Waterloo Student Conference in Statistics, Actuarial Science and Finance
When Sacha Tihanyi was halfway through an undergraduate economics degree, he recognized a budding passion for mathematics. “I had many broad interests in school, but something flipped for me during that program,” he remembered. “I chose Waterloo because it had a reputation as the best mathematics school in Canada.”
As a master’s student, Chris Salahub completed a project focused on reducing the risk for cyclists on the busy streets of Zurich. His passion for leveraging data to solve real-world problems is what drew him back to the Faculty of Mathematics, where he previously earned a bachelor’s degree, to pursue a PhD in statistics.
“I like to think I’m a curious person,” said Jennifer Haid (BMath’04), a native of the Waterloo Region. When she learned about Waterloo’s Math Day from her high school math teacher, she decided to attend in hopes of learning about a career path that would leverage her aptitude for mathematics in the business world. “I remember watching a professor deliver a presentation about actuarial science and thinking two things: It was challenging, and I could do it,” she shared.
Katia Naccarato has never shied away from exploring an unfamiliar path, hitting a dead end, and trying a different one. Before she enrolled in the Master of Actuarial Science (MActSc) program at the Faculty of Mathematics, she was laser-focused on pursuing a career in medicine. Her current trajectory looks nothing like she expected, but she’s confident she’s heading in the right direction.
Samantha Wallis’s enthusiasm for statistics is matched only by her longtime passion for visual arts. As she considers her path forward after graduating with a degree in mathematics, Wallis thinks about how to meld her two interests into a single career. While she hasn’t landed on a definitive answer, she has a strong hunch where she will go next.
Network Response Regression for Modeling Population of Networks with Covariates
Multiple-network data are fast emerging in recent years, where a separate network over a common set of nodes is measured for each individual subject, along with rich subject covariates information. Existing network analysis methods have primarily focused on modeling a single network, and are not directly applicable to multiple networks with subject covariates.
In this talk, we present a new network response regression model, where the observed networks are treated as matrix-valued responses, and the individual covariates as predictors. The new model characterizes the population-level connectivity pattern through a low-rank intercept matrix, and the parsimonious effects of subject covariates on the network through a sparse slope tensor. We formulate the parameter estimation as a non-convex optimization problem, and develop an efficient alternating gradient descent algorithm. We establish the non-asymptotic error bound for the actual estimator from our optimization algorithm. Built upon this error bound, we derive the strong consistency for network community recovery, as well as the edge selection consistency. We demonstrate the efficacy of our method through intensive simulations and two brain connectivity studies.
Join Zoom Meeting
Meeting ID: 844 283 6948
Passcode: 318995
Statistics and Actuarial Science PhD candidate Yilin Chen is one of two students to claim the 2020 Huawei Prize for Best Research paper by a Mathematics Graduate Student. The $4,000 prize affirms the value of Chen’s efforts to establish a framework for analyzing nonprobability survey samples in her winning paper: Doubly Robust Interference with Nonprobability Survey Samples.
Please note: This seminar has been cancelled.
The Extended Reproducibility Phenotype - Re-framing and Generalizing Computational Reproducibility
Computational reproducibility has become a crucial part of how data analytic results are understood and assessed both in and outside of academia. Less work, however, has explored whether these strict computational reproducibility criteria are necessary or sufficient to actually meet our needs as consumers of analysis results. I will show that in principle they are neither. I will present two inter-related veins of work. First, I will provide a conceptual reframing of the concepts of strict reproducibility, and the actions analysts take to ensure it, in terms of our ability to actually trust the results and the claims about the underlying data-generating systems they embody. Second, I will present a generalized conception of reproducibily by introducing the concepts of Currency, Comparability and Completeness and their oft-overlooked importance to assessing data analysis results.