Mathematics Department Fall Colloquium:
Student Summer Research Projects
Friday, September 9 at 3:00 PM
Cheray 101
Utilizing the Octet Truss in the Design of Lateral Transfer Retroreflectors
presented by James Coolidge ’12 and Kevin O’Brien ’11
Abstract: The purpose of our research was to improve upon the design of the Lateral Transfer Retroreflector (LTR) from the GLAS program for the NASA ATLAS (Advanced Topographical Laser Altimeter System) satellite. Our goal was to accomplish this while incorporating the octet truss geometry into the design. The primary focus was on minimizing the effects of thermal changes on the alignment of the optics, while creating a practical and functional design. We did the research on Saint Michael's campus under a Vermont Space Grant Consortium Student Mentoring Grant.
Signature Recognition using Wavelets and Other Methods
presented by Katie Heath ’14
Abstract: In this presentation we will utilize the Haar Wavelet Function and Fourier Series to perform a Discrete Fourier transform, and a Haar Wavelet decomposition. We use these and other similar methods to discover algorithms for identifying forged signatures. In addition, we will look at a few more sophisticated methods used by professionals for signature verification and recognition. This research was conducted at Texas A&M University as part of their pre-REU program.
Possession is 9/10’s of the Law: Banking, Footie, and how to “Excel” at Multilevel Logistic Regression
presented by Eric Parziale ’13
Abstract: Over the summer I took an Internship in the Financial Accounting Division at Credit Suisse. My team's main role was to calculate the banks' required capital (needed to protect the bank against risk). Much of this work involved the use of Microsoft Excel, which proved a wonderful tool in my summer statistical analysis in which Prof. Phil Yates and I teamed up to compete in a soccer-statistics competition. As part of the competition, the soccer-statistics company, STAT DNA, distributed 300+ games worth of raw soccer match data (in over 500,000 rows and 10 columns), and asked the competitors to choose an aspect of soccer to analyze. The Excel-training really paid off to make the data usable. We used the statistics program "R" to construct a multilevel linear regression model, and we were able to identify which players and teams are most successful in terms of pass-completion. Throughout the summer, I learned to "Look at the big picture", to figure out how different fields and ideas link-together, and to (attempt) to understand why certain processes are more effective than others.
Rhythmic Clustering of Drum Breaks
presented by Jaqui Tuthill ’12
Abstract: Many rap songs have long segments of a looped one-measure drum sample as the underlying beat, layered with instruments and vocals. Since the library of drum breaks is large, this project clusters the breaks in order to reduce the time needed to find the source of the break. We propose a method of clustering these drum breaks based only on rhythmic structure, ignoring features like tempo and timbre. The clustering creates a taxonomy of common rhythmic events for each cluster. Each break can identify with a taxonomy and then need only be compared to the breaks in the cluster rather than the library. This is done using template matching and partitioning around medoids. This work was done as part of the 2011 Claremont Colleges REU funded by NSF Grant DMS 07-55540.
Ninth Annual Pi Mu Epsilon Lecture
Frank Morgan, Atwell Professor Mathematics
Williams College, Vice President American Mathematical Society
Soap Bubbles and Mathematics
Friday, September 23 at 3:15 PM
Cheray 101
Abstract: A round soap bubble provides the least-area way to enclose a given volume of air, as was proved mathematically by Schwarz in 1884. Similarly, the double bubble that forms when two soap bubbles come together provides the least-area way to enclose two given volumes of air, although we did not prove that until 2000. The show will include a guessing contest with demonstrations, explanations, and prizes.
Sponsored by Saint Michael’s College Mathematics Department
& the VT Alpha Chapter of Pi Mu Epsilon National Honor Society
with support from NSF S-STEM