Short Course Schedule

Monday, June 30, 2025

Instructors

 
Likun Zhang
Likun Zhang

University of Missouri

Jordan Richards
Jordan Richards

University of Edinburgh

Reetam Majumder
Reetam Majumder

University of Arkansas

Monday, June 30, 2025

Location: 

Session 1: Foundations of Extreme Value Theory (EVT; keep this short)

Introduction to univariate EVT

  • Generalized Extreme Value (GEV) and Generalized Pareto (GPD) distributions

Modeling non-stationarity via covariates

  • Generalized linear modeling of GEV/GPD parameters
Session 2: Introduction to Neural Networks with Keras

Overview of TensorFlow/Keras basics

  • Hands-on primer tailored to statisticians
Gaussian regression as a motivating bridge
Session 3: Semi-parametric quantile regression (SPQR) for density estimation
Modeling the full conditional distribution
Applications in probabilistic forecasting and tail risk assessment

Case Study: Bias correction of climate model output

  • Hands-on exercise using CMIP6 and nClimGrid data
Looking forward: density estimation for weather and climate extremes

Session 4: Modern ML/DL Approaches for Dependent Extremes

Transition to flexible GEV/GPD regression frameworks
Other use cases: dependence modelling
Other options for GPD regression: decision trees, GBEX, EVGAM, erf

Overview of current research frontiers in deep learning for spatial-temporal extremes

  • Neural estimators, physics-informed architectures, generative AI: VAEs and GAN
Discussion of challenges: uncertainty quantification, extrapolation, and interpretability

Ongoing work and insights from our forthcoming book chapter on ML for extremes

  • Opportunities for collaboration and continued learning


Conference Schedule

Monday, June 30, 2025

TimeTitleLocation
8-8:30 AMCheck-InLeadership Lounge
8:30-8:45 AM

Welcoming remarks

Dr. Cooper Drury
Dean of the College of Arts & Science

Dr. Scott Holan
Chair of the Department of Statistics

Leadership Auditorium
8:45-10:45 Am

Short Course (Only for Short Course registered participants)

  • See the above drop-down menu for details
Leadership Auditorium
10:45-11 AMBreak (Coffee/Refreshments)Leadership Lounge
11-12:30 PM

Session 1

Advances in Modeling and Uncertainty Assessment for Heat and Fire Extremes

  • Alan Rhoades: Snow-eater heatwaves of western North America

  • Nathan Lenssen: Climate Projection Uncertainty of Future Heat Extremes

  • Reetam Majumder: Semi-parametric bulk and tail regression using spline-based neural networks

 

Leadership Auditorium
12:30-1:45 PM

Lunch (on your own)

First Floor Seating Area
1:45-2:45 PM

Session 2

Modern Methods for Multivariate Extreme Value Modeling (I)

  • Emma Simpson: A comparison of generative deep learning methods for multivariate angular simulation 

  • Callum Murphy-Barltrop: Inference for multivariate extremes via a semi-parametric angular-radial model

Leadership Auditorium
2:45-3 PMBreak (Coffee/Refreshments)Leadership Lounge
3 - 4 PM

Session 3

Modern Methods for Multivariate Extreme Value Modeling (II)

  • Dan Cooley: Vector Spaces of Regularly Varying Random Variables (and maybe Inner Products) 

  • Benjamin Shaby: Bayesian model averaging of risk set probabilities using a geometric representation of multivariate extremes

Leadership Auditorium
4-5:30 PMPoster SessionLeadership Lounge
 

Tuesday, July 1, 2025

TimeTitleLocation
8:30-10 AM

Session 4

Quantifying Risk of Cascading and Compound Extremes in Critical Systems

  • Miguel de Carvalho: A Kolmogorov–Arnold Neural Model for Cascading Extremes 

  • Mary Salvana: A Self-Organized Criticality Model of Extreme Events and Cascading Disasters 

  • Mitchell Krock: Uncertainty quantification for critical energy systems during compound weather extremes via probabilistic simulation of climate data

Leadership Auditorium
10-10:15 AMBreak (Coffee/Refreshments)Leadership Lounge
10:15-12:15 PM

Session 5

Next-Gen Weather Generators for Extremes: From Stochastic Models to Huge Fourier Neural Ensembles

  • William Collins: Statistical Properties of Huge Ensembles of Climate Hindcasts 

  • Mark Risser: Surface temperature extremes produced by huge ensembles of summer 2023 hindcasts 

  • Michael Stein: What If I'm Not Only Interested in Extremes? Case Studies

Discussant: Kotamarthi, Rao

Leadership Auditorium
12:15-1:30 PM

Lunch (on your own)

First Floor Seating Area
1:30-3 PM

Session 6

Targeting the Tail: Conditional, Censored, and Threshold-Based Approaches

  • Max Thannheimer: Bayesian Inference for Functional Extreme Events Defined via Partially Unobserved Processes 

  • Christian Rohrbeck: A clustering framework for conditional extremes models

  • Simone Padoan: Statistical Prediction of Peaks Over a Threshold

Leadership Auditorium
3-3:30 PMBreak (Coffee/Refreshments)Leadership Lounge
3:30-5 PM

Session 7

Modeling Precipitation Extremes: Calibration, Comparison, and Decomposition

  • Jonathan Koh: Tail calibration of probabilistic forecasts 

  • Joshua North: A probabilistic model based approach for matrix factorization of extreme data 

  • Thomas Opitz: How to assess differences in heavy tails - Application to precipitation data

Leadership Auditorium
6 PM

Conference Dinner

Blufftop Bistro (Free for all registered participants)

  • Charter bus will be arranged to take guests from Tiger Hotel to the Bistro at 6PM
14020 W. Hwy BB Rocheport, MO 65279

Wednesday, July 2, 2025

TimeTitleLocation
8:30-10 AM

Session 8

Advances in Predictive Modeling and Tail Risk Assessment for Natural Hazards

  • Kate Saunders: Data-driven recommendations for enhancing real-time natural hazard warnings 

  • Moe Campos: A Structured Penalization Framework for Correcting Reporting Bias in Catastrophe Risk Models 

  • Jordan Richards: Generative modelling of joint environmental extremes using normalising flows

Leadership Auditorium
10-10:15 AMBreak (Coffee/Refreshments)Leadership Lounge
10:15-12:15 PM

Session 9

Scalable Inference and Spatial Analysis for Climate and Environmental Extremes

  • Benjamin Seiyon Lee: Comparative Analysis of Spatial Extremes Models and Scalable Inference for Large Spatial Datasets 

  • Noel Cressie: Spatio-Temporal Filtering for Mapping and Monitoring CO2 from space

  • RaphaĂ«l Huser: Neural Bayes estimators for fast inference with complex models in climate data applications

Discussant: Richard Smith

Leadership Auditorium
1:30-2:15 PM

Closing Remarks

  • We conclude the workshop and announce the Best Poster Award
Leadership Auditorium