Abstract: Gaussian process (GP) has been a cornerstone of machine learning in a Bayesian context and is one of a few machine learning toolboxes that comes with uncertainty quantification. Using GP as the underlying surrogate model, Bayesian optimization (BO) methods have been a successful approach for active learning and optimization with applications across multiple disciplines. Of particular importance to this talk, we will explore the applications to multiscale computational solid mechanics and materials science. To this end, numerous multiscale Integrated Computational Materials Engineering (ICME) models have been developed over a spectrum of nano-scale, mesoscale, and macro-scale, including density functional theory, molecular dynamics, kinetic Monte Carlo, phase-field, crystal plasticity finite element model.
In the first half of the talk, we will discuss a generic and versatile BO approach to tackle a handful of optimization problems, including known and unknown constraints, multi-objective, multi-fidelity, parallelization on high-performance computers, Big Data, and high-dimensional problems. In the second half of the talk, we will discuss the applications of GP/BO to several ICME models, in the materials design under uncertainty context and in the spirit of the Material Genome Initiative (2011). In particular, using ICME applications as forward models in the process-structure-property relationship, we will discuss how GP/BO fits in as the data-driven fourth-paradigm for materials design using ICME models.
Speaker’s Bio: Anh Tran is currently a limited-term senior member of technical staff of Uncertainty Quantification and Optimization in Sandia National Laboratories, Albuquerque, New Mexico. He obtained his B.S., M.S., and Ph.D. in Mechanical Engineering from Georgia Institute of Technology in December 2011 and December 2018, respectively. He also held an M.S. in Applied Mathematics from Georgia Southern University in May 2014. His research interests include optimization, uncertainty quantification, and machine learning, with applications to multi-scale materials science.
Last Updated: March 22, 2022 - 7:33 am