Just another WordPress site - Ruhr-Universität Bochum

M.Sc. Eslam Ibrahim

Doctoral Candidate

Department:

Group: Data-Driven Methods for Atomistic Simulations

ICAMS

Contact


Ruhr-Universität Bochum
IC 02-729
Universitätsstr. 150
44801 Bochum

Tel.: +49 234 32 22775
E-Mail: eslam.saadibrahim@rub.de

Research

Artistic t-SNE maps of ice phases and water configurations at 300 K. Cover art of the issue created by Eslam Ibrahim.
ICAMS 2025

Here we show the significance of our work, highlighting the power of ACE descriptors. By projecting the ACE descriptors in two dimensions using nonlinear dimensionality reductions, implemented using TSNE, we could show the structure similarity between diverse ice and liquid phases. This allowed us to efficiently start fitting a potential from ice phases based on solely DFT static calculations, which is orders of magnitude faster and more efficient than running ab-initio molecular dynamics (AIMD) for each phase at different temperatures and densities. Later, using active learning to select a few liquid configurations, we had evidence that described the liquid in excellent agreement with AIMD for the reference DFT functional. View the article. https://pubs.acs.org/doi/10.1021/acs.jctc.4c00802

Competences

machine learning potentials,
water and water interfaces,
ice,
local atomic environments,
ab initio molecular dynamics

Publications
Talks and Posters
Short CV

Professional Experience

Research visiting student at KAUST, CPMS group
Since 2022: Fellow at IMPRS-SusMet
Since 2022: PhD candidate at ICAMS, AMS department
2024 Guest researcher at CIC nanoGUNE, Donostia / San Sebastián, Spain

Education

2022. M.Sc., Materials Science and Simulation, Ruhr-Universität Bochum

back