From Models to Decisions: Harnessing Physics and Machine Learning for Reliable Predictions of Material and Biological Systems

Recent advancements in computational engineering, combined with the rapid growth of machine learning techniques, have transformed the integration of experimental data to significantly enhance the predictive capabilities of continuum mechanics models for physical and engineering systems. However, a critical challenge in utilizing these models for high-consequence decision-making is ensuring the reliability of their predictions, particularly in the face of significant data uncertainties and inherent modeling errors. This talk introduces the Occam-Plausibility Algorithm (OPAL), a comprehensive framework for model validation and selection under quantified uncertainty. Leveraging Bayesian inference and the notion of model plausibility, OPAL systematically identifies the “optimal” model with reliable prediction from a large family of possible models with varying fidelity and complexity. The effectiveness of this framework will be demonstrated through its application in modeling plastic deformation in polycrystalline metallic materials and in constructing reliable deep-learning surrogate models for porous ceramic aerogels. Furthermore, the presentation will introduce efficient algorithms critical for leveraging validated finite element models in predicting and optimizing biological and engineering systems. Case studies will illustrate these advancements, including the early prediction of subject-specific brain tumor growth and the risk-averse design of high-performance building insulation. Finally, the presentation will outline a vision for addressing critical challenges in predictive digital twins and their potential across diverse scientific and engineering domains.

Danial Faghihi is an assistant professor in the Department of Mechanical Engineering at the University at Buffalo (UB), with an affiliation in the Department of Civil, Structural and Environmental Engineering and the Institute for Artificial Intelligence and Data Science. Before joining UB in 2019, he was a research scientist at the Oden Institute for Computational Engineering and Sciences at the University of Texas at Austin. He holds a PhD in structural engineering and mechanics from Louisiana State University. His research focuses on predictive computational modeling of complex materials and biological systems, with an emphasis on scalable computational frameworks at the intersection of finite element modeling, scientific machine learning, and high-performance computing. In 2022, he received the National Science Foundation CAREER Award and has authored over 40 journal articles in computational and applied mechanics.