anders |

U | 11/10/08 | Pricing and the Central Limit Theorem | Scot Adams, Professor of Mathematics, University of Minnesota Twin Cities |

G | 11/06/08 | MACHINE LEARNING VIA BILEVEL OPTIMIZATION | Gautam Kunapuli, Postdoc, University of Wisconsin - Madison |

U | 10/30/08 | Free Math! The Sage software project for free and open-source computation | Marshall Hampton, UMD Mathematics Department |

G | 10/24/08 | NECESSARY AND SUFFICIENT CONDITIONS FOR THE ASYMPTOTIC DISTRIBUTION OF THE LARGEST ENTRY OF A SAMPLE CORRELATION MATRIX | Deli Li, Department of Mathematical Sciences, Lakehead University, Thunder Bay, Ontario, Canada P7B5E1 |

G | 10/16/08 | FUNCTIONAL LOCALIZATION OF CALCIUM HANDLING MACHINERY | Steven J. Cox, Department of Computational and Applied Mathematics, Rice University, Houston, Texas |

U | 10/09/08 | Actuarial Summer Intern Experiences | Jeff Anderson, Dan Boeder, Shannon Glonek, Brian Hinkle |

U | 10/03/08 | My Life with the Queen | William A. Lokke, Former Deputy Director of the Lawrence Livermore National Laboratory, Departmental nominee to the SCSE Academy, Livermore, California |

U | 9/25/08 | Provably Hard and Provably Impossible Tasks | Steve Rosenberg, Associate Professor, University of Wisconsin--Superior, Mathematics and Computer Science Department |

U | 9/18/08 | Sex ratios in parasitic wasps | Dick Green, Department of Mathematics UMD |

**Sex ratios in parasitic wasps**

Dick Green

Department of Mathematics, UMD

**Abstract:**

The study of sex ratios and, more generally, sex allocation has provided some of the most successful examples of modern evolutionary theory. Many Hymenoptera (bees, ants and wasps) have the ability to determine the sex of their offspring: unfertilized eggs become sons while fertilized eggs become daughters. I will describe several ways in which parasitic wasps use their ability to choose the sex of their offspring in order to solve problems presented by nature. I am particularly interested in how mated female wasps should allocate the sexes of their offspring to hosts of different sizes when some females are unmated and produce only sons. This talk uses mathematical and statistical ideas that are familiar to anyone who has taken the first courses in those subjects.

**Time:**Thursday, September 18, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus center 130

**Provably Hard and Provably Impossible Tasks**

Steve Rosenberg

Associate Professor

University of Wisconsin--Superior

Mathematics and Computer Science Department

**Abstract:**

Did you ever feel frustrated by a math problem? Are some problems really "harder" than others? Yes! By making the notion of task precise and then measuring quantitatively how efficiently each type of task can be accomplished, we arrive at a notion of the intrinsic level of difficulty of a problem. We can even prove that some types of problem are impossible to solve, in the sense of finding a procedure that works every time. For those in the know: we will be talking about formal algorithms, time-complexity, and the Halting Problem. In some small way, these results can comfort us in the future when we feel that our problems are too hard! No special knowledge of computer science is assumed.

**Time:**Thursday, September 25, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus center 130

**My Life with the Queen**

William A. Lokke

Former Deputy Director of the Lawrence Livermore National Laboratory

Departmental nominee to the SCSE Academy

Livermore, California

**Abstract:**

Mathematics is often called the queen of science. Besides its traditional role in structuring theory, mathematics in its pure form can also be directly beneficial to the other sciences. Alan Turing.s insight in considering Hilbert.s decision problem, which led to today.s pervasive computer, delivered perhaps the greatest gift the field has yet given to science. In tracing my history of service to science.s king, physics, at California.s Lawrence Livermore National Laboratory, I will point out the crucial role computing has played in my career and in the laboratory.s many programs - particularly nuclear weapons. I will present a model for making most effective use of computing, describe some interesting challenges awaiting its application, and admit to certain difficult mathematical obstacles to be overcome.

**Time:**Friday, October 3, 2008 3:30- Colloquium

**Location:**Chemistry 150

**Actuarial Summer Intern Experiences**

Jeff Anderson

Dan Boeder

Shannon Glonek

Brian Hinkle

UMD

**Abstract:**

Four UMD students will present about their experiences as actuarial summer interns this summer in the Twin Cities. They will talk about topics ranging from the job search, to interviewing, to the projects they worked on over the summer. There will also be a question and answer period at the end of the presentations. The interns and their companies are Jeff Anderson, Securian Financial; Dan Boeder, CG Consulting; Shannon Glonek, Thrivent Financial; and Brian Hinkle, Securian Financial.

**Time:**Thursday, October 9, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus Center 130

**FUNCTIONAL LOCALIZATION OF CALCIUM HANDLING MACHINERY**

Steven J. Cox

Department of Computational and Applied Mathematics

Rice University

Houston, Texas

**Abstract:**

Calcium, the most important of the second messengers, sculpts and records synaptic input and modulates the excitability of both nerve and muscle. We exploit the recent ability to dynamically monitor cytosolic calcium, throughout rat hippocampal pyramidal cells in slice, with sub-millisecond temporal resolution and sub-micron spatial resolution in the construction of a functional map of calcium and potassium channel density.

In the process we pose and solve a number of inverse problems associated with dye recordings following focal uncaging of intracellular calcium and suprathreshold current injection. In particular, we 1) Infer from the change in cytosolic dye-buffered calcium fluorescence the dye binding rates and the calcium pump rate. 2) Infer from (1) the calcium current associated with back propagating action potentials and infer from this the associated channel densities.

**Time:**Thursday, October 16, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus center 130

**NECESSARY AND SUFFICIENT CONDITIONS FOR THE ASYMPTOTIC DISTRIBUTION OF THE LARGEST ENTRY OF A SAMPLE CORRELATION MATRIX**

Deli Li

Department of Mathematical Sciences

Lakehead University

Thunder Bay, Ontario, Canada P7B5E1

**Abstract:**

Let be a double array of nondegenerate i.i.d. random variables and let be a sequence of positive integers such that is bounded away from 0 and . In this paper we give the necessary and sufficient conditions for the asymptotic distribution of the largest entry of the sample correlation matrix where denotes the Pearson correlation coefficient between and . Write and . Under the assumption that for some , we show that the following six statements are equivalent:

where and . The equivalences between (i), (ii), (iii), and (v) assume that only . Weaks laws of large numbers for are also established and these are of the form and ,respectively. The current work thus provides weak limit analogues of the strong limit theorems of Li and Rosalski as well as a necessary and sufficient conditions for the asymptotic distribution of obtained by Jiang. Some open problems are also posed.

**Time:**Friday, October 24, 2008 4:00- Colloquium

**Location:**Solon Campus center 130

**Free Math! The Sage software project for free and open-source computation**

Marshall Hampton

UMD Mathematics Department

**Abstract:**

Three years ago William Stein released the first version of Sage, a free and open source computational platform based on the popular scripting language Python. Sage unifies a huge collection of mathematical software projects into a coherent and powerful system for mathematics, statistics, and scientific computation. In this talk I will give a brief overview of its present capabilities and future directions.

**Time:**Thursday, October 30, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus center 130

**MACHINE LEARNING VIA BILEVEL OPTIMIZATION**

Gautam Kunapuli

Postdoc

University of Wisconsin - Madison

**Abstract:**

Support Vector Machines have been applied with great success in machine learning. Support vector classification and regression are usually formulated and regarded as convex machine learning problems. However, the use of these methods entails solving several other related problems such as parameter and feature selection in conjunction with the central machine learning problem.

In fact, these problems are far from being convex: they depend on hyper-parameters like the regularization parameter and the kernel parameter whose optimal values need to be determined judiciously in order to get the best generalization behavior. When viewed from the perspective that the hyper-parameters are also variables in the model, the machine learning task becomes non-convex.

A novel approach is proposed where in cross validation is formulated as a continuous bilevel program which is a mathematical program whose constraints are functions of optimal solutions of another mathematical program. This powerful approach unifies parameter and feature selection into the central task of learning such that it is possible to simultaneously determine, not only the optimal model, but also the optimal parameters and features. Numerical results demonstrate that this approach is superior to conventional grid search approaches.

**Time:**Thursday, November 6, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus center 130

**Pricing and the Central Limit Theorem**

Scot Adams

Professor of Mathematics

University of Minnesota Twin Cities

**Abstract:**

We'll discuss "risk-neutral" pricing and how that concept combines with the Central Limit Theorem to lead to Black-Scholes.

Note: Professor Adams is the director of the Financial Mathematics program at UMNTC and will be happy to discuss that program with interested students

**Time:**Monday, November 10, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus center 130

**The daily life of an actuary**

Brad Diaz

Actuary at Ingenix/Redden and Anders

**Abstract:**

Brad Diaz is a UMD graduate and an actuary at a health care consulting firm that is part of United Health Group. He will talk about the everyday responsibilities of a health care actuary, the implications of a consulting role in the health care sector, and the importance of actuarial exams. There will also be time for questions at the end.

**Time:**Thursday, November 20, 2008 3:00-4:00 Colloquium

**Location:**Solon Campus center 130

**GLOBAL EXISTENCE AND STABILITY OF A QUASILINEAR WAVE EQUATION**

Professor Qiong Zhang

Department of Mathematics

Beijing Institute of Technology

**Abstract:**

A degenerate nonlinear dissipative wave equations of Kirchhoff type with boundary damping is studied. We prove that this system has a unique global solution if the initial data satisfied suitable assumption. We also reach the polynomial decay of solution of the system.

**Time:**Friday, December 12, 2008 4:15-

**Location:**Solon Campus center 130

**GENERALIZED ADDITIVE MODEL (GAM)**

Huimin Liu

MS Candidate: Program in Applied and Computational Mathematics

Advisor: Kang James

Department of Mathematics and Statistics

UMD

**Abstract:**

For years, researchers have been relying on Generalized Linear Model (GLM) to do their everyday research. As great as GLM is, it does not work in every situation; it fails to capture the nonlinearity in some cases, and therefore possibly not leading to satisfactory results. Generalized Additive Model (GAM) is introduced to compensate some of the disadvantage associated with GLM. In my presentation, the basic theory of GAM by Hastie and Tibishirani will be introduced. The advantage of GAM will be illustrated by examples. The fitting of GAM using SAS will be discussed. In addition, Generalized Additive Logistic Model together with an example will be given.

**Time:**Thursday, December 18, 2008 2:00 p.m. - 3:00 p.m.

**Location:**Solon Campus center 130

**Using Mathematics to Create Symmetry Patterns**

Joe Gallian

University of Minnesota Duluth

**Abstract:**

We use video animations to illustrate how mathematics can be used to create computer generated symmetry patterns. Discrete math, exponential functions, logarithms and modular arithmetic are used to transform basic images into symmetry patterns. These methods were used to create the image for the 2003 Mathematics Awareness Month poster. The talk is intended for a general audience.

**Time:**Thursday, February 5, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**Benford's law, or the first digit phenomenon**

Dick Green

UMD Department of Mathematics and Statistics

**Abstract:**

When written in decimal form, almost all real numbers have approximately equal concurrences of the digits 0, 1, 2,..., 9. The leading digit cannot be 0, but one might expect that numbers encountered in the world would be equally likely to begin with each of the digits, 1, 2, 3,..., 9. However, for many kinds of data (stock prices, street addresses, river lengths) the leading digit is more often 1 than 2, more often 2 than 3 and so on. In 1938, Frank Benford proposed a distribution for first digits that follows the distribution of their common logarithms. I will describe Benford's law and explain why it is often likely to hold, at least approximately. I will talk about other empirical laws related to Benford's, including those of Zipf and Lotka, and I will say a few words about eponymous "laws" like "Benford's".

**Time:**Thursday, February 19, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**CLASSICAL MARKOV CHAINS: A UNIFYING FRAMEWORK FOR UNDERSTANDING AVIAN REPRODUCTIVE SUCCESS**

Mathew Etteson, Ecologist

U.S. Environmental Protection Agency

Duluth, Minnesota

**Abstract:**

Traditional methods for monitoring and analysis of avian nesting success have several important shortcomings, including 1) inability to handle multiple classes of nest failure, and 2) inability to provide estimates of annual reproductive success (because birds can, and typically do, make multiple nest attempts in a year). Both of the above limitations may be relaxed by treating the avian nesting process as a Markov chain. In the first case the transition matrix is assumed unknown and maximum likelihood estimates for the transition probabilities are easily obtained. In the second case the asymptotic behavior of Markov chains may be used to gain remarkable insight from fairly simple hypotheses about the transition probabilities. With the above underpinnings, I will describe and demonstrate time-hetergeneous Markov chain models that we are currently developing at the EPA lab here in Duluth. Our primary objective in developing these models is to provide a regulatory tool for integrating laboratory testing data on adverse reproductive effects with avian life history information to inform ecological risk assessments.

**Time:**Thursday, February 26, 2009 3:00--

**Location:**130 Solon Campus Center

**Some Math Behind M.C. Escher's "Circle Limit" Patterns**

Doug Dunham

UMD Department of Computer Science

**Abstract:**

The Dutch graphic artist M.C. Escher was the first person to create designs that could be interpreted as repeating patterns in the hyperbolic plane. He called them "Circle Limit" patterns since they were drawn inside a circle and had motifs that grew ever smaller as they approached that bounding circle. Some mathematicians have analyzed the symmetries of these patterns, giving their symmetry groups, which we will explain. The "Circle Limit IV" pattern seems to be mis-oriented frequently. We will show the correct orientation. Finally, we will show how to make your own circle limit patterns using mathematics and computer graphics.

**Time:**Thursday, March 5, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**MODELING OF THE ROBOTIC "PRESSING AND SLICING" CUT OF BIO-MATERIALS**

Dr. Debao Zhou

Department of Mechanical and Industrial Engineering

UMD

**Abstract:**

The applications of robotics are becoming more and more common in non-traditional industries such as the medical industry and food industry where the technology of cutting biomaterial is used. Automating this process using robotic devices with closed-loop force control has shown promise. The control of the force trajectory directly relates to the internal stress in the material during cutting. The ability to model the stress distribution in the bio-materials being cut would provide a better understanding of the influencing factors and help predict the required cutting force for the design of the cutting mechanism and for automating the cutting operations. This presentation concentrates on the modeling of the stress distribution when a blade acts on certain bio-material and the application of the cutting principles to biomaterial cutting. An analytical expression for the 3D cutting stress tensor has been obtained and simulated. The influence of the blade edge shape and the slicing angle are discussed. An optimal slicing angle can be formulated to maximize the feed rate while minimizing the cutting forces. Moreover, the method offers a means to predict cutting forces between the blade and the biomaterials, and a basis for design of robust force control algorithms for automating the cutting of biomaterials.

**Time:**Thursday, March 12, 2009 3:00--

**Location:**130 Solon Campus Center

**Behind the Scenes of Mobius Transformations Revealed**

Jonathan Rogness

Assistant Professor

University of Minnesota Twin Cities

**Abstract:**

"Mobius Transformations Revealed" is a short film which illustrates Mobius Transformations and shows how moving to a higher dimension makes them easier to understand. The movie was intended both to explain a certain theorem and to show the beauty of mathematics to a general public audience. After winning an award from the National Science Foundation and Science Magazine, it has been viewed online over 1.5 million times on YouTube alone, and been featured in stories by Minnesota Public Radio and the Associated Press. In this talk we'll explore some of the mathematics behind the movie, give a more detailed version of the theorem it illustrates, and talk about some open problems related to this description of Mobius Transformations.

**Time:**Thursday, March 26, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**MATHEMATICAL CONTEST IN MODELING 2009**

Energy and the Cell Phone: Brian Christner, Matt Holst, Aaron Potvien

Designing a Traffic Circle: Jeff Anderson, Carl Green, Jerod Lass

Undergraduate Students

Department of Mathematics & Statistics

UMD

**Abstract:**

Each February, a nationwide 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. The teams will discuss the contest problems, their proposed solutions, and their overall experience with the competition.

**Time:**Thursday, Apr. 2, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**Math and the Markets**

Joseph Artim

Instructor

Department of Finance and Management Information Sciences

UMD

**Abstract:**

In this talk I will give an overview of the UMD LSBE Financial Markets Program, discuss the mathematics used in finance, and describe opportunities for math majors in the finance industry.

**Time:**Thursday, Apr. 9, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**PENALIZED REGRESSION WITH NETWORKED PREDICTORS AND ITS APPLICATION TO eQTL ANALYSIS**

Professor Wei Pan

School of Public Health

University of Minnesota

Minneapolis, Minnesota

**Abstract:**

We consider the problem of conducting penalized regression analysis with predictors whose relationships are described a priority by a network. A class of motivating examples is to model a quantitative or categorical phenotype using gene expression profiles while accounting for coordinated functioning of genes in the form of biological pathways or networks. We introduce our new method and compare with some existing ones. We will discuss an application of the new method to express quantitative trait loci (eQTL) analysis. This is joint work with Drs. Benhuai Xie and Xiaotong Shen.

**Time:**Friday, Apr. 10, 2009 3:00--

**Location:**130 Solon Campus Center

**INTERACTIVE DRAWINGS OF ALGEBRAIC SURFACES**

Professor Joel Roberts

School of Mathematics

University of Minnesota

Minneapolis, Minnesota

**Abstract:**

In algebraic geometry we study geometric objects that are defined by systems of polynomial equations. Algebraic plane curves and algebraic surfaces in 3 dimensional space are instances that can be visualized. In this talk we will exhibit pictures of some curves and surfaces, illustrating some basic properties and also some features that seem fairly surprising. These figures will include several types of surfaces in R^3. Many of the surface pictures are interactive, in that they are posted on a webpage where the viewer -- using nothing more than a Java-enabled browser -- can rotate the figure continuously by dragging it with the mouse.

**Time:**Thursday, Apr. 16, 2009 3:00--

**Location:**130 Solon Campus Center

**UNDERGRADUATE HONORS COLLOQUIUM**

Jeff Anderson: "What Makes a Player 'Most Valuable': A Statistical Analysis?"

Amy Schmidt: "Development of Mathematica Software Tools for Analysis of Differential Equations Systems"

Xi Xi: "Does the Capital Asset Pricing Model Work In The Current Financial Turmoil?"

Jerod Lass: "Analysis of Traffic Circles" (to be presented in a Computer Science Colloquium, Tuesday, April 28, 3:00 p.m. in SCC 21)

**Abstract:**

Four undergraduate students in the Department of Mathematics and Statistics are graduating with honors this year. Part of the honors requirement is to complete a research project under the supervision of a department faculty member. Three of the four recipients will present their research in this colloquium. (The fourth will present at a separate Compute Science colloquium.)

**Time:**Thursday, Apr. 23, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**Analysis of Traffic Circles**

Jerod Lass

UMD Math/CS Major

**Abstract:**

The 2009 Mathematical Contest in Modeling featured a problem with the goal of finding the best algorithm/configuration for a traffic circle, i.e. number of lanes, stopping/yielding method, etc. The UMD team consisting of Jeff Anderson, Carl Green, and Jerod Lass attempted this problem during the contest, which gave teams 96 hours to generate a solution.

Since then, further research and analysis has been conducted using the simulation model that began in the days of the contest. A more developed solution will be discussed along with the challenges and future of this problem.

**Time:**Tuesday, Apr. 28, 2009 3:00--4:00

**Location:**21 Solon Campus Center

**SCALING LAWS AND SIMILARITY CONCEPTS APPLIED TO THE ANALYSIS OF STABILITY OF SURFACE AND UNDERGROUND EXCAVATIONS IN ROCK MASSES**

Dr. Carlos Carranza-Torres

Associate Professor of Geotechnical Engineering

Department of Civil Engineering

UMD

**Abstract:**

The presentation focuses on a classical problem of geotechnical engineering that involves the determination stability conditions for tunnels and slopes using cohesive-frictional models ---such as closed-form solutions, limit equilibrium models and numerical models. The presentation discusses how the application of scaling laws in classical shear-strength models ---e.g., Mohr-Coulomb models for soils Hoek-Brown models for rocks--- allows to obtain, through similarity rules, generality in the applicability of results from the above mentioned methods. The similarity of the problems discussed here is that manifested by factors of safety for the case of tunnels and slopes and convergence curves for the case of tunnels. The core of the presentation involves classical geotechnical engineering concepts that will be explained in simple terms and simple examples to a non-geotechnical engineering audience.

**Time:**Thursday, Apr. 30, 2009 3:00--

**Location:**130 Solon Campus Center

**DEVELOPING A STORM WATER UTILITY**

Derek Carlson

Undergraduate Student

UMD

**Abstract:**

Derek Carlson is an undergraduate from UMD and a civil engineer for the City of Superior. He will talk about some of the everyday duties working as a civil engineer. The focus of the presentation will describe the development of the bill database for the storm water utility in the City of Superior. There will be a time at the end for questions.

**Time:**Thursday, May. 7, 2009 3:00--4:00

**Location:**130 Solon Campus Center

**Optimum Mixed Volume for Polynomial Sets**

Annette Katich

MS Candidate: Program in Applied and Computational Mathematicst

Advisor: Marshall Hampton

Department of Mathematics and Statistics

UMD

**Abstract:**

A fundamental problem of algebra is determining the number of solutions to a system of polynomials. We know that for polynomials in one variable with complex coefficients, there is a complete and simple answer; however, for multivariate polynomials the answers are less clear. Currently, there are two primary methods for exact counts of solutions of a particular multivariate polynomial system, resultants and Groebner bases. The trouble is that the current methods are computationally intensive and infeasible for many applications. The mixed volume of a multivariate polynomial system provides an upper bound on the number of isolated non-zero solutions and is less computationally intensive. We investigate ways to manipulate the planar, circular, restricted four-body problem, a form of the n-body problem, into alternate equivalent forms with more optimal (lower) mixed volumes.

**Time:**Tuesday, June 16, 2009 3:30PM -- 4:30PM

**Location:**130 Solon Campus Center

**Effects of Monetary Rewards on Reliability of Health Assessment Surveys**

Lucas Streng

MS Candidate: Program in Applied and Computational Mathematicst

Advisor: Ronald Regal

Department of Mathematics and Statistics

UMD

**Abstract:**

Completing a health assessment is a process that many people who have health insurance have likely gone through at least once in their lives. To encourage participation in the assessment, an employer will usually offer the employees an incentive for the completion of an assessment. This gives rise to the question of: Does increasing the incentive cause a change in the way participants are responding to these surveys? For instance, does increasing the incentive cause the chances of seeing a yes response on the survey to decrease? To investigate this question, a logistic model with random effects was used. The results of this analysis and other questions of interest will be discussed in this talk.

**Time:**Thursday, June 18, 2009 10:00AM -- 11:00AM

**Location:**130 Solon Campus Center

**Creating Custom Matrices for Organisms with an A+T Rich Genome**

Seth Slettedahl

MS Candidate: Program in Applied and Computational Mathematicst

Advisor: Marshall Hampton

Department of Mathematics and Statistics

UMD

**Abstract:**

One of the main causes of malaria is the protozoan parasite Plasmodium Falciparum. P. Falciparum has an interesting property in that its genome is composed mainly of A¡¯s and T¡¯s. This makes it very difficult to define the proteins in P. Falciparum. This is why custom scoring matrices are important. We hope to show that the process of creating custom scoring matrices is not only manageable but helpful in the field of Bioinformatics, especially with high throughput genome annotation.

**Time:**Wednesday, June 24, 2009 3:00PM -- 4:00PM

**Location:**130 Solon Campus Center

**Penalized Model-Based Clustering**

Wei Zhang

MS Candidate: Program in Applied and Computational Mathematicst

Advisor: Kang James

Department of Mathematics and Statistics

UMD

**Abstract:**

Clustering is the identification of groups of observations that are cohesive and separated from other groups. Most clustering done in practice is based largely on heuristic but intuitively reasonable procedures, such as hierarchical agglomerative clustering and K-means clustering. Clustering algorithms based on probability models offer a principled alternative to heuristic-based algorithms. In particular, this paper mainly reviews the standard model-based clustering and penalized model-based approach with an penalty. Performances of both methods are compared by applying them to real gene expression data. We also extend the current penalized method to the one with an penalty while . The results will be discussed in this talk.

**Time:**Friday, June 26, 2009 2:00PM -- 3:00PM

**Location:**130 Solon Campus Center

**Numerical Analysis of Cooling Effects of a Cylinder Head Water Jacket**

Qingzhao Wang

MS Candidate: Program in Applied and Computational Mathematicst

Advisor: Steven Trogdon

Department of Mathematics and Statistics

UMD

**Abstract:**

The Computational Fluid Dynamics (CFD) code Fluent is used to solve the Navier-Stokes equations for the cooling water flow in the cylinder head of a single cylinder of a diesel engine. This region that is occupied by the cooling water is referred to as the water jacket. The Navier-Stokes equations are particularized for the k-epsilon turbulence model in which the "Law of the Wall" is utilized with "standard" wall functions. The aim of this study is to examine the effects various boundary conditions have on the overall heat transfer characteristics of the water jacket with special attention paid to those regions of the flow where temperature effects could be problematic.

**Time:**Saturday, August 15, 2009 10:00AM -- 10:50AM

**Location:**130 Solon Campus Center

**Active Control of Impulsive Noise**

Ronghua Zhu

MS Candidate: Program in Applied and Computational Mathematicst

Advisor: Zhuangyi Liu, Xun Yu

Department of Mathematics and Statistics

UMD

**Abstract:**

This presentation is about the comparison of performance among four different adaptive algorithms for suppressing impulsive noise. They are the filtered-x least mean square (FXLMS) algorithm, least mean M-estimate (LMM) algorithm, filtered-x least mean M-estimate (FXLMM) algorithm, and a modified FXLMM algorithm. A designed noise, which is composed of impulsive noise and background noise, is used as the reference input signal in the simulation. Extensive simulations are executed to demonstrate the effectiveness of these four algorithms. From the convergence speed and stability point of view, FXLMM achieves the best performance among them. Besides, some stability and modeling error analysis are presented.

**Time:**Monday, August 17, 2009 2:00PM -- 3:00PM

**Location:**130 Solon Campus Center

**Spatially Correlated Mixture Models with Application in Genomic Hypotheses Testing**

Jie Ren

MS Candidate: Program in Applied and Computational Mathematicst

Advisor: Kang James

Department of Mathematics and Statistics

UMD

**Abstract:**

An important task in genomic studies is the detection of genes satisfying certain conditions, such as being regulatory targets of a transcription factor (TF). With high-throughput data, e.g. microarray data, this is usually formulated as a simultaneous hypotheses testing problem. Gaussian mixture model (GMM) can be used in such a problem. However, standard GMM assumes that all the genes have an identical and independent distribution a priori, which contradicts the biological fact that genes work coordinately as dictated by gene networks. Wei and Pan (2008) proposed a spatially correlated mixture model to integrate gene network information into statistical analysis of genomic data. We applied the standard model, Wei and Pan¡¯s model, and a similar model with modifications to a real ChIP-chip data set. By comparing the ROC curves, we found that both the two spatial models had better performance than did the standard model. The modified spatial model gave an even higher statistical power, as well as the ability to discovery more potential target genes.

**Time:**Friday, September 4, 2009 1:00PM -- 2:00PM

**Location:**130 Solon Campus Center