Regensburg 2025 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
SYAI: Symposium AI in (Bio-)Physics
SYAI 1: AI in (Bio-)Physics
SYAI 1.2: Invited Talk
Thursday, March 20, 2025, 10:00–10:30, H1
Realizing Schrödinger's dream with AI-enabled molecular dynamics — •Alexandre Tkatchenko — Department of Physics and Materials Science, University of Luxembourg
The convergence between accurate quantum-mechanical (QM) models (and codes) with efficient machine learning (ML) methods seem to promise a paradigm shift in molecular simulations. Many challenging applications are now being tackled by increasingly powerful QM/ML methodologies. These include modeling covalent materials, molecules, molecular crystals, surfaces, and even whole proteins in explicit water (https://www.science.org/doi/abs/10.1126/sciadv.adn4397). In this talk, I attempt to provide a reality check on these recent advances.
In particular, I will introduce the recently developed SO3LR force field (https://doi.org/10.26434/chemrxiv-2024-bdfr0), trained on a diverse set of 4 million neutral and charged molecular complexes computed at the PBE0+MBD level of quantum mechanics, ensuring a comprehensive coverage of covalent and non-covalent interactions. SO3LR is characterized by computational and data efficiency, scalability to 200 thousand atoms on a single GPU, and reasonable to high accuracy across the chemical space of organic (bio)molecules. SO3LR is applied to study units of four major biomolecule types, polypeptide folding, and nanosecond dynamics of larger systems such as a protein, a glycoprotein, and a lipid bilayer, all in explicit solvent. Finally, I discuss the future challenges toward truly general molecular simulations by combining ML force fields with traditional atomistic models.
Keywords: Molecular dynamics; Machine learning; Force Fields; Quantum mechanics; Biophysics