Machine Learning Design of Single-Atom Catalysts for Carbon Dioxide Reduction

Executive Summary

This project will use high-throughput simulations and machine learning to accelerate the design of highly selective catalysts to produce valuable products from carbon dioxide. Recycling carbon dioxide to produce high-value products can reduce greenhouse gas emissions and provide a grid-scale energy storage system.

Publications, Presentations, and Patents

Schwalbe-Koda, Daniel. 2023. "mkite: A distributed computing platform for high-throughput materials simulations." Computational Materials Science 230: 112439. https://doi.org/10.1016/j.commatsci.2023.112439

Vita, Joshua A., and Schwalbe-Koda, Daniel. 2023. “Data efficiency and extrapolation trends in neural network interatomic potentials.” Machine Learning: Science and Technology 4 (3): 035031. https://doi.org/10.1088/2632-2153/acf115

Daniel Schwalbe-Koda, “Interatomic Potentials Enabled by Machine Learning” (Presentation, FairMAT-IKZ Winter School, virtual, January 2023).

Daniel Schwalbe-Koda,“Adversarial Sampling and Extrapolation Trends in Neural Network Potentials” (Presentation, Seminars on Machine Learning in Quantum Chemistry and Quantum Computing for Quantum Chemistry, virtual, March 2023).

Daniel Schwalbe-Koda,“Data Efficiency and Extrapolation Trends in Neural Network Interatomic Potentials” (Presentation, Telluride Science & Innovation Center, Telluride, CO, June 2023).

Daniel Schwalbe-Koda,“mkite: a Distributed Computing Platform for High-Throughput Materials Simulation” (Presentation, 2023 ACS Fall Meeting, San Francisco, CA, August 2023).

Daniel Schwalbe-Koda,“Data Efficiency and Extrapolation Trends in Neural Network Interatomic Potentials” (Presentation at the 2023 ACS Fall Meeting, San Francisco, CA, August 2023).

Daniel Schwalbe-Koda, Zhou, Fei, and Lordi, Vincenzo, “Identifying and Quantifying Uncertainty in Fitted Interatomic Potentials for Molecular Dynamics” (Presentation, UMN Materials + DS Workshop, Minneapolis, MN, September 2023).