Abstract Scope |
Due to the exponential growth in compositions, there exists the fundamental challenge of exploiting the enormous unique opportunities for designing high-entropy alloys (HEAs). Although the use of data-driven methods, such as supervised machine learning (ML), in HEA research, has achieved various degrees of success in predicting solid-solution phases, there exist inherent limitations such as available datasets and the effectiveness of selected features. High-entropy intermetallic phases, which exist either as composites or single-phase alloys, have the potential to deliver superior structural and functional properties. Nevertheless, the prediction of high-entropy intermetallics(HEI) raises yet another level of challenge. In this talk, we will present various ML models that utilize a combination of data-guided (phenomenological) and physics-based (adaptive) features, complemented with features engineering and experimental results, to explore the complex compositional landscape of HEI. We focus on two types of intermetallics namely Heusler and B2 ordered body-centered cubic phases. |