Archival

Efficient sampling of molecular systems at thermodynamic equilibrium is a hallmark challenge in statistical physics. This challenge has …

Numerous problems in bioinformatics and computational biology can be framed as a task of learning a mapping from one state of a …

Predicting the intermediate trajectories between an initial and target distribution is a central problem in generative modeling. …

Simulation-free training frameworks have been at the forefront of the generative modelling revolution in continuous spaces, leading to …

Scalable sampling of molecular states in thermodynamic equilibrium is a long-standing challenge in statistical physics. Boltzmann …

Accurately predicting experimentally-realizable 3D molecular crystal structures from their 2D chemical graphs is a long-standing open …

Diffusion language models have emerged as a powerful alternative to autoregressive models, enabling fast inference through flexible and …

Sampling efficiently from a target unnormalized probability density remains a core challenge, with relevance across countless …