Feb 27, 2023 · Our results demonstrate that diffusion models fine-tuned with differential privacy can produce useful and provably private synthetic data, even ...

The ability to generate privacy-preserving synthetic versions of sensitive image datasets could unlock numerous ML applications currently constrained by ...

The results demonstrate that diffusion modelsfine-tuned with differential privacy can produce useful and provably private synthetic data, ...

Generative models trained with differential privacy (DP) on sensitive data can sidestep this challenge, providing access to synthetic data instead. However, ...

We provide pre-trained checkpoints for all models presented in the paper here. You can sample from the checkpoint using the following command.

Our results demonstrate that diffusion models fine-tuned with differential privacy can produce useful and provably private synthetic data, even in applications ...

Missing: Generate | Show results with:Generate

Differentially Private Diffusion Models Generate Useful Synthetic Images (Ghalebikesabi et al., 2023); PPSyn | Partition-based differentially private ...

Discover how fine-tuning ImageNet pre-trained diffusion models with over 80M parameters achieves state-of-the-art results on CIFAR-10 and Camelyon17 datasets, ...