Business & Industrial Statistics Section Workshop 2016
Uncertainty Quantification and Optimization for Complex Models with Gaussian Processes
The Business and Industrial Statistics Section (BISS) is pleased to partner with the CANSSI Collaborative Research Team on Computer and Physical Models in Earth, Ocean and Atmospheric Sciences.
Sunday, May 29, 2016 from 9 am to 4 pm, lunch included — Thistle 246
Derek Bingham, Simon Fraser University
Jason Loeppky, The University of British Columbia-Okanagan
Complex computer models are ubiquitous in virtually every area of Science. Practitioners and Statisticians are routinely faced with using these virtual models to study real world phenomena. This workshop is focused on using these models in a scientific and disciplined way to help understand the processes being modelled. The computational complexities of the model place various restrictions on the way the model can be used. This workshop is focused on the practical aspects of using a non-parametric regression techniques, specifically a Gaussian process to build statistical models for various aspects of the process being studied.The workshop will begin with a brief review of Bayesian inference and a quick introduction to Hamiltonian Markov Chains. We will then introduce the Gaussian process emulator and illustrate the use of a GP emulator using the RStan package in R. The remainder of the course will be focused on applications of the GP model for solving real world problems with a specific emphasis in combining multiple sources of data for model calibration. The course will feature hands-on activities to illustrate the practical uses of the model. Participants are encouraged to bring laptops with R and RStan preinstalled. Additional packages and data will be announced closer to the date of the workshop.
Derek Bingham is a professor in the Department of Statistics and Actuarial Science at Simon Fraser University. His research interests lie in the area of the design and analysis of experiments in the physical and engineering sciences. He has made contributions in the development of statistical methodology for the design and analysis of experiments on complex computer simulators. Much of his research has been motivated through scientific collaborations. Recent work on computer experiments and uncertainty quantification is the direct result of collaboration with scientists at Argonne National Laboratory, Los Alamos National Laboratory, the University of Michigan’s Center for Radiative Shock Hydrodynamics and Texas A&M’s Center for Exascale Computing. He is currently a team member of the CANSSI Collaborative Research Team project Statistical modeling of the world: Computer and physical models in earth, ocean, and atmospheric sciences.
Jason Loeppky is an Associate Professor of Statistics at The University of British Columbia's Okanagan campus. He is a professional, accredited Statistician (P.Stat.,Statistical Society of Canada). He has been developing methodology for complex models since 2004 and has applied this methodology to problems in engineering and biology. His research is funded by the Natural Sciences and Engineering Research Council, and he is the recent recipient of the prestigious NSERC Discovery Accelerator supplement.
Financial Support Available for Student Participants
Students who attend this workshop will be eligible for financial support, including workshop registration and partial travel funding. To apply, register for the workshop and email your application to Derek Bingham.
Your application should include a receipt indicating registration in the BISS workshop, contact information for your supervisor, and a brief statement of interest. Travel funding will be up to $100 (Ontario), $200 Quebec, $300 Manitoba, Saskatchewan, Alberta, Atlantic Canada and $350 British Columbia. Students must participate in the workshop to receive travel funding and reimbursement of the workshop registration. The application deadline is April 1, 2016. Funding decisions will be made by April 15, 2016. Organizers reserve the right to limit funding.