UOttawa Math+Stat Colloquium

Welcome to the University of Ottawa Colloquium in Mathematics and Statistics (UOCMS). This series brings together researchers from diverse institutions to present lectures on current developments in mathematics and statistics.  A colloquium talk is one that is geared towards a general mathematical audience.

We proudly acknowledge the support of the Department of Mathematics and Statistics, the Tutte Institute for Mathematical Computing (TIMC) and the Centre de recherches mathématiques (CRM).

Join us for cookies and coffee in the lounge (STEM 540) starting at 2:30; our colloquiua take place in STEM 664 from 3 to 4 pm.  

• Next talk:  Bei Jiang (University of Alberta) October 30, 2026

This talk will be part of the SCOPE seminar and is supported by CANSSI

• Bio: Dr. Bei Jiang is a full Professor in the Department of Mathematical and Statistical Sciences at the University of Alberta, a Canada CIFAR AI Chair, and a Fellow of the Alberta Machine Intelligence Institute (Amii). Her research focuses on the statistical foundations of trustworthy AI, including data privacy, algorithmic fairness, uncertainty quantification, federated learning, and statistical learning for complex and heterogeneous data. She has published extensively in leading statistics and machine learning venues, including the Journal of the American Statistical Association, Annals of Statistics, Journal of Machine Learning Research, NeurIPS, ICML, and ICLR. Dr. Jiang currently serves as Co-Editor of Statistics Surveys, Associate Editor for the Journal of the American Statistical Association, and Area Chair for NeurIPS 2026. She is a recipient of the 2025 COPSS Emerging Leader Award and the University of Alberta Faculty of Science Research Award.
• Title: Achieving AI Fairness Through Synthetic Data
• Abstract: Algorithmic decision systems are increasingly used in socially sensitive domains, raising concerns about bias inherited from historical data. Fair synthetic data generation provides a pre-processing strategy for bias mitigation, but existing methods often rely on black-box generative models whose tuning parameters have limited theoretical interpretation. We propose fDA, a Data-Augmented predictive framework for generating Fair synthetic labels. The framework combines a fairness-enforcing model, which specifies a fair reference conditional law satisfying the desired fairness constraint, with a faithfulness-preserving model, which generates auxiliary variables from observed labels to retain controlled originallabel information. Synthetic labels are sampled from the predictive distribution induced by jointly modeling these components, coupled with a tuning mechanism. For continuous labels, the Gaussian working specification yields explicit calibration of the effective noise level to a target population upper bound on unfairness. For both continuous and ordinal labels, the predictive distribution is fully specified for each tuning value, enabling empirical calibration of the achieved fairness–faithfulness trade-off. Theoretically, when the auxiliary variable becomes fully informative, synthetic labels converge to the original labels in probability and distribution; when it becomes non-informative, the mechanism reduces to the fair reference law. Experiments on simulated and real datasets show interpretable trade-offs and improved faithfulness over GAN-based baselines.
• Location: STEM 364

Some upcoming speakers:

• Alain Valette (University of Neuchatel) November 9, 2026

• Prof Valette earned his PhD under the supervision of Lucien Waelbroeck and Alain Connes. He has a long and distinguished history of important mathematical contributions, in a large range of areas including  operator algebras, graph theory and geometric group theory. He has written several books with different coauthors, notably on Kazhdan's Property (T), expanding graphs and the Baum-Connes conjecture.  These reference books have been crucial in the modern development of these topics.
• Title: The Wasserstein distance on metric graphs and trees
• Abstract:  Let (X,d) be a metric space. The Wasserstein distance, or earthmover distance, is a metric on the space of probability measures on X, that originates from optimal transportation: if \mu, \nu are probability measures on X, the Wasserstein distance between \mu and \nu intuitively represents the minimal amount of work necessary to transform \mu into \nu. Since the definition of the Wasserstein distance involves an infimum, it is not expected in general that there is a closed formula for this distance. Such a closed formula however exists for metric trees (i.e. combinatorial trees where the length of an edge can be any positive real number), and this closed formula has an interesting history that makes it suitable for a colloquium talk: it appeared first in computer science papers (Charikar 2002), then surfaced again in bio-mathematics (Evans-Matsen 2012), before catching the interest of pure mathematicians. In joint work with G. Skandalis, we consider the Wasserstein distance on finite metric graphs. Linearizing the problem by embedding the Wasserstein space into the space of functions with total mass 0 on vertices, equipped with transportation cost norm, we relate the latter space to the space of cycles, a classical object from algebraic graph theory. This re-proves and explains the closed formula for trees : they have no cycle !

• Chris Kapulkin (Western University) November 20, 2026

  • Prof Kapulkin earned his PhD from the University of Pittsburgh in 2014.  Dr.~Kapulkin works in homotopy theory, applying its techniques to numerous other areas of mathematics, including formal logic, theoretical computer science, combinatorics, category theory, and interactive theorem proving.  He has made a number of significant contributions to homotopy type theory, including proving the homotopy canonicity conjecture first proposed by Voevodsky in 2010.  He won the Florence Bucke Prize in Science in 2024 and the Coxeter--James Prize of the Canadian Mathematical Society in 2025, and is an Associate Editor of the Canadian Journal of Mathematics and the Canadian Mathematical Bulletin.

• Archer Yang (McGill) November 27, 2026

  • • Prof. Yang is an Associate Professor of Statistics at McGill University and an Associate Academic Member of Mila, the Quebec AI Institute. He is also affiliated with McGill’s programs in Computer Science and Quantitative Life Sciences. Professor Yang received his PhD from the University of Minnesota, where he was advised by Professor Hui Zou. His research lies at the intersection of statistical machine learning and uncertainty quantification, with interests spanning high-dimensional inference, reliable AI, scalable computational methods, and biomedical data science, including applications to drug discovery.
    • Tite: TBA
    • Abstract: TBA

• Zelalem Negeri (Waterloo) Dec 4, 2026

  • • Bio: Zelalem Negeri is an assistant professor in the University of Waterloo’s Department of Statistics and Actuarial Science. Before that, he completed a two-year postdoctoral fellowship at McGill University after earning his Ph.D. in Statistics at McMaster University. His research interest lies at the intersection of statistical methodology, computational statistics, and biostatistics, with an emphasis on robust methods for meta-analysis of diagnostic test accuracy studies.
    • Tite: One-stage binormal random-effects modelling for threshold-based diagnostic test accuracy meta-analysis
    • Abstract: Meta-analysis of diagnostic test accuracy is a rapidly growing area of research. It often includes studies that report binary test results obtained by applying different numerical thresholds to an underlying continuous biomarker. Standard bivariate and hierarchical summary receiver operating characteristic models synthesize sensitivity and specificity but do not explicitly use the reported threshold values, which leads to clinically uninterpretable diagnostic accuracy estimates. In contrast, existing multiple-threshold methods model cutoff-dependent accuracy curves from reported threshold-specific summaries. They are not formulated as a one-stage marker-level likelihood for a single selected cutoff per study, based on the observed two-by-two counts. In this talk, we present a novel one-stage binormal random-effects model for meta-analysis with study-specific thresholds and a single two-by-two table per study, which provides threshold-specific sensitivity and specificity as well as conditional and marginal summary receiver operating characteristic curves. We will demonstrate and validate the proposed method via extensive simulation studies and real-life data examples.

• Emily Riehl (Johns Hopkins) February 2027

  • Prof Riehl earned her PhD from the University of Chicago in 2009.  She  is the Kelly Miller Professor of Mathematics at Johns Hopkins University, working on higher category theory, homotopy type theory, and computer formalization. She has held prestigious visiting positions all over the world. She has held numerous distinctions and awards, and most recently was an invited sectional speaker in topology at the 2026 ICM.  She is a community leader, having served as a founding board member of Spectra, and currently serving on the Executive of the AMS Council.

• Mathilde Gerbelli-Gauthier (University of Toronto) Winter term 2027

  • Prof Gerbelli-Gauthier earned her PhD from the University of Chicago in 2020 and is now an Assistant Professor at the University of Toronto, working in number theory and representation theory. She is an excellent public speaker and has already won numerous distinctions for both her research and teaching.  In addition to her contributions to asymptotic and statistical aspects of automorphic representations, and the cohomology of arithmetic groups, she is active in the community, organizing conferences and seminars, and serving on panels and conducting outreach activities.

• Jean-Luc Guermond (Texas A&M University) May 7th, 2027

  • TBA
  • Title: TBA
  • Abstract: TBA

Would you like to nominate a speaker?  Please contact Monica Nevins, Arian Novruzi or Dingke Tang.