Salvato in:
Dettagli Bibliografici
Autori principali: Bond-Taylor, Sam, Willcocks, Chris G.
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:https://arxiv.org/abs/2303.18242
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909124202070016
author Bond-Taylor, Sam
Willcocks, Chris G.
author_facet Bond-Taylor, Sam
Willcocks, Chris G.
contents This paper introduces $\infty$-Diff, a generative diffusion model defined in an infinite-dimensional Hilbert space, which can model infinite resolution data. By training on randomly sampled subsets of coordinates and denoising content only at those locations, we learn a continuous function for arbitrary resolution sampling. Unlike prior neural field-based infinite-dimensional models, which use point-wise functions requiring latent compression, our method employs non-local integral operators to map between Hilbert spaces, allowing spatial context aggregation. This is achieved with an efficient multi-scale function-space architecture that operates directly on raw sparse coordinates, coupled with a mollified diffusion process that smooths out irregularities. Through experiments on high-resolution datasets, we found that even at an $8\times$ subsampling rate, our model retains high-quality diffusion. This leads to significant run-time and memory savings, delivers samples with lower FID scores, and scales beyond the training resolution while retaining detail.
format Preprint
id arxiv_https___arxiv_org_abs_2303_18242
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle $\infty$-Diff: Infinite Resolution Diffusion with Subsampled Mollified States
Bond-Taylor, Sam
Willcocks, Chris G.
Machine Learning
Computer Vision and Pattern Recognition
This paper introduces $\infty$-Diff, a generative diffusion model defined in an infinite-dimensional Hilbert space, which can model infinite resolution data. By training on randomly sampled subsets of coordinates and denoising content only at those locations, we learn a continuous function for arbitrary resolution sampling. Unlike prior neural field-based infinite-dimensional models, which use point-wise functions requiring latent compression, our method employs non-local integral operators to map between Hilbert spaces, allowing spatial context aggregation. This is achieved with an efficient multi-scale function-space architecture that operates directly on raw sparse coordinates, coupled with a mollified diffusion process that smooths out irregularities. Through experiments on high-resolution datasets, we found that even at an $8\times$ subsampling rate, our model retains high-quality diffusion. This leads to significant run-time and memory savings, delivers samples with lower FID scores, and scales beyond the training resolution while retaining detail.
title $\infty$-Diff: Infinite Resolution Diffusion with Subsampled Mollified States
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2303.18242