FRIDAY, NOVEMBER 10
Isaac Leiterman – 5:15 pm (St Norbet College)
Room: GMS 1097
Title: Predicting Low-Probability River Floods using Extreme Value Theory
Abstract: he devastation that can follow in the wake of extreme flooding events necessitates methods for accurately predicting events of such magnitude. In this presentation, we will discuss the application of Extreme Value Theory (EVT) to historical flooding data for seven Wisconsin rivers. We assert that this methodology provides advantages over existing techniques applied by the USGS to these rivers. Due to EVT’s wide array of applications in insurance, finance, and hydrology, it could become an important tool in predicting these low-probability flooding events and we hope to show the viability of modeling using EVT.
Isaac enjoys hiking and collecting rocks.
Sydney Vitalbo – 5:35 pm (St Norbert College)
Room: GMS 1097
Title: Exploration and Visualization of Research Concepts in the GAIA Database Union Presentation and Beta Invariant
Abstract: s academia takes an increasingly interdisciplinary approach towards research, the need to understand the relationships among authors and publications becomes apparent. The Görg lab at CU Anschutz previously created an interactive multivariate network approach (KaleidoNETS) that visualizes co-authorship connections along with author metadata, such as demographic and personnel data. This project expands that approach by integrating publication metadata from OpenAlex, specifically authors’ focus area, to support exploration and interpretation of author data and relationships in the context of the previously collected metadata.
While in Denver doing this project Sydney went skydiving!
Emily Landolt- 5:55 pm (St Norbert College)
Room: GMS 1097
Title: Malthusian growth models of EZ54 laboratory collected data show community-dependent growth rates
Prochlorococcus (Pro) is the smallest, most abundant, photosynthetic organism on the planet. Over time it has evolved to maximize its surface area:volume ratio to allow for the most efficient nutrient uptake. This resulted in Pro cells being very small with a minimal, streamlined genome. As a consequence, Pro lost important catalase and peroxidase genes that detoxify the environment of toxic reactive oxygen species (ROS) like hydrogen peroxide (HOOH) that can damage cells. Pro resides in the open ocean and must rely on other community organisms, referred to as ‘helper’ organisms, that have retained catalase and peroxidase genes to detoxify the environment from HOOH to the benefit of Pro. Among these helper organisms are the heterotrophic bacteria known as EZ54 (Thalassospira sp.) and EZ55 (Alteromonas sp.). Pro, EZ54, and EZ55 were isolated from a community identified in the ocean, and the interactions between Pro and EZ55 have been explored in depth. While EZ55 has been characterized as an excellent helper, little is known about EZ54. To explore the community interactions between EZ54, EZ55, and Pro, a tripartite experiment was set up alongside monoculture and dipartite controls. The abundance of each organism was sampled over an 18-day period and plotted based on which community it was grown in. Malthusian growth models were run in Python on lab-collected growth curve data to quantify growth rates. My results indicate that EZ54 growth rates are community-dependent and that EZ54 grows best when cultured with both EZ55 and Pro. In contrast, the growth rates of Pro and EZ55 did not vary based on their community structure. Future work will be done to assess how well EZ54 is able to detoxify the environment of ROS, like HOOH, and to determine how effective EZ54 is as a ‘helper’ organism for Prochlorococcus.
Emily has read over 30 books this year.
Liu Scott & Johnny Dahl – 6:25 pm (Lawrence University)
Room: GMS 1097
Title: Coding Genus 2 Fibrations
Abstract:
Tyler Blom – 6:25 pm (SNC)
Room: GMS 1096
Title: Sums of Pairs: A Pure and Pleasant Proof
Abstract: This talk will be a discussion of a proof in recreational number theory that I found over the summer based on a problem from Stan Wagon. The conjecture to be proven was that for any even integer, there are always three integers whose sums of pairs of them make the first three powers of that even integer.
Tyler likes painting and writing music.
Sarah Kulas – 6:45 pm (SNC)
Room: GMS 1097
Title: Present Bias in Group Work
In this talk, we discuss how procrastination is affected in a group setting. We begin by reviewing why a single individual procrastinates using findings by Kleinberg and Oren. The general idea is that a person inflates the cost of having to do a task in the present because doing work right now feels much harder than putting off that same work until tomorrow. Yet, when we get to tomorrow, this idea still applies and we end up putting off the task until the last minute when we are forced to do it for a much larger cost. We investigated this procrastination problem by using a fan graph with a source vertex, sink vertex, and weighted edge costs. To find the optimal path for doing a task requires finding the cheapest path from the source to the sink. In this talk, our goal is to find the most expensive path that a group could take.
Sarah enjoys running and is a member of the Cross Country and Track team here at SNC.
Emma Conran – 6:45 pm (Carthage College)
Room: GMS 1096
Title: Patterns in Political Polarization
The present political climate of the United States suggests that political partisanship has reached an all-time high. How can this be quantified? We investigate polarization in the U.S. Senate with methods previously used to study the U.S. House of Representatives. By measuring agreements both within and between members of political parties, we create a network structure defined by voting records. These networks display the range of senator behaviors, allowing us to track changes in partisanship on both an individual and a party level. We observe shifts in behavior over decades of voting history, highlighting trends and outlier events. Some of our findings may contradict what you hear on the news! This work was completed as part of the Carthage College Summer Undergraduate Research Experience.
Emma can touch their feet to the back of their head.
Dylan Schuster- 7:05 pm (SNC)
Room: GMS 1097
Title: Classifying Character Degree Graphs with Seven Verticesnt Bias in Group Work
In this talk, we discuss graphs with seven vertices in an effort to classify which of them appear as the prime character degree graphs of finite solvable groups. This classification is complete for the disconnected graphs. Of the 853 non-isomorphic connected graphs, we were able to demonstrate that twenty-two occur as prime character degree graphs. Forty-four graphs remain unclassified.
A non-mathematical fact (or set of facts) about Dylan is that he plays football at St. Norbert College, is slightly ambidextrous, and loves a good smoothie.
SATURDAY, NOVEMBER 11
Benjamin McFarland – 9:00 am (Ripon College)
Room: GMS 1096
Title: Predicting Results of ICA on Overcomplete Data
Abstract: The Independent Component Analysis (ICA) algorithm takes a recording with multiple channels and decomposes the recording into components that are as independent as possible. For example, ICA could be used on a two microphone recording of two people talking over each other. ICA returns an unmixing matrix which when multiplied to the recordings gives us two components estimating the two individual voices. Because ICA returns a matrix, it always returns the same number of components as channels. If there are more voices than components then there aren’t enough components for each voice. So ICA cannot fully separate the voices. We call data with an equal number of channels and sources complete, and data with more channels than sources overcomplete. ICA is currently used in many applications where there are likely more channels than sources, such as EEG and fMRI. Likewise, if there are fewer channels than sources, such as ECoG, ICA is not readily used.
Our aim is to predict how ICA decomposes overcomplete data. We have found that ICA groups sources together based on relative amplitude over the channels. We hypothesize that this grouping of sources can be modeled by the Least Square Mean (LSM) solution for the unmixing matrix, using a binning matrix to identify the groupings. Our findings show a pattern in the way ICA groups sources into components. This may allow researchers to better interpret components from overcomplete data. “
Benjamin enjoys crocheting and making wooden spoons.
George Burchfield – 9:00 am (Carthage College)
Title: Who’s on Top and Why? Exploring Digraphs to Represent Outcomes of Games or Events
Abstract:
Surveys show that 70% of Americans follow at least one sport. Many people look at sports rankings (everything from figure skaters to race car drivers to running backs) out of curiosity, to select fantasy teams, and for betting. What is the math behind these rankings? Check out my exploration of digraphs to represent outcomes of games or events. Digraphs can further be represented using matrices to make a quantified team ranking system. The system can be applied to professional sports leagues like the NFL as a more concrete way of ranking teams.
Georege likes to make humorous flyers for fictional events and hang them up around school or at restaurants.
Alyssa Ebeling- 9:20 am (Wisconsin Lutheran College)
Room: 1096
Title: Exploring Snort as a Tool for Analyzing Network Quality
Abstract: The Department of Homeland Security is invested in ensuring the quality and security of industry software. To this end, Montana State University has developed a code quality analysis tool, PIQUE, that uses static analysis tools to score overall code quality on a scale from 0 to 1. This REU project focuses on adapting a new model of PIQUE specific to microservice ecosystems. We explore the ability of a popular intrusion prevention/detection system, Snort, to measure and quantify network quality.
Alysa runs cross country at WLC, and their brother just got me hooked on the Stormlight Archive series.
Andrew Radicker – 9:20 am (Bradley University)
GMS 1097
Title: Graph Coloring, The Hadwiger-Nelson Problem, and Mycielski’s Construction
Abstract: The Hadwiger-Nelson Problem asks how many colors we would need to color each point in the coordinate plane so that no two points that are unit-distance apart share the same color. We will discuss how advancements have been made to this problem over the years, including how upper and lower bounds have been established through the use of graph theory. We will also explore corollary ideas like Mycielski’s construction and cones over graphs, which are related to the Hadwiger-Nelson Problem, but have not seen much research.
Andrew enjoys puzzles, yo-yo-ing, chess, and escape rooms with his friends
Andrew Valentini – 9:40 am (Carthage College)
GMS 1097
Title: Analyzing Causes of Gravitational Wave False Alarms
Abstract: The prompt discovery of gravitational wave events is performed by a variety of matched-filtering search pipelines. Due to the non-Gaussian nature of the Advanced LIGO detector, these pipelines often react to transient noise sources, commonly referred to as glitches, and mistake them for gravitational wave events. Analyzing the properties of these glitches and a pipeline’s reaction to them is thus crucial to improve pipeline efficiency and event validation accuracy. In this study, we perform such an analysis on one of these pipelines, GstLAL, with the intent of providing a battery of tests and models by which a given gravitational wave candidate may be evaluated against to provide a higher degree of confidence in the identification of the event origin in addition to offering a valuable perspective on the responsiveness of this pipeline to instrumental noise.
Andrew plays the guitar.
Peyton Snidanko – 9:40 am (Bradley University)
GMS 1096
Title: Fast Imaging Functional for Periodic Structures with Limited Data
This project focuses on the inverse problem of reconstructing the shape of periodic structures from scattered wave data. This inverse problem is motivated by applications of nondestructive evaluation of photonic crystals. A new imaging functional is proposed for reconstructing the shape of the periodic structure using only limited data from above or below. This imaging functional is fast, simple to implement, and robust against noise in the data.
Peyton enjoys reading, dancing, and board games.
Anna Bauer- 10:00 am (St Norbert College)
GMS 1097
Title: AI Powered Molecular Analysis
Abstract: In this integration between mathematics and biochemistry, protein structure is investigated. To predict a previously unknown protein structure we leveraged an artificial intelligence powered methodology called AlphaFold, a deep learning technique. This computationally intensive program grants probabilistic models in synthesizing the most likely protein structure based on specific chemical properties of the individual components of the protein. This method allows us to predict this structure based on amino acid side chain interactions locally and globally.
Anna’s favorite season is fall, especially the chai tea and vibrant colors.
Hasif Ahmed- 10:00 am (Lawrence University)
GMS 1096
Title: Accelerating Diffusion Models using Stochastic Solvers in Particle Physics
Abstract: Our research investigates the optimization of sampling techniques applied to particle physics datasets, focusing on numerical methods and comprehensive data analysis. The unique challenges encompassed by particle physics data, such as continuous coordinates, stochastic dimensionality, and permutation invariance, distinguish it from standard datasets. The prevalent deep generative models, mainly designed for image data, fall short for these datasets. Our approach centers on a novel neural network simulation called Fast Point Cloud Diffusion (FPCD). The core aim is to boost the efficiency and speed of diffusion models, assessed via minimized Wasserstein Distances between authentic and generated data distributions and reduced sampling time. Our proposed Stochastic Solver 0165EM250 samples 8 times faster than blackbox Scipy Solver Benchmark Runge-Kutta Solver 45 without sacrificing average accuracy across the particle dataset
Hasif is actually one year older (biologically) than what shows on their birth certificate.
