Adaptive Discovery and Mixed-Variable Bayesian Optimization of Emerging Material Systems

Abstract: Design of new materials is characterized by several challenges such as high-dimensionality of the atomic structure-composition variable space, formidable cost of directly using high-fidelity simulations for design optimization, dispersity in literature-reported similar materials and synthesis methods, complex physical mechanisms and mixed qualitative and quantitative design variables that lead to a disjointed design space.

Even though machine learning (ML) techniques have been employed to expedite materials innovation, existing methods treat ML and design optimization as two separate processes, failing to resolve the fundamental challenges associated with high dimensionality and mixed-variable complexity. We have developed an ML enhanced mixed-variable material design optimization framework to efficiently extract useful information from existing data in literature and physics-based simulations to guide the autonomous search for optimal materials.

In this talk, we will introduce state-of-the-art statistical inference and AI methods for designing emerging materials systems, such as microelectronics, polymer nanocomposites, Metal-Organic Frameworks, metamaterials, etc. We will discuss the paradigm shifts in materials discovery and present research developments in novel data-driven and AI enabled design methods that integrate machine learning, mixed-variable Latent Variable Gaussian Process modeling, Bayesian optimization, topology optimization and the concept of digital twins, for autonomous materials discovery.