Distinguished lectures in applied mathematics and statistics
Overview
This series of conferences on topics in applied mathematics began in 2024. Each year, the CRM welcomes one or two distinguished speakers. Each edition is devoted to a different theme. The theme for 2025 was numerical analysis. The theme for 2026 will focus on the statistical science of space and time.
Jean-Fran莽ois Coeurjolly (Universit茅 Grenoble Alpes)
叠颈辞驳谤补辫丑测:听Jean-Fran莽ois Coeurjolly has been a professor at the Jean Kuntzmann Laboratory (LJK) of University Grenoble Alpes, France, since 2020, after spending several years in the Department of Mathematics at UQAM (from 2016 to 2020). His research interests focus on the simulation and inference of stochastic processes, in particular irregular processes (fractional and multifractional), random fields, and spatial and spatio-temporal point processes. He is also very interested in multidisciplinary collaborations and has applied his work to climate science, cognitive science, environmental studies, economics, computer graphics, and experimental design, among others. He notably leads the Statistics Department of LJK as well as the CNRS research network MAIAGES dedicated to image processing and stochastic geometry.
Tuesday, May 26, 2026, 2:00 p.m to 3:00 p.m. (UQAM, room PK-5115)
(Presentation in French, slides in English)
Scientific conference (STATQAM)
罢颈迟濒别:听Spatial median for a point process and its link with the median of a perturbed Poisson distribution
This work was carried out, among others, in collaboration with Christophe Biscio (Aalborg University, Denmark) and Jo毛lle Rousseau Tr茅panier (M.Sc. student at UQAM, Data Scientist)..)
础产蝉迟谤补肠迟:听Estimating the intensity of a stationary point process is usually the first (and simplest) problem considered when analyzing point pattern data. It allows one to estimate, for instance, the number of trees in a forest or the number of lightning strikes over a given territory per unit volume. In this talk, we begin by showing that this estimator is, by construction, highly non-robust in the presence of outliers, that is, spatial regions where either no points or excessive concentrations of points are observed. We then show that a robust version can be developed by defining an analogue of a spatial median for a point process. The remainder of the talk investigates the theoretical properties of this new estimator and focuses in particular on a simple but non-trivial question: the study of the median of a perturbed Poisson distribution.
Thursday, May 28, 2026, 2:00 p.m to 3:00 p.m. (CRM, room 6214)
(Presentation in French, slides in English)
Public conference
罢颈迟濒别:听The point at the interface between time and space.
础产蝉迟谤补肠迟:听Point processes or point patterns are datasets used to model interacting objects-so-called points. A point may correspond to the firing time of a neuron, the location of a tree in a forest, a particle in an ideal gas, or the time and location of a lightning strike over a given territory, among others. Although such examples have appeared in many classical works, their use has increased significantly since the 2000s with the rise of georeferenced data acquisition techniques. Through a variety of examples (including forestry, economics, cognitive science, experimental design, genomics, and climate science) studied in recent years, we will illustrate several methodological challenges and recent contributions needed to account for the complexity of these applications: strong inhomogeneity, strong dependence, and high dimensionality (in terms of data size and multivariate structure), etc.
Peter F. Craigmile (Hunter College)
Biography:聽Peter F. Craigmile is a Professor of Statistics in the Department of Mathematics and Statistics at Hunter College, City University of New York. His research interests include time series and longitudinal analysis, spatial statistics, and spatio-temporal modeling. He works on building scientifically relevant hierarchical statistical models, applied to areas such as climate science, public health, psychology, environmental health, and medicine. He has extensive experience in collaborative, team-based interdisciplinary research, working with other statisticians and practitioners on numerous research projects at the local, national, and international levels. Professor Craigmile is a fellow of the American Statistical Association, the Institute of Mathematical Statistics, and the Royal Statistical Society.
Tuesday, May 26, 2026, 3:30 p.m to 4:30 p.m. (UQAM, room PK-5115)
(Presentation in English)
Scientific conference
Title:聽Modeling Nonstationary Time Series using Locally Stationary Basis Processes
Peter F. Craigmile, Department of Mathematics and Statistics, Hunter College, City University of New York. (This is joint research with Shreyan Ganguly.)
Abstract:聽Methods of estimation and forecasting for stationary models are well known in classical time series analysis. However, stationarity is an idealization which, in practice, can at best hold as an approximation, but for many time series may be an unrealistic assumption. We define a class of locally stationary processes which can lead to more accurate uncertainty quantification rather than making an invalid assumption of stationarity. This class of processes assumes model parameters to be time-varying and parameterizes them in terms of a transformation of basis functions that ensures the processes are locally stationary. We develop methods and theory for parameter estimation in this class of models and propose a test that allows us to examine certain departures from stationarity. We assess our methods using simulation studies and apply these techniques to the analysis of an electroencephalogram time series. We conclude with a discussion of a spatio-temporal extension.
Thursday, May 28, 2026, 3:30 p.m to 4:30 p.m. (CRM, room 6214)
(Presentation in English)
Public conference
Title:聽Stories in Statistical Climatology
Peter F. Craigmile, Department of Mathematics and Statistics, Hunter College, City University of New York. (This is joint research with Peter Guttorp and Thordis Thorarinsdottir.)
Abstract:聽Climate impacts are real and continue to affect our world. Thus, the study of climate is of great interest to the scientific community, policy makers, and the general public. In statistical climatology, we develop and use statistical methodologies to investigate how climate processes interact and evolve over space and time.
Through a selection of statistical stories, we explore how statisticians contribute to understanding the behavior of the climate system, using measures such as temperature and precipitation. We also introduce salient features of climate processes and climate data that should be incorporated into statistical models. Finally, we discuss the role of careful dataset selection and highlight the role that climate model simulations can play in studying our world.