| Abstract Scope |
Accelerated refinement of multiscale material models (ranging from atomistic to the macroscale) demands the development and implementation of novel high throughput strategies in both experimentation and physics-based simulations, and their seamless integration using the emergent AI/ML (artificial intelligence/machine learning) toolsets. This talk presents a novel information gain-driven Bayesian machine learning framework with the following salient features: (i) explicit consideration of the physics parameters as inputs (i.e., regressors) in the formulation of the material physics models needed to drive materials discovery, design and development workflows; (ii) Bayesian design of experiments strategies that maximize the expected information gain; (iii) versatile feature engineering for multiscale material internal structure using the formalism of n-point spatial correlations; (iv) amenable to a broad suite of surrogate model building approaches; and (vi) Markov chain Monte Carlo (MCMC)-based computation of posteriors for physics parameters using available experimental observations (usually disparate, incomplete, and uncertain). |