Deep Learning for Materials and Additive Manufacturing: Towards AI-Enabled Closed-Loop Systems

Artificial Intelligence (AI) is rapidly transforming both materials discovery and advanced manufacturing, yet these domains have traditionally evolved along largely independent research trajectories. Materials informatics has primarily focused on predicting material properties and discovering new compositions, whereas AI in Additive Manufacturing (AM) has largely focused on process monitoring, defect detection, and parameter optimization. Recent advances in DL architectures, multimodal data integration, and autonomous experimentation are beginning to bridge these domains by enabling closed-loop materials-manufacturing systems in which materials design, manufacturing processes, and real-time sensing are integrated through data-driven feedback. This review examines the emerging role of Deep Learning (DL) as a unifying framework connecting materials chemistry and AM. We summarize key DL architectures-including convolutional neural networks, graph neural networks, generative models, transformer-based systems, and multimodal foundation models-and discuss their applications across materials discovery, microstructure characterization, process monitoring, and manufacturing optimization. Emphasis is placed on how these models can integrate heterogeneous data sources spanning atomic-scale simulations, microstructural imaging, and manufacturing process monitoring. We further explore how integrating Artificial Intelligence (AI) with AM technologies enables closed-loop workflows in which ML models guide materials design, optimize manufacturing parameters, and iteratively update predictive models using experimental feedback. Such systems have the potential to significantly accelerate materials development and improve manufacturing reliability. Finally, we discuss key research challenges and future directions, including physics-informed learning, data infrastructure development, and autonomous experimentation systems that may enable fully integrated AI-driven materials-manufacturing ecosystems.