Atomic Property

Predicting atomic properties is crucial for materials science and chemistry, aiming to bypass computationally expensive methods like quantum calculations. Current research focuses on leveraging machine learning, particularly graph neural networks and message-passing neural networks, often employing transfer learning and multi-task learning strategies to improve prediction accuracy, especially with limited data. These advancements enable efficient prediction of both global material properties and individual atomic characteristics, such as magnetic moments or charge transfer, improving the speed and efficiency of materials discovery and design.

Papers

April 20, 2024

Transfer Learning for Molecular Property Predictions from Small Data Sets
Thorren Kirschbaum, Annika Bande
Transfer Learning Molecular Property Prediction Small Datasets Molecular Property Molecular Descriptor Quantum Chemical Calculation Atomic Property

October 25, 2023

From Molecules to Materials: Pre-training Large Generalizable Models for Atomic Property Prediction
Nima Shoghi, Adeesh Kolluru, John R. Kitchin, Zachary W. Ulissi, C. Lawrence Zitnick, Brandon M. Wood
Known Molecule Material Response Generalist Model Chemical Domain Atomic Property

February 4, 2022

Multi-task graph neural networks for simultaneous prediction of global and atomic properties in ferromagnetic systems
Massimiliano Lupo Pasini, Pei Zhang, Samuel Temple Reeve, Jong Youl Choi
Neural Network Multi Task Joint predictioN Material Property Ab Initio Atomic Property Ferromagnetic Particle

Atomic Property

Papers

Transfer Learning for Molecular Property Predictions from Small Data Sets

From Molecules to Materials: Pre-training Large Generalizable Models for Atomic Property Prediction

Multi-task graph neural networks for simultaneous prediction of global and atomic properties in ferromagnetic systems