MUMS Transition Workshop and SPUQ: May 14-17, 2019


This workshop was held at Sitterson Hall, Room 014 on the campus of the University of North Carolina at Chapel Hill.


The SAMSI program on Model Uncertainty: Mathematical and Statistical has brought statisticians and applied mathematicians together with disciplinary scientists from a variety of fields, to better understand the effects of modeling and uncertainty on predictions, typically called Uncertainty Quantification (UQ). This joint workshop was designed to bring together SAMSI researchers in UQ and UQ researchers who could not participate in the MUMS program, both to share the results of SAMSI research and to assess the overall state of UQ. The first half of the workshop highlighted research from the MUMS program year, addressing foundational questions in UQ, the strength and weaknesses of models, and computational methods and applications of UQ. The second half focused on non-SAMSI researchers, under the auspices of the second workshop on Statistical (and mathematical) Perspectives on Uncertainty Quantification (SPUQ), following up on the first SPUQ conference, held in 2017 at Georgia Tech University.

** A Poster Session was held on the evening of Thursday, May 16, 2019. **

Schedule and Supporting Media

Printed Schedule
Speaker Titles/Abstracts
Participant List
Workshop Posters

Tuesday, May 14, 2019
Sitterson Hall, Room 014, University of North Carolina, Chapel Hill, NC

Time Description Speaker Slides
8:50-9:00am Welcome David Banks, SAMSI
Bruce Pitman, University at Buffalo SAMSI
9:00am-12:00pm Data Fusion Working Group (with a break)
A Review of Model Calibration Methods with an Application by Fusing Multiple Sources of Data from the Eruption of the Kilauea Volcano in 2018 Mengyang Gu, Johns Hopkins University
Bayesian CUSP Catastrophe Model for Sudden Changes Zhuoqiong He, University of Missouri
Data Fusion for Correlated, Shape-Restricted Curves with Varying Support Paul Speckman, University of Missouri
Bayesian Model Selection for a Linear Model with Grouped Covariates Dongchu Sun, University of Missouri
Bayesian Analysis for one-way MANOVA and a 3-Level Hierarchical Model Chengyuan Song, ECNU and SAMSI
Bayesian Smoothing Spline with a Generalized Constraint Operator Cong Lin, ECNU and SAMSI
12:00-1:30pm LUNCH
1:30-2:15pm UQ in Materials Working Group
SAMSI/NSF SEAS Materials Science Hackathon Results Ralph Smith, N.C. State University
2:15-5:15pm Prediction Uncertainty and Extrapolation Working Group
Stochastic Simulators: Issues, Methods, Unresolved Questions Evan Baker, University of Exeter
Embedding a Discrepancy in the Computer Model Pierre Barbillon, AgroParisTech and SAMSI
Some Strategies to Quantify Uncertainty for Extrapolation in Physical Systems Aaron Danielson, Simon Fraser University
Estimating Ocean Circulation Structure: Deterministic and Stochastic Simulators Radu Herbei, Ohio State University
Practical Bayesian Optimization for Agent Based Transportation Simulators Laura Schultz and Vadim Sokolov, George Mason University
5:30pm Shuttle to Hotel

Wednesday, May 15, 2019
Sitterson Hall, Room 014, University of North Carolina, Chapel Hill, NC

Time Description Speaker Slides
9:00-11:45am Reduced Order Models Working Group
Reduced Order Modeling of a Biphasic Cartilage Mixture Model under Dynamic Compressive Loading Mansoor Haider, N.C. State University
Parameter Subset Selection for Coupled Flow and Deformation Modeling Sue Minkoff, University of Texas
Dimension Reduction and Global Sensitivity Metrics using Active Subspaces for Coupled Flow and Deformation Modeling Hyunjung Lee, Marquette University
A Coupled Parallel Partial Emulator for Flow and Deformation Modeling Elaine Spiller, Marquette University
Parameter-Dependent Surrogate Model Development and Control Design for PZT Bimorph Actuators Employed for Micro-Air Vehicles Nikolas Bravo, N.C. State University
11:45am-1:15pm LUNCH
(Working Group Leaders in Room FB009 – to Discuss Final Report)
1:15-3:30pm Storm Surge Hazard and Risk Working Group
Taylor Asher, UNC-Chapel Hill
Some Thoughts on Estimating Input Distribution of Storm Surge Simulations Whitney Huang, University of Victoria
Emulation for Forecasting Storm Surge Matthew Plumlee, Northwestern University
Emulation for Large Storm Surge Simulation Ensemble Won Chang, University of Cincinnati
An Emulator Approach for Quantifying the Risk Due to Storm Surge Pulong Ma, SAMSI
3:30-4:00pm BREAK
4:00-5:45pm Foundations of Model Uncertainty Working Group (with a break)
Variable Selection in the Discrepancy Function Associated with a Simulator Pierre Barbillon, AgroParis Tech and SAMSI
Are Reported Likelihood Ratios Well Calibrated? Jan Hannig, UNC-Chapel Hill
Deep Fiducial Inference and Approximate Fiducial Computation Gang Li, UNC-Chapel Hill
Model Selection in the Context of Computer Models Rui Paulo, Universidade de Lisboa
6:00pm Shuttle to Hotel

Thursday, May 16, 2019
Sitterson Hall, Room 014, University of North Carolina, Chapel Hill, NC

Time Description Speaker Slides
8:50-9:00am Opening Remarks SPUQ Elaine Spiller, Marquette University and SAMSI
Transformation and Additivity in Gaussian Process Roshan Joseph, Georgia Institute of Technology
9:00-10:00am Keynote Address: Navier-Stokes, Spatial-temporal Kriging and Combustion Stability: a prominent example of physics-based analytics Jeff Wu, Georgia Institute of Technology
10:00-10:30am BREAK
10:30am-12:00pm Invited Session on Design in UQ
Robust Experimental Design for Model Calibration William Brenneman, Proctor and Gamble
A Sequential Design Approach for Calibrating a Dynamic Population Growth Model Devon Lin, Queens University
Design of Experiments for Calibration of Computational Models David Woods, University of Southampton
12:00-1:30pm LUNCH
1:30-3:00pm Invited Session on Exascale and Dimensional Analysis
Some Pieces of Exascale Uncertainty Quantification Earl Lawrence, Los Alamos National Laboratory
Dimensional Analysis in Computer Experiments Will Welch, University of British Columbia
Gradient-Free Construction of Active Subspaces for Dimension Reduction Brian Williams, Los Alamos National Laboratory
3:00-3:30pm BREAK
3:30-5:00pm Invited Session on “Perspectives on Sensitivity in UQ”
Mathematical Perspective Pierre Gremaud, N.C. State University
Statistical Perspective Max Morris, Iowa State University
5:00-5:15pm Discussion Ralph Smith, N.C. State University
5:15-5:30pm Floor Discussion
5:30-7:00pm Poster Session and Reception
7:15pm Shuttle to Hotel

Friday, May 17, 2019
Sitterson Hall, Room 014, University of North Carolina, Chapel Hill, NC

Time Description Speaker Slides
9:00-10:30am Invited Session on Modeling and Learning in UQ
Computer Experiments with Binary Time Series and Applications to Cell Biology: Modeling, Estimation, and Calibration Ying Hung, Rutgers University
Simulation Experiments and Uncertainty Quantification in Remote Sensing Emily Kang, University of Cincinnati
Gaussian Process Model Assisted Active Learning of Physical Laws Lulu Kang, Illinois Institute of Technology
10:30-11:00am BREAK
11:00am-12:30pm Invited Session on Getting Inside the Black Box
Adaptive Step-Size Selection for State-Space Probabilistic Differential Equation Solvers Oksana Chkrebtii, Ohio State University
Representing Model Inadequacy in Reduced Models of Interacting Systems Rebecca Morrison, Massachusetts Institute of Technology
Opening up the Black Box: Gaussian Process Modeling using Information from Partial Differential Equation Models Matthias Tan, City University of Hong Kong
12:30-2:00pm LUNCH
2:00-3:00pm Invited Session on Emulation and Calibration
Competing Complexities in Bayesian Inverse Problems: Models and Distributions Simon Cotter, University of Manchester
Transformation and Additivity in Gaussian Process Roshan Joseph, Georgia Institute of Technology
3:00-3:30pm BREAK
3:30-4:50pm New Researchers Invited Session
Clustering based gaussian process emulation and calibration of a stochastic agent-based model Arindam Fadikar, Virginia Polytechnic Institute and State University
Structural model discrepancy in Nuclear Energy Density Functional Simulators Michael Grosskopf, Los Alamos National Laboratory
cmenet: a new method for bi-level variable selection of conditional main effects Simon Mak, Georgia Institute of Technology
4:50-5:00pm → Farewell and Shuttle to Airport

Questions: email