Machine Learning Paves the Way for High Entropy Compounds Exploration: Challenges, Progress, and Outlook

Area of Science:

• Materials Science
• Computational Chemistry
• Data Science

Background:

• High entropy compounds (HECs) offer vast compositional space and tunability, attracting significant research interest.
• The intricate structure-property relationships in HECs challenge conventional experimental and computational methods.
• Machine learning (ML) provides a powerful framework to navigate the complexity of HECs.

Purpose of the Study:

• To highlight the indispensable role of machine learning in advancing high entropy compound research.
• To explore the microscopic and macroscopic applications of ML in understanding HEC properties.
• To underscore the potential of ML in accelerating the design and discovery of novel functional HECs.

Main Methods:

• Utilizing various ML algorithms, including traditional methods and deep neural networks.
• Emphasizing the critical importance of comprehensive data collection, feature engineering, and rigorous model validation (e.g., cross-validation).
• Applying ML for Hamiltonian modeling, property prediction, phase structure analysis, and simulation potential development.

Main Results:

• ML enables atomic-level investigations by modeling system Hamiltonians.
• ML facilitates the analysis of macroscopic material properties like hardness, melting point, and ductility.
• ML aids in predicting phase structures, stability, and designing functional materials.

Conclusions:

• Machine learning is an essential tool for understanding and exploiting the potential of high entropy compounds.
• ML accelerates the exploration of HECs, paving the way for Artificial Intelligence-assisted material discovery.
• Further research into ML applications for magnetic and device materials is warranted.