Leslie Horace Wins Student Award at CoDA 2025
Leslie Horace Wins Student Award at CoDA 2025
Leslie Horace, a New Mexico Consortium (NMC) student scientist and Los Alamos National Laboratory (LANL) contractor, recently attended the Conference on Data Analysis (CoDA), held on February 25, 2025 in Santa Fe, New Mexico, where she won an award for her presentation.
This conference, held annually, highlights data-driven problems of interest to the Department of Energy. Presentations at this event featured research from the Department of Energy national laboratories, academia, and industry.
Horace is a graduate student at Georgia Tech working on a MS degree in Computer Science with a focus on Artificial Intelligence and Machine Learning. At LANL, she is currently working on the application of machine learning techniques for surrogate modeling for model-simulation application code and the application of machine learning techniques for the prediction of energy and power demands in HPC datacenters.
The NMC is pleased to announce that her poster at CoDA,titled “Machine Learning Surrogates for Simulating Electrostatic Potential of Heterogeneous Materials”, won an honorable mention award in the student poster session.
Simulating complex physical systems with diverse material properties and dynamic interactions often requires solving 2D or 3D partial differential equations (PDEs) using complex iterative methods. This is crucial for various industries, such as heat transfer in industrial components, electrostatics in semiconductor manufacturing, and chemical reactions in battery materials.
Nevertheless, the high-resolution calculations required over extensive parameter spaces result in high costs, extended processing times, and limited scalability. This research shows that machine learning surrogate modeling offers a different strategy by approximating complex physical behaviors in high-fidelity data. Network models can extract patterns from a subset of simulation samples, finding a way around many of the restrictions of traditional methods.
Horace and her colleague’s work focuses on developing machine learning models to address two critical challenges: spatial data compression and temporal evolution of the simulation. Her poster presentation covered the applications, models, and performance results obtained across several dimensions.
Horace’s mentor William M. Jones (CCU/LANL HPC-DES) says, “I’ve been working with Leslie since October 2023 and since that time, she has made major contributions to ongoing projects with two different HPC groups at LANL. She’s an asset to our teams, and I look forward to her continued involvement with NMC and LANL.”
Congratulations Leslie on your excellent work!