Collection 

Machine Learning Interatomic Potentials in Computational Materials

Submission status
Closed
Submission deadline

Machine learning interatomic potentials (MLIPs) have become an essential tool to enable long-time scale simulations of materials and molecules at unprecedented accuracies. The aim of this collection is to showcase cutting-edge developments in MLIP architectures, data generation techniques, and innovative sampling methods that push the boundaries of accuracy, efficiency, and applicability in atomic-scale simulations.

This Collection welcomes the following topics, including but not limited to:

  1. Architectural Advances
    1. New model architectures that enhance the accuracy, interpretability, or transferability of interatomic potentials.
    2. Novel algorithms for reducing model complexity while maintaining predictive power.
  2. Data Generation and Processing
    1. High-quality data generation strategies, including ab initio data curation and active learning techniques that minimize the need for computationally expensive data.
    2. Approaches for ensuring data diversity and coverage to improve MLIP robustness, especially for rare events.
    3. Advanced sampling methods that improve MLIP efficiency in handling diverse materials and molecular configurations.
  3. Applications
    1. In-depth studies demonstrating MLIPs in real-world applications, including material design, chemical reactions, and biological molecule simulations.
    2. Novel applications where MLIPs offer unique insights beyond traditional potential models or ab initio methods.
machine learning

Editors

Shyue Ping Ong, PhD, University of California San Diego, USA
Dr. Ong received his Ph.D. in Materials Science and Engineering from the Massachusetts Institute of Technology (MIT) in 2011. He was subsequently appointed as a Senior Research Associate at MIT. Dr. Ong’s research and teaching vision is to be a leader in bringing forth a data-driven future for materials design. During his PhD, Dr. Ong was part of a research team that developed a high-throughput (HT) computational framework for materials design that led to the discovery of several novel Li-ion battery cathode materials. In Oct 2011, this framework was spun off into the Materials Project (www.materialsproject.org), an open science project to make the calculated properties for all known inorganic materials publicly available to researchers to accelerate materials innovation. This project is now a cornerstone of the Materials Genome Initiative and has garnered significant press worldwide. Intersecting the disciplines of materials science and information science, Dr. Ong’s research will develop new materials informatics approaches to create and analyze rich materials data and apply them to the design of new energy materials. He will teach courses that will train new practitioners in up-to-date computational materials science and informatics approaches. Dr. Ong will also continue his efforts with the Materials Project to promote open access to materials information access and analysis codes.

Bingqing Cheng, PhD, University of California Berkeley, USA
Dr. Cheng is an Assistant Professor at the Colleage of Chemistry, UC Berkeley. She did a PhD (09/2014–02/2019) in Materials Science at École Polytechnique Fédérale de Lausanne (EPFL), supervised by Michele Ceriotti, a Master's degree in The University of Hong Kong, and a joint Bachelor's degree in The University of Hong Kong & Shanghai Jiao Tong University. She is interested in developing methods to extend the scope of atomistic simulations, in order to understand and predict materials properties that are hard to access. The group deploys and designs a combination of techniques encompassing machine learning, enhanced sampling, path-integral molecular dynamics, and free energy estimation. The systems of study include energy materials, aqueous systems, and matter under extreme conditions.


Xiangguo Li, PhD, Sun Yat-sen University, China
Dr. Xiangguo Li is an Associate Professor at the School of Materials, Sun Yat-sen University. He obtained the PhD in Materials Science at University of Florida in 2014. He was an assistant professor at University of Wisconsin-Madison from 2020 to 2021. The current research focuses on combining machine learning methods and physical theory knowledge in the prediction of material properties and the discovery and exploration of new materials. Specifically, at the mesoscopic near-micrometer scale, he is one of the earliest scholars to engage in research in the emerging field of machine learning potential field. He has developed a series of machine learning potential fields for alloy materials, and used them to reveal a potential intrinsic mechanism for the high strength of high-entropy alloys, making important contributions to the extension of machine learning potential fields to multi-element alloy systems; at the mesoscopic nanoscale, he revealed the electronic structure and transport properties of the interface of two-dimensional materials, providing a theoretical basis for the optimization of quantum devices. In addition, a series of material property databases have been established through high-throughput computing; at the microscopic scale, through the method of first-principles calculations, a quantum technology based on molecular deformation has been proposed and a new magnetic mechanism has been discovered.