Published December 22, 2025 | Version 1.0
Dataset Open

Research Data Supporting "Inverse Design of Bespoke Interatomic Potentials via Information-Matching Active Learning"

Authors/Creators

  • 1. ROR icon Brigham Young University
  • 2. ROR icon Sandia National Laboratories California
  • 3. ROR icon Lawrence Livermore National Laboratory
  • 4. ROR icon University of Minnesota System
  • 5. ROR icon University of California, Los Angeles
  • 6. ROR icon National Ignition Facility

Description

Overview

This dataset contains the research data supporting the study “Inverse Design of Bespoke Interatomic Potentials via Active Learning by Information-Matching.” The data were generated as part of a computational investigation of active learning strategies for the inverse design of interatomic potentials using an information-matching approach, with the goal of targeting accurate predictions of plastic strength.

Accompanying GitHub repository: https://github.com/yonatank93/information-matching_eam-strength

Specific requirements

Content

  • eam-snap - Containing scripts and data for the cases fitted to EAM and SNAP proxy ground truth. These 2 cases are combined into a single folder because they use the same dataset.
  • dft - Containing scripts and data for the case fitted to DFT ground truth.
  • download_additional_data.py - A utility script to download the additional data.

Contact

For questions about the code or data, please contact Yonatan Kurniawan at kurniawanyo@outlook.com.

Files

Files (26.0 GB)

Name Size
md5:e254d860588cab4937e092b22b911f76
26.0 GB Download

Additional details

Funding

United States Department of Energy
Predictive Atomistic Materials Simulations with Uncertainty Quantification 23-SI-006

Software

Repository URL
https://github.com/yonatank93/information-matching_eam-strength
Programming language
Python