This review examines the capabilities of artificial neural networks (ANNs) in elucidating the complex, nonlinear relationships between alloy composition, microstructure, processing conditions, and resultant properties. ANNs can effectively link microstructural evolution with material performance, enabling the prediction of mechanical properties, thermal behavior, corrosion resistance, and phase stability, as well as Additive manufacturing (AM) and welding process parameters. Regarding biomedical applications, the emergence of high-entropy alloys (HEAs) as promising metallic biomaterials is discussed, with a focus on biocompatibility prediction. The predictive modeling of passive-layer stability, ion dissolution rates, and cell interactions has enabled the design of HEAs tailored for implants. Elements such as Ti, Zr, Ta, and Mo exhibit bio-inertness, and their presence in HEAs is strategically leveraged using machine learning algorithms. The growing role of ANNs in optimizing metallurgical practices, from casting and welding to additive manufacturing, is also evaluated. While ANNs offer significant potential to streamline material discovery, they have limitations in terms of model generalization, data quality, and transparency. The integration of data-driven models with physics-informed approaches is discussed as a route toward more accurate and robust predictions. Cross-disciplinary research efforts must be prioritized to harness the power of ANNs for next-generation materials design.
Dewangan et al. (Wed,) studied this question.