The Mathematics and Computer Science Seminar topics range widely, but typically focus on original research, technical exposition, snapshots of working life, or teaching.
Seminar Schedule (Fall 2018 - Spring 2019)
Seminar attendees are invited to gather at 15 minutes prior to the talk to partake of light refreshments and to socialize.
|
9/24, 2018 |
Honest Talk about Graduate School in Math, CS, and Physics 4pm, Thompson 391 |
|
10/1, 2018 |
Geodesics on Platonic Solids 4pm, Thompson 391 Abstract: In joint work with David Aulicino and Pat Hooper, we study the problem of finding closed geodesics passing through exactly one vertex on the surfaces of Platonic solids. We show that there are no such trajectories on any of the solids except the dodecahedron, on which there are 31 equivalence classes of such trajectories. The talk will be elementary and accessible to undergraduates. |
|
10/22, 2018 |
Modeling Sales Opportunities for Microsoft 4pm, Thompson 391 Abstract: This talk traces the development of a data science project at Microsoft over the course of 2.5 years – from initial problem statements to fully productized system, from idea formulation to patent filings. We’ll touch on technical challenges large and small, competing business goals, and difficult tradeoffs. Most importantly, we’ll discuss how people from different teams with different jobs and different backgrounds contributed to this project, and how this diversity of ideas was necessary for project success. |
|
11/5, 2018 |
Applications of topology for information fusion 4pm, Thompson 391 Abstract: In the era of "big data" we are often overloaded with information from a variety of sources. Information fusion is important when different data sources provide information about the same phenomena. For example, news articles and social media feeds may both be providing information about current events. In order to discover a consistent world view, or a set of competing world views, we must understand how to aggregate, or "fuse", information from these different sources. In practice much of information fusion is done on an ad hoc basis, when given two or more specific data sources to fuse. For example, fusing two video feeds which have overlapping fields of view may involve coordinate transforms; merging GPS data with textual data may involve natural language processing to find locations in the text data and then projecting both sources onto a map visualization. But how does one do this in general? It turns out that the mathematics of sheaf theory, a domain within algebraic topology, provides a canonical and provably necessary language and methodology for general information fusion. In this talk I will motivate the introduction of sheaf theory through the lens of information fusion examples. This research was developed with funding from the Defense Advanced Research Projects Agency (DARPA). The views, opinions and/or findings expressed are those of the author and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government. Approved for Public Release, Distribution Unlimited. |
|
11/26, 2018 |
Opportunities in Scientific Computing: Outsourcing the Computation So We Can Do the Thinking... 4pm, Thompson 391 Abstract: In the age of Numpy, WolframAlpha, and R (not to mention calculators), manual computation has gone the way of the dinosaurs. My research is in the field of scientific computing. That is, I study how we can use computers to help us answer scientific problems that would be exceptionally tedious or impossible to solve by hand. In this talk, I’ll share a few open projects in scientific computing that I am working on. First: How can we simulate precisely how a protein folds from one conformation to another? We’ll make use of probability theory to reframe this question in a more general way, and use concepts from statistics and parallel computing to think about computationally efficient ways to answer this. Second: How can we simulate systems that are way too big to even fit on our computer? This will draw on concepts from partial differential equations, Fourier analysis, and more! |
|
12/3, 2018 |
Deep Neural Networks, and Their Application to Scalograms of Laboratory Mouse Vocalizations 4pm, Thompson 391 Abstract: This will be a talk adapted from the presentation I made in 2017 at the 2017 IEEE International Conference on Bioinformatics and Biomedicine. I will start out with an introduction to deep neural networks: what they are and how they were adapted from the previous “shallow” neural networks. In particular a convolutional neural network (“CNN”) behaves much like the mammalian visual cortex, and has had great success in recent years in tasks with a visual or spatial component (such as in Google’s AlphaGo project). I will then show how we adapted a CNN to the task of identifying mouse vocalizations. This is a laborious task that traditionally takes hundreds of person-hours and yields inconsistent results. I will show our success using CNNs, and detail possible next steps that can be made. |
|
2/19, 2019 |
A Higher Multivariate Chain Rule 4pm, Thompson 395 Abstract: How do you take the repeated derivative of composed functions? For single-variable functions, the first few derivatives are readily computable by hand, but each consecutive derivative requires additional uses of the chain and product rules and quickly becomes overwhelming. The situation is even worse for the multivariate case. There has to be a better way than computing higher derivatives by hand. The solution for single-variable functions is to define a class of polynomials—Bell polynomials—which do all of the algebra and combinatorics for you, and use them to define the so-called Faá di Bruno Formula for taking repeated derivatives of composed, single-variable functions. However, there has been, to date, no generalization of Bell polynomials for use in the multivariate Faá di Bruno formula. In this talk, we introduce generalizations to Bell polynomials and Faá di Bruno formula to apply them to finding repeated derivatives of multivariate functions. |
|
2/25, 2019 |
Achievement Unlocked: An Inside Look at the Games Industry 4pm, Thompson 395 Abstract: Many a youngster grows up wondering what it might be like to make video games for a living. Turns out it’s not all fun and games, but it’s not boring either. Come learn about one developer’s journey to a career in gaming and some of the interesting challenges that face video game makers. |
|
3/11, 2019 |
Survey Data Science 4pm, Thompson 395 Abstract: For survey stats practitioners, it sometimes feels like the sky is falling. Response rates are declining. Data collection costs are increasing. Federal budgets are shrinking. But if the Pacific Northwest has taught me anything it is that the sky isn’t falling; no, the sky is raining data! With the advent of big data and advancements in technology, a wealth of additional data sources, such as satellite imagery (a.k.a. "sky data"), are available to supplement survey data. My talk will address how to combine these data sources using predictive models to produce potentially more efficient estimators. Drawing on my collaborations with the U.S. Bureau of Labor Statistics and the U.S. Forest Service, we’ll see how big data can help us estimate the number of bartenders in the US and the average number of trees per acre in Daggett County, Utah. |
|
4/8, 2019 |
Windows Internals: Multi-Core, Virtualization, and Persistent Memory 4pm, Thompson 395 Abstract: Come and hear how Windows has adapted to take advantage of the latest technologies in computing. We will discuss some of the driving factors that push Windows to continuously innovate. We will explain some of the techniques used to help deliver great battery life. We’ll also talk about how Windows “virtualizes” your motherboard hardware so you can run multiple operating systems at the same time (and why you would want to do such a thing). Lastly we’ll touch on some new storage technologies on the near-horizon that will change the fundamentals of how we write software. Bio Sketch: Matt earned his Computer Science degree from Texas A&M and has spent much of his career working on the Windows operating system. Matt has extensive experience at the lowest layers of the Windows kernel, focusing on hardware/software interfaces. He has been involved in bringing Windows to multiple new silicon architectures, from small IoT devices to large servers, and everything in between. Matt currently manages engineering teams that cover many technologies; everything from security, energy efficiency, system firmware, storage, and file systems. |
|
4/10, 2019 |
A Life in a (Work)-Day 4pm, Thompson 391 Abstract: Matthew Bogert (BS'17) and Mark Gilbert (BS'17) will be speaking about their roles at Workday, a company dedicated in delivering quality cloud-based solutions for businesses ranging from small to Fortune 500. Topics include:
|
|
4/16, 2019 |
Schröder numbers and equivariant operads 4pm, Thompson 395 Abstract: The Schröder numbers are a classical integer sequence counting a particular class of lattice paths. In the 1990's, they re-emerged in combinatorics through their role in counting a particular class of pattern-avoiding permutations. In joint work with Angélica Osorno and a team of Reed undergraduates, they appeared again in the structure of an equivariant operad motivated by stable homotopy theory. In this talk, I will explore the history and significance of the Schröder numbers, and show how they solved a problem in equivariant infinite loop space theory. |
|
4/22, 2019 |
Hecke L-functions: An Introduction & Two Research Projects 4pm, Thompson 395 Abstract: One of the most famous open questions in all of mathematics is the Riemann hypothesis, which is a conjecture regarding the locations of nontrivial zeros of the Riemann Zeta function. The Riemann Zeta function is a specific example of an L-function, which are functions of great interest in number theory. L-functions are important objects of study in number theory for precisely the same reason that the Riemann Zeta function is important: they provide information about the distribution of prime numbers. Studying the distribution of zeros of L-functions can tell us about the distribution of prime numbers in not only the ring of integers but also in other rings or fields. I will provide an introduction to the study of L-functions and discuss two research projects involving Hecke L-functions which I have worked on during my past two REUS (Research Experiences for Undergraduates). One project involved the study of regions of the complex plane in which a certain family of L-functions is never zero, and the other involved using L-functions and a (provably incorrect, yet still useful) conjecture called the L-functions Ratios Conjecture to analyze the distribution of prime elements in certain number fields. These projects touch on important areas of both analytic and algebraic number theory, and provide examples of what collaborative student research in number theory can look like. |
|
4/29, 2019 |
Mathematical Models for Hosts and Parasitoids in the Insect World 4pm, Thompson 395 Abstract: We provide a framework for understanding difference equation models for host– parasitoid species interactions in the insect world. We give context for the biological motivation of these models, which were selected to compare different combinations of standard functional forms for density-dependent growth of the host species and the effects of parasitism. These models combine simple and well-understood individual components, but these particular combinations yield some unexpected dynamics and rich mathematical behavior, including chaos. |