Saved in:
Bibliographic Details
Main Authors: Williams, Christopher, Campbell, Andrew, Doucet, Arnaud, Syed, Saifuddin
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.07877
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912151845732352
author Williams, Christopher
Campbell, Andrew
Doucet, Arnaud
Syed, Saifuddin
author_facet Williams, Christopher
Campbell, Andrew
Doucet, Arnaud
Syed, Saifuddin
contents Denoising diffusion models (DDMs) offer a flexible framework for sampling from high dimensional data distributions. DDMs generate a path of probability distributions interpolating between a reference Gaussian distribution and a data distribution by incrementally injecting noise into the data. To numerically simulate the sampling process, a discretisation schedule from the reference back towards clean data must be chosen. An appropriate discretisation schedule is crucial to obtain high quality samples. However, beyond hand crafted heuristics, a general method for choosing this schedule remains elusive. This paper presents a novel algorithm for adaptively selecting an optimal discretisation schedule with respect to a cost that we derive. Our cost measures the work done by the simulation procedure to transport samples from one point in the diffusion path to the next. Our method does not require hyperparameter tuning and adapts to the dynamics and geometry of the diffusion path. Our algorithm only involves the evaluation of the estimated Stein score, making it scalable to existing pre-trained models at inference time and online during training. We find that our learned schedule recovers performant schedules previously only discovered through manual search and obtains competitive FID scores on image datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2412_07877
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Score-Optimal Diffusion Schedules
Williams, Christopher
Campbell, Andrew
Doucet, Arnaud
Syed, Saifuddin
Machine Learning
Denoising diffusion models (DDMs) offer a flexible framework for sampling from high dimensional data distributions. DDMs generate a path of probability distributions interpolating between a reference Gaussian distribution and a data distribution by incrementally injecting noise into the data. To numerically simulate the sampling process, a discretisation schedule from the reference back towards clean data must be chosen. An appropriate discretisation schedule is crucial to obtain high quality samples. However, beyond hand crafted heuristics, a general method for choosing this schedule remains elusive. This paper presents a novel algorithm for adaptively selecting an optimal discretisation schedule with respect to a cost that we derive. Our cost measures the work done by the simulation procedure to transport samples from one point in the diffusion path to the next. Our method does not require hyperparameter tuning and adapts to the dynamics and geometry of the diffusion path. Our algorithm only involves the evaluation of the estimated Stein score, making it scalable to existing pre-trained models at inference time and online during training. We find that our learned schedule recovers performant schedules previously only discovered through manual search and obtains competitive FID scores on image datasets.
title Score-Optimal Diffusion Schedules
topic Machine Learning
url https://arxiv.org/abs/2412.07877