Machine learning (ML) and computational modeling have transformed metallurgy into a predictive and data-driven field with accelerated alloy design, optimized processing pathways, and improved microstructural properties. This review discusses the emerging trends in atomistic, mesoscale, and continuum modeling studies, such as Density Functional Theory, molecular dynamics, Monte Carlo simulations, phase-field modeling, crystal plasticity, and finite element analysis. More physics-based models are being complemented by ML models, such as classical statistical learning, deep neural networks, and physics-informed neural networks (PINNs). The combination of models has made it possible enabled significant advancements in designing inverse alloys, predicting microstructures, characterizing defects, and controlling processes in real time in casting, thermomechanical processing, heat treatment, and additive manufacturing. Based on a wide body of peer-reviewed literature, this review critically evaluates the strengths, limitations, and strategies for integrating ML into metallurgical systems. High-performance aluminum alloys, energy-relevant materials, and new ideas for digital twins have been emphasized. Major challenges, such as data quality, multiscale coupling, quantification of uncertainty, and interpretability of the model, are addressed, and future research directions based on autonomous and mechanism-aware metallurgical systems are described.
T et al. (Thu,) studied this question.