no code implementations • 6 Feb 2024 • Christian Horvat, Jean-Pascal Pfister
In the original formulation of the diffusion model, this vector field is assumed to be the score function (i. e. it is the gradient of the log-probability at a given time in the diffusion process).
1 code implementation • NeurIPS 2021 • Christian Horvat, Jean-Pascal Pfister
Here we propose a novel method - called Denoising Normalizing Flow (DNF) - that estimates the density on the low-dimensional manifold while learning the manifold as well.
2 code implementations • 25 May 2021 • Christian Horvat, Jean-Pascal Pfister
We also show that, if the embedding dimension is much larger than the manifold dimension, noise in the normal space can be well approximated by Gaussian noise.