LESets Machine Learning Model: A Revolutionary Approach to Accurately Predicting High-Entropy Alloy Properties by Capturing Local Atomic Interactions in Disordered Materials

Graph Neural Networks (GNNs) are a powerful tool for predicting material properties by capturing intricate atomic interactions within various materials. They encode atoms as nodes and chemical bonds as edges, allowing for a detailed representation of molecular and crystalline structures. High-entropy alloys (HEAs) present a challenge in modeling due to their combinatorial complexity and lack of periodic atomic order. Existing methods struggle to capture the nuanced interactions within HEAs, making it difficult to predict their properties accurately. The LESets model is a novel approach designed to accurately predict the properties of complex materials like HEAs. It represents HEAs as a collection of local environment (LE) graphs, capturing the detailed atomic interactions within the alloy. The LESets model constructs a graph for each local environment in a HEA, representing each local environment within the alloy as a separate graph. It then aggregates these local environment graphs to form a global representation of the HEA, enabling accurate prediction of various material properties. The LESets model outperformed traditional models in predicting the mechanical properties of HEAs, achieving a higher coefficient of determination (R2) and lower mean absolute error (MAE) across multiple random data splits. It demonstrated robustness to data variations, making it a foundational tool for materials science research and development. LESets Machine Learning Model offers a revolutionary approach to accurately predicting High-Entropy Alloy Properties by capturing local atomic interactions in disordered materials. It provides a competitive advantage in materials science research and development, offering customizable AI solutions aligned with specific business needs.