Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
DY: Fachverband Dynamik und Statistische Physik
DY 33: Machine Learning in Dynamics and Statistical Physics I
DY 33.5: Vortrag
Donnerstag, 20. März 2025, 10:30–10:45, H47
Efficient mapping of phase diagrams with conditional Boltzmann Generators — •Maximilian Schebek1, Michele Invernizzi2, Frank Noé1,2,3,4, and Jutta Rogal1,5 — 1Fachbereich Physik, Freie Universität Berlin, 14195 Berlin — 2Fachbereich Mathematik und Informatik, Freie Universität Berlin, 14195 Berlin — 3Department of Chemistry, Rice University, Houston, 77005, Texas, USA — 4AI4Science, Microsoft Research, 10178 Berlin — 5Department of Chemistry, New York University, New York, NY 10003, USA
The accurate prediction of phase diagrams is of central importance for both fundamental and applied material sciences. However, the computational prediction of the relative stability between phases based on their free energy is a daunting task, as traditional free energy estimators require a large amount of uncorrelated equilibrium samples over a grid of thermodynamic states. In this work, we develop deep generative machine learning models based on the Boltzmann Generator approach for entire phase diagrams, employing normalizing flows conditioned on the thermodynamic states that they map to. By training a single model to transform the equilibrium distribution sampled at only one reference thermodynamic state to a wide range of target temperatures and pressures, we can efficiently generate equilibrium samples across the entire phase diagram. We demonstrate our approach by predicting the solid-liquid coexistence line for a Lennard-Jones system in excellent agreement with state-of-the-art free energy methods while significantly reducing the number of energy evaluations needed.
Keywords: Sampling; Generative modelling; Free energies; Statistical Mechanics; Machine learning