Uncertainty Quantification, Model Validation, and Prediction Science

Featuring Professor Roger Ghanem
Departments of Aerospace and Mechanical Engineering and Civil Engineering
University of Southern California

Location: Engineering Gateway E3161
Free and open to the public

Abstract:
This talk will present some current challenges that stochastic analysis must address as it becomes a cornerstone of prediction science.  Hilbert space decompositions of stochastic processes will be used as the analysis vehicle of choice to delineate and characterize these challenges.  Particular attention will be devoted to the definition of an error budget to be used in the adaptive refinement of prediction instruments.  Through a suitable identification of these instruments with operators on product spaces, the adaptive refinement in question can be viewed as defining a constructive path for model validation. Evaluating the error budget entails defining novel statistics that permit the propagation of the effect of data limitations to the stochastic predictions.  Novel procedures for estimating these statistics are described and challenges in their implementation are noted. 

Furthermore, the need for reduced-order analysis and models is reiterated for the accurate and proper application of stochastic analysis to physically realistic problems.  Procedures for the construction of such models are described together with related theoretical and computational challenges.  Since issues of reduced-order analysis pertain to stochastic multi-scale representations, challenges and on-going research in multi-scale stochastic analysis will also be described.

About the Speaker:
Dr. Ghanem has a Ph.D. in civil engineering from Rice University.  He has served on the faculty of SUNY Buffalo, Johns Hopkins University, and is currently on the faculty at USC in the departments of aerospace and mechanical engineering and civil engineering.  Dr. Ghanem is the co-author of a book on Stochastic Finite Elements, and of more than 200 technical publications in the fields of dynamics and stochastic analysis.  He currently serves as the Chairman of the Executive Committee of the ASCE Engineering Mechanics Division.

He specializes in the area of uncertainty quantification and management.  Specifically, Dr. Ghanem is working on the development of representations of uncertainty that permit the quantitative stochastic characterization of related model-based predictions. Over the past 15 years, he has pioneered the methods of Polynomial Chaos for Uncertainty representation and propagation.  He is currently working on links between multi-scale behavior and uncertainty quantification with applications drawn from throughout science and engineering.