Project Heads
Frank Noé, Cecilia Clementi
Project Members
Justin Kowalski
Project Duration
01.10.2026 − 30.09.2029
Located at
FU Berlin
We propose to develop a mathematical and machine-learning (ML) framework for a transferable machine-learned force field which is simultaneously able to generate its equilibrium distribution via a denoising diffusion-model process. If successful, the method will combine recent breakthroughs in the ML for the electronic structure problem and the sampling problem and enable new applications in drug discovery and materials design.
External Website
Related Publications
Related Pictures