Advancements in Machine Learning for the Development of Cementitious Composites Toward an Intelligent and Green Lifecycle: A State-of-the-Art Review

As fundamental construction materials, cementitious composites face significant challenges under conventional development approaches, including carbon-intensive production, resource-intensive experimentation, and inefficient design processes. With the emergence of machine learning (ML) as a transformative solution to these limitations, this study presents a state-of-the-art review of existing research to highlight its potential in advancing the development of cementitious composites with intelligent and green lifecycles. The review first provides a foundational introduction to ML concepts and then proposes a novel four-quadrant classification framework to systematically organize current ML applications in the field. The ML-driven innovations integrate the component–structure–process–performance relationships of cementitious composites through sustainable material selection, effective characterization, accurate performance prediction, and optimized inverse design, collectively promoting a paradigm shift toward intelligent and green lifecycles. Furthermore, critical implementation challenges are examined across technical, methodological, and operational dimensions, together with corresponding solution strategies. This review ultimately offers both a conceptual framework and practical implementation guidelines for the development of next-generation sustainable construction materials.