Ambient Diffusion: Learning Clean Distributions from Corrupted Data

arXiv2305.19256MaRDI QIDQ6438543

Author name not available (Why is that?)

Publication date: 30 May 2023

Abstract: We present the first diffusion-based framework that can learn an unknown distribution using only highly-corrupted samples. This problem arises in scientific applications where access to uncorrupted samples is impossible or expensive to acquire. Another benefit of our approach is the ability to train generative models that are less likely to memorize individual training samples since they never observe clean training data. Our main idea is to introduce additional measurement distortion during the diffusion process and require the model to predict the original corrupted image from the further corrupted image. We prove that our method leads to models that learn the conditional expectation of the full uncorrupted image given this additional measurement corruption. This holds for any corruption process that satisfies some technical conditions (and in particular includes inpainting and compressed sensing). We train models on standard benchmarks (CelebA, CIFAR-10 and AFHQ) and show that we can learn the distribution even when all the training samples have $90\mathrm{\%}$ of their pixels missing. We also show that we can finetune foundation models on small corrupted datasets (e.g. MRI scans with block corruptions) and learn the clean distribution without memorizing the training set.

Has companion code repository: https://github.com/giannisdaras/ambient-diffusion

This page was built for publication: Ambient Diffusion: Learning Clean Distributions from Corrupted Data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6438543)