We introduce a measure of the information content stored in a neural network in the context of diffusion models.

A method for the prediction of molecular dynamics trajectories using long time steps

Molecular dynamics in smoothed noisy spaces to accelerate conformational ensemble generation