English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The ability to generate privacy-preserving synthetic versions of sensitiveimage datasets could unlock numerous ML applications currently constrained bydata availability. Due to their astonishing image generation quality, diffusionmodels are a prime candidate for generating high-quality synthetic data.However, recent studies have found that, by default, the outputs of somediffusion models do not preserve training data privacy. By privatelyfine-tuning ImageNet pre-trained diffusion models with more than 80Mparameters, we obtain SOTA results on CIFAR-10 and Camelyon17 in terms of bothFID and the accuracy of downstream classifiers trained on synthetic data. Wedecrease the SOTA FID on CIFAR-10 from 26.2 to 9.8, and increase the accuracyfrom 51.0% to 88.0%. On synthetic data from Camelyon17, we achieve a downstreamaccuracy of 91.1% which is close to the SOTA of 96.5% when training on the realdata. We leverage the ability of generative models to create infinite amountsof data to maximise the downstream prediction performance, and further show howto use synthetic data for hyperparameter tuning. Our results demonstrate thatdiffusion models fine-tuned with differential privacy can produce useful andprovably private synthetic data, even in applications with significantdistribution shift between the pre-training and fine-tuning distributions.

Differentially Private Diffusion Models Generate Useful Synthetic Images