Congratulations to the Winner of the SCM ParAMS Machine Learning Challenge 2025!
The SCM ParAMS Challenge 2025 invited WATOC 2025 attendees to improve the accuracy of the M3GNet machine learning potential for modeling polymerization reactions by fine-tuning training parameters, data, and strategies within ParAMS. After reviewing all contributions, we are pleased to announce Javier Domínguez as the winner.
Congratulations, Javier—and thank you to all participants for your enthusiasm and contributions!
Interview with Javier
Q: Who are you?
I’m a third-year PhD student at the Donostia International Physics Center (DIPC) in San Sebastian, Spain. I work with Dr. Eduard Matito on developing new approaches to density functional theory.
Q: What is your research about?
I’m focused on creating better density functional approximations using model systems, specifically the harmonium atom. Essentially, I’m trying to build new functional components by taking advantage of exact solutions from these well-understood models.
The goal is to tackle some persistent issues in current density functional approximations, particularly how they handle different types of electron correlation. The harmonium atom works well for this because we can tune its confinement parameter to explore various correlation scenarios, which helps us develop more reliable functionals.
Q: What do you think about ParAMS?
I was pleasantly surprised by how user-friendly it is. The interface makes what could be a complex parametrization process quite manageable. I’m particularly interested in its capabilities for parametrizing DFT and DFTB methods.
I think it has real potential to be valuable for the computational chemistry community. I may incorporate it into my workflow once I reach the parametrization phase of my research!
Thank you, Javier, and good luck with your research!
