MateriAlZ Seminar: Ashif Sikandar Iquebal
Friday, August 25, 2023, 11:00 am MST
Ashif Sikandar Iquebal
Assistant Professor of Industrial Engineering
School of Computing and Augmented Intelligence
Arizona State University
"Advancing and Accelerating Materials Characterization through Stochastic Inverse Modeling"
Zoom link | Passcode: 961011
MateriAlZ Seminar website | YouTube | Twitter
Abstract
A wide range of problems in science and engineering necessitates estimating critical quantities of interest (QoIs) through indirect measurements. A pertinent example lies within materials science, where pursuing comprehensive material structure and properties involves either exorbitantly expensive experiments limited to laboratories or costly destructive testing. For instance, the definitive method for appraising elastoplastic properties entails destructive tensile testing, while microstructure characterization demands intricate electron backscatter diffraction with high fidelity. These challenges fueled the research on estimating QoIs using indirect measurements, leading to developments in solving ill-posed inverse problems. Yet, a fundamental limitation of classical inverse problems is that they consider material properties to be deterministic, lacking uncertainty quantification. Bayesian inverse models attempt to overcome this issue but assume that the variability in the indirect measurements arises from measurement noise, thereby failing to account for the variability in the QoIs.
In this talk, we will explore the existing research on inverse problems and how they are limited in accurately estimating the QoIs and their variabilities. Subsequently, we will present our research on stochastic inverse problems that reformulates the classical inverse problem by considering the variability in the QoIs. This new approach leads to accurately estimating not just the QoIs but also the variabilities therein. Advances in stochastic inverse problems also open venues beyond material characterization, such as discovering the physics of complex processes via indirect measurements. We will show examples to demonstrate these applications.
Bio
Dr. Iquebal is an assistant professor of Industrial Engineering in the School of Computing and Augmented Intelligence at ASU. Prior to this, he obtained his B.S in Industrial Engineering from IIT Kharagpur, India and M.S. in Statistics and Ph.D. in Industrial Engineering from Texas A&M University. His research aims to bridge the gap between advanced manufacturing and statistical learning. More specifically, he is interested in stochastic inverse problems, active learning, and graphical models for accelerating materials characterization, discovering process physics, and generating causal inference. He received the Pritzker Doctoral Dissertation Award from the Institute of Industrial and Systems Engineering in 2021. His research papers have been winners/finalists for six best student paper/poster awards at INFORMS, IISE, IEEE and the American Statistical Associationconferences. His research is funded by DoD, NIH, and industry.