Self-driving manufacturing: accelerating materials discovery with adaptive closed-loop processing

Conventional human-centered process operations and explicit physics-based models face scalability limitations in high-dimensional, partially observable semiconductor and advanced materials processing. This review critically surveys the evolution of automation in materials processing, from heuristic-driven control through AI-assisted optimization to implicit, adaptive process modeling, where process-outcome relationships are learned directly from data and continuously refined through closed-loop feedback. We provide a unifying framework for developing scalable, robust, and knowledge-generating autonomous manufacturing systems.