The department’s research strength is centered on excellence in theory and methods. In addition, many faculty members in the department have very successful collaborative research programs.
Andrei Badescu, Associate Professor, Tenure stream. B.S. and MSc Economics – Academy of Economical Studies Bucharest, PhD in Actuarial Science at Western University. I am interested in studying the surplus process of an insurance company under various model assumptions that includes Markovian arrival processes, Phase-type claims and dependencies among them. Bridging Ruin Theory to the area of Stochastic Claim Reserving by making use of the theory of Marked Point Processes is a new topic that interests me. Microlevel Claim Reserving is another new and very active topic that I am working on. Modeling Aggregate Claim Processes via mixture distributions is another area of interest with various applications in Actuarial Science, Operational Risk etc. The connections between Stochastic Fluid Models and Ruin Problems are also of high interest to me.
Jerry Brunner, Associate Professor, Statistics, Tenure Stream. B.A De Paul University (Psychology), Ph.D., University of Chicago (Psychology), M.Sc., SUNY-Buffalo (Statistics), Ph.D., SUNY-Buffalo (Statistics). Statistical models incorporating measurement error, with emphasis on parameter identifiability and model mis-specification in structural equation models. More recently I have been working on estimating the replicability of published research, and the related problem of detecting patterns of bad statistical practice using statistical analysis of the statistics reported in published articles.
Radu Craiu, Full Professor (Statistics), Tenure-Stream. B.S. & M.S. in Mathematics, University of Bucharest, Romania; PhD in Statistics, University of Chicago. A constant in my research is the development of efficient statistical computing methods, especially Markov chain Monte Carlo algorithms. I am also interested in methods for dependent data, copula-based models, model selection and dimension-reduction. As a member of the Strategic Training for Advanced Genetic Epidemiology (STAGE) team I maintain a constant interest in developing statistical methods for genetics.
David Duvenaud, Assistant Professor, Computer Science and Statistics, Tenure Stream. B.Sc.Hons in Computer Science, University of Manitoba; M.Sc. in Computer Science, University of British Columbia; Ph.D. in Engineering, University of Cambridge. My research focuses on developing new model classes in which efficient inference is possible, with a focus on latent-variable models with meaningful structure. I have a focus on generative models, which let us generate data such as images and text to answer questions about the model.
Michael Evans, Full Professor, Statistics, Tenure Stream. BSc (Western), MSc, PhD (Toronto). Generally concerned with the development of a theory of statistical reasoning based on measuring statistical evidence appropriately. This leads to the checking of the subjective elements in a statistical analysis against the data, to see if reasonable choices have been made, and a theory of inference using these ingredients, based upon measuring the change in belief induced by the data.
Andrey Feuerverger, Full Professor, Tenure Stream. B.Sc. McGill, PhD UC Berkeley. Statistical theory, Mathematical statistics, Transform methods, Spectral analysis, novel applications
D A S Fraser, Professor, Emeritus. PhD Princeton. Bayes methodology can be based on genuine priors or default priors (Efron 2013): For the default (Laplace) case, many have thought that inserting a mathematical object into the conditional probability formula would give a real probability but at most it can give reproducibility which is confidence. Statistics needs to update its logical methods (Statistical Science 2016).
Alison Gibbs, Associate Professor, Teaching-Stream. B.Math. in Applied Mathematics, University of Waterloo; B.Ed., University of Western Ontario; M.Sc. and Ph.D. in Statistics, University of Toronto. My research interests include the effect of innovative teaching methods on student learning of statistics, particularly online and inverted classrooms. In addition, I regularly participate in collaborative research projects with University of Toronto colleagues from a variety of academic units.
Sebastian Jaimungal, Full Professor, Tenure Stream. B.A.Sc. (Toronto), MSc, PhD (UBC). My current research interests tie in probability theory with financial markets. I am interested generally in how agents can act optimally in a variety of settings, how multiple agents interact with one another and play stochastic games, and how data driven approaches to financial modeling can be married with traditional stochastic analysis methods.
Keith Knight, Full Professor, Tenure Stream. B.Sc. (UBC), M.S., Ph.D. (Washington). My research has focuses on estimation in models where standard assumptions do not hold, such as estimation of extremes and estimators that are solutions of linear or convex programs.
Dehan Kong, Assistant Professor, Statistics, Tenure Stream. B.S. in Mathematics, Nankai University; Ph.D. in Statistics, North Carolina State University. My research interests focus on big data related areas such as neuroimaging data analysis, high dimensional data analysis, functional data analysis and machine learning. I am interested in developing novel and rigorous statistical procedures to answer relevant and important scientific questions arising from the real data.
Sheldon Lin, Full Professor, Tenure Stream. BSc (Xiamen), MSc (Fuzhou), MMath (Waterloo), PhD (Alberta). Application of non-Gaussian mixture models in general insurance; Prediction of IBNR claims and stochastic claim reserving; efficient nested simulation for large variable annuity portfolios using spatio-temporal models.
Radford Neal, Professor, Tenure stream. BSc Calgary, MSc Calgary, PhD Toronto. I am currently working on improving the implementation of the R programming language, on fast exact summation, on reinforcement learning in partially observed environments, and on Markov chain Monte Carlo methods. I have also done work on other issues in statistical computation, on non-parametric Bayesian modeling, on error-correcting codes, on data compression, on philosophical issues in probabilistic inference, and on applications in bioinformatics and other fields.
Jeremy Quastel, Professor, Tenure Stream. PhD NYU/Courant Inst 1990. Stochastic PDE, Kardar-Parisi-Zhang equation and universality class
Nancy Reid, Full Professor (Statistics), Tenure-Stream. B.Math Statistics Waterloo; MSc Mathematics UBC; PhD Statistics Stanford. My research focusses mainly on the theory of inference, with emphasis on methods based on the likelihood function and approximations to likelihood developed for complex models and data. Current projects include composite likelihood and inference for change points.
Jeffrey Rosenthal, Full Professor (Statistics), Tenure-Stream. B.Sc. in Mathematics, Physics, and Computer Science from U of T; PhD in Mathematics from Harvard University. My research interests include probability theory, Markov chains, stochastic processes, and statistical computation especially Markov chain Monte Carlo computer algorithms.
Daniel Roy, Assistant Professor, Dept of Statistical Sciences, Computer and Mathematical Sciences (UTSC), Computer Science (status only), Tenure-Stream. S.B., M.Eng., in EECS; Ph.D. in Computer Science, MIT. I focus on theoretical questions and ones that usually have some role in the foundations of machine learning or statistics. Recent work has made progress on the foundations of statistical network analysis by introducing a new class of random graphs that are sufficiently sparse to model real world data. Other work revisits the foundations of statistical decision theory using tools from mathematical logic to draw tight connections between frequentist and Bayesian notions of optimality.
Jamie Stafford, Full Professor and Chair, Tenure-Stream. PhD in Statistics, University of Toronto. My research activity concerns the analysis of interval censored data, correct use of flexible modern regression tools in the analysis of complex survey data, as well as spatial data methodology and data analysis.
Lei Sun, Full Professor (Statistics and Biostatistics), Tenure-Stream. B.S. in Mathematics, Fudan University; PhD in Statistics, University of Chicago. Develop statistical methods and computational tools that address analytical challenges arising from genetic studies of complex human traits. Current research projects include robust joint location-scale scale and multivariate association tests, stratified FDR control and multiple hypothesis testing, bias correction and modeling in selective inference, and collaborative work on genetic studies of Cystic Fibrosis, Type 1 Diabetes and Psychiatric Disorders.
Nathan Taback, Assistant Professor, Teaching-Stream. BSc, MSc, PhD (Toronto). My research interests include: applications of statistical methods in medicine, inference for observational data, and statistical education and communication.
Stanislav Volgushev, Assistant Professor, Statistics, Tenure Stream. Diplom in Mathematics (Msc equivalent) Ruhr university Bochum, Germany; Phd in Mathematics, Ruhr University Bochum, Germany. A core topic of my research are non- and semi-parametric methods that deal with distributional effects in data. This includes methods such as quantile regression and copulas that focus on models and inference for conditional and joint distributions, respectively. Recently, I have also started working on computationally feasible inference procedures for large-scale data sets. One particular focus is on the construction of novel bootstrap procedures which continue to be computationally feasible for large data sets. I also work on theoretical properties of recently proposed approaches to parallel computation such as the divide and conquer approach.
Bethany White, Associate Professor, Teaching Stream, Teaching Stream. BScH (Acadia), MMATH, PhD (Waterloo). My research interests relate to the impact of structured technology-enabled activities and course formats on students learning and attitudes of statistics.
Fang Yao, Full Professor (Statistics), Tenure-Stream. B.S. in Statistics, University of Science of Technology of China; MS in Statistics, UCDavis; PhD in Statistics, UCDavis. My research focuses on methodological developments in modeling functional and high-dimensional data with complex structures. Current research interest mainly includes estimation and inference for large-scale functional regression models in high dimensions, learning and inference on representation of functional objects residing in infinite-dimensional spaces, Gaussian sequences approach to functional data, complex data models that interface imaging, manifold and graphical/network structures as random objects.
Vicki Zhang, Assistant Professor, Teaching-Stream. B.Eng (Electrical Engineering and computer science), Shanghai Jiaotong University; Masters in statistics, University of California, Santa Barbara; ABD in statistics, UCSB (completed PhD qualification exams). Fellow of Society of Actuaries (FSA). Associate of Canadian Institute of Actuaries (ACIA). Certified enterprise risk analyst (CERA). My research mainly falls into one of the two categories: (1) researches that examine the effectiveness of different pedagogical approaches in actuarial science and insurance finance; (2) researches that examine the various insurance products and the evolution of insurance regulation, and the economic and social consequences
Zhou Zhou, Associate Professor, Tenure stream. B.S. Peking University. Ph.D. University of Chicago. My major research interest, in general, involves statistical inference of data with complex structures. Such data structures include, but are not limited to non-stationary temporal dynamics and dependence structures which vary both abruptly and smoothly over time, stochastic processes which exhibit nonlinear temporal associations and dependencies, functional or longitudinal data where the object of interest lies in certain infinite dimensional Hilbert space and statistical associations which are constrained to certain complex geometric spaces such as cones or manifolds. The analysis of such data inevitably involves certain non- or semi-parametric statistical procedures since parametric specification of the complex data generating mechanism is difficult, if not impossible.