Azadeh Sheidaei, Iowa State University
Khiem Nguyen, University of Glasgow
Mohammad Saber Hashemi, Iowa State University
This mini-symposium focuses on computational inverse problems whose targets are modeling and ultimately designing microstructural materials at the microscale and the mesoscale with manufacturability considerations using state-of-the-art machine learning techniques. It is a highly multidisciplinary field of great importance with applications in computational mechanics and materials, including automatic discovery of governing equations, surrogate modeling, and advanced materials design. However, many inverse problems are known to be challenging due to, e.g., high-dimensional feature spaces, complicated or unspecified models, limited observed data, stochasticity, and computationally intensive physical solver or simulator. Recent advances in machine learning provide alternatives owing to its several ingredients: (1) deep neural networks as flexible frameworks for representing high-dimensional and complicated functions and mappings, (2) generative models for creating more relevant but expensively fabricated data, (3) transformer-based networks as powerful tools to infer long-range relationships in input data and to digest and combine mixed-format input data seamlessly, and (4) advanced online optimization algorithms considering physical constraints and invariants through expert knowledge in physics-informed learning. It is only natural to leverage these developments in the physics-aware study of inverse problems.
We are particularly interested in research that involves inverse modeling and design problems in computational mechanics and materials. Some examples of these material systems are cellular, composites, biological tissue, implants, rocks, and multifunctional materials.
