Huziel E. Sauceda

TITLE:

Machine Learning in condensed matter: from molecular systems to materials

ABSTRACT:

Machine learning (ML) encompasses a wide range of algorithms and models, which have been prominently applied to condensed matter physics. Some applications range from atomistic simulations, generative quantum and classical distributions, physicochemical properties learning, among many others. In this talk, we will present some examples of how ML models have advanced our understanding of molecular systems and their complex interactions. In particular, we will focus on how combining machine learned force fields (MLFFs) and quantum dynamics, reveals the intricate nature of molecular systems and materials, as well as evince the limitations of many electronic structure methods. Furthermore, we will show how MLFFs enable the study of materials under realistic simulation conditions to generate predictive observables comparable with experimental results.

BIO:

Prof. Sauceda earned his Ph.D. in physics from the Instituto de Física, UNAM (IFUNAM), specializing in computational nanoscience. He then held a postdoctoral appointment at the Fritz Haber Institute of the Max Planck Society (Berlin), where he developed and applied machine-learning models for materials and molecular simulations. Afterward, he joined the Machine Learning and Big Data group at the Technical University of Berlin, serving as group leader until returning to IFUNAM in late 2021.
He now directs the Machine Learning for Simulations group at IFUNAM, mentoring undergraduate and graduate researchers. His team’s interests include: machine-learned interatomic potentials; learning classical and quantum propagators; quantum molecular dynamics; machine learning for many-body quantum systems; and battery physics. Across these areas, the group emphasizes rigorous benchmarking, open, reproducible workflows, and the integration of domain knowledge with modern data-driven methods.