Assessing the risk of extreme events is important in fields such as environmental science, finance, and insurance. While univariate extreme value theory (EVT) provides a well-established framework for modeling and extrapolating rare events, the multivariate setting is more challenging under asymptotic independence between variables.

The Heffernan–Tawn (HT) conditional extremes model addresses this by characterizing a response variable’s distribution given that a covariate exceeds a high threshold. However, existing implementations estimate a single pair of location and scale normalizing parameters across the entire covariate space, which limits conditional extreme quantile estimation accuracy where the dependence structure varies across the conditioning space, particularly in data-sparse regions. To address this, the present work proposes a locally adaptive estimator for the HT model based on random-forest-derived adaptive neighbourhoods and weighted quantile regression. The proposed approach preserves the theoretical structure of the conditional extremes model while replacing restrictive global estimation with flexible, locally adaptive inference. 

A simulation study compares the proposed random-forest-weighted (RF) estimator against the global HT benchmark across five bivariate dependence structures spanning asymptotic independence and both strong and weak asymptotic dependence. The RF estimator achieves a consistent RMSE reduction over HT across all structures and target quantile levels ($\tau = 0.90, 0.95, 0.99$). The improvements are particularly pronounced under weak asymptotic dependence and at the most extreme quantile levels.

To join this seminar virtually, please request Zoom connection details from hr.ops@stat.ubc.ca. 

This talk is part of the Institute of Applied Mathematics (IMS) Seminar series. SCOPE connects researchers and trainees with leading work in statistical sciences, biostatistics, and data science from Canada and beyond. Organized with CANSSI Ontario’s partner universities, the series features accomplished researchers advancing theory, methodology, and innovative applications across interdisciplinary challenges.

Abstract:

Markov Chain Monte Carlo (MCMC) is a powerful algorithmic framework for sampling from complex probability distributions. Standard MCMC methods struggle with high-dimensional distributions containing well-separated modes, becoming trapped in local regions. Replica-exchange/Parallel tempering (PT) addresses this by using intermediate annealing distributions to bridge a tractable reference (e.g., Gaussian) and an intractable target distribution. However, classical PT is inflexible, fragile, hard to tune, and prone to performance collapse on challenging inference tasks.

This talk introduces non-reversible parallel tempering (NRPT), which provably dominates classical PT algorithms. We show that NRPT undergoes a sharp algorithmic phase transition with increased parallelism, becoming robust, easy to tune, and scaling efficiently to GPUs. I will then demonstrate how to further accelerate PT using neural transports such as normalising flows and diffusions. We demonstrate this framework across a variety of examples in Bayesian inference and inference-time control for diffusion models.

This talk is part of the Canadian Statistical Sciences Institute's (CANSSI) SCOPE Seminar series. SCOPE connects researchers and trainees with leading work in statistical sciences, biostatistics, and data science from Canada and beyond. Organized with CANSSI Ontario’s partner universities, the series features accomplished researchers advancing theory, methodology, and innovative applications across interdisciplinary challenges.

Abstract:

Unseen planets, brown dwarfs and stellar companions leave a faint imprint on the motion of their host star. The Hipparcos and Gaia space missions, together spanning 25 years, make this imprint measurable for over a hundred thousand nearby stars.

I will describe our reconstruction of the companions of 106,508 stars, which uses a catalogue we recently released (G23H). Each reconstruction is a trans-dimensional posterior over the number of companions and their orbits. Unlike a catalogue of detections, a posterior for every star, including the non-detections, can support demographic inference without detection thresholds. We are now developing a joint population model in which the population and the individual systems inform each other.

This project shows where Bayesian modelling stands today. Some things are now feasible at scale: exploring multimodal, ill-conditioned posteriors over mixed discrete and continuous parameters with variational non-reversible parallel tempering, implemented in our open-source library Pigeons, and iterating quickly on models written in domain-specific modelling languages. Other challenges remain open, from quantifying the effect of misspecification to managing the complexity of large workflows.

Joint work with William Thompson (NRC Herzberg), Dori Blakely (University of Victoria), Trevor Campbell (UBC) and the wider G23H team.

REGISTER HERE

In data science education, competency is built through hands-on manipulation of code and data. Traditional lecture slides often lack this kind of active learning, presenting static lines of code alongside their output instead. When live coding is needed, instructors often need to switch between the slides and a separate environment such as RStudio — one that students may need to install themselves just to reproduce the demonstration. This presentation demonstrates a workflow for creating slides that embed live, editable R code students can modify and execute directly within the slide itself, removing the friction of switching between lecture material and a separate coding environment. We use Quarto, the live-revealjs extension, and GitHub Pages to build a simple, reproducible example course website that instructors can clone and adapt for their own use.

To join this seminar virtually, please request Zoom connection details from hr.ops@stat.ubc.ca.

This talk is one of the Statistics Teaching and Learning in Statistics and Data Science Seminar Series. 

Statisticians who use Bayes factors often complain that, somewhat awkwardly, their hypothesis testing and estimation results can disagree: a Bayes factor may favour the null hypothesis while the 95% credible interval excludes the null value. In this talk, I show that this apparent contradiction disappears once testing and estimation are required to use the same "implied prior model odds." Upon careful consideration, we see that using a Bayes factor for decision-making implicitly commits one to a mixture model and, when one's null is a point null, implicitly specifies a spike-and-slab prior. Such priors are not without their challenges. The resulting model-averaged posterior has an atom at the null value, its CDF has a jump, and credible intervals at conventional levels such as 95% may not exist at all. Moreover, if the frequentist p-value is held fixed while the sample size grows, the range of probability levels at which a credible interval can be defined steadily shrinks. This is a lesser-known correlate of the Jeffreys–Lindley paradox, one that carries the paradox from testing into interval estimation. All of this leads to a simple question: is a spike-and-slab prior (or, equivalently, a Bayes factor against a point null) ever a good idea?  

To join this seminar virtually, please request Zoom connection details from hr.ops@stat.ubc.ca.

Reported cases provide a delayed and incomplete view of infection incidence and so make it difficult to recover the true timing and magnitude of an epidemic. To use such surveillance data to estimate latent infections, we develop a trend filtering deconvolution approach. We use this to retrospectively estimate daily COVID-19 infections across U.S. states prior to the Omicron period using seroprevalence data in an antibody prevalence model. We adapt it for the Omicron period by using viral shedding and wastewater concentration data that allow for infection estimation when case and seroprevalence surveillance data become less reliable. Then we study the statistical properties of the estimator by extending trend filtering theory to the case where the design arises from delay distributions as a Toeplitz convolution operator and develop a backfitting extension to handle multiple overlapping infection curves, such as those from different variants.

To join this seminar virtually, please request Zoom connection details from hr.ops@stat.ubc.ca. 

Since the widespread availability of generative artificial intelligence (GenAI), particularly Large Language Models (LLMs), fundamental questions have emerged about the future of coding in data science. Some predict that data scientists will no longer need traditional coding skills, while others question whether LLMs might replace data scientists entirely. However, these discussions have largely proceeded without empirical evidence of how practicing data scientists actually use these tools.

This study addresses this gap by surveying trained, practicing data scientists to understand if and how they integrate LLMs into their workflows, particularly for writing and editing code and performing other data science tasks. Building on our recent investigation of data science educators' perspectives on LLMs, this research examines real-world usage patterns among practitioners to bridge the gap between current practice and educational preparation.

Our findings will contribute to the data science community in two critical ways. First, by documenting how data scientists are actually working with LLMs four years after their initial release, we provide actionable insights that allow practitioners to learn and adopt effective strategies for integrating these tools into their work. Second, we inform data science education by evaluating whether current pedagogies adequately prepare students for this evolving landscape.

This research will help answer key pedagogical questions: Should coding education emphasize code reading, tracing, and editing over writing large amounts of de novo code? Should greater focus be placed on writing high-quality documentation and specifications, given their value as prompt context for LLMs? Should testing receive increased emphasis to enable verification of LLM-generated code? By grounding these questions in empirical evidence of practitioner behaviour, we aim to provide data-driven guidance for evolving data science curricula.

To join this seminar virtually, please request Zoom connection details from hr.ops@stat.ubc.ca.

This talk is one of the Teaching and Learning in Statistics and Data Science Seminar Series. 

Data science underpins modern AI and many advances in healthcare, yet human judgment permeates every stage of the data science life cycle. These judgment calls introduce hidden uncertainties that go well beyond sampling variability and drive many of the risks associated with AI.

We introduce veridical data science, grounded in three fundamental principles—Predictability, Computability, and Stability (PCS)—to make such uncertainties explicit and assessable and to aggregate reality-checked algorithms for better results. The PCS framework unifies and extends best practices in statistics and machine learning and is illustrated through healthcare applications, including identifying genetic drivers of heart disease, reducing cost of prostate cancer detection, improving uncertainty quantification beyond standard conformal prediction, and proposing, Green Shielding, a new user-centric framework for safeguarding users of AI. 

To join this seminar virtually, please request Zoom connection details from hr.ops@stat.ubc.ca.

Estimation frameworks for statistical inference are preferred to hypothesis testing when quantifying uncertainty and precise estimation are more valuable than binary decisions about statistical significance. Study design for estimation-based investigations often uses precision criteria to select sample sizes that control the length of interval estimates with respect to a sampling distribution. In this work, we define a distribution that characterizes the probability of obtaining a sufficiently narrow interval estimate as a function of the sample size. This distribution can be used to determine the smallest sample size needed to ensure an interval estimate is sufficiently narrow. We prove that this distribution is approximately normal in large-sample settings for many data generation processes. However, this approximate normality may not hold for studies with moderate sample sizes, particularly when incorporating prior information or obtaining asymmetric interval estimates. Thus, we also propose an efficient simulation-based approach to approximate the distribution for the sample size by estimating the sampling distribution of interval estimate lengths at only two sample sizes. Our methodology provides a unified framework for design with precision criteria in Bayesian and frequentist settings.

To join this seminar virtually, please request Zoom connection details from hr.ops@stat.ubc.ca.