pith. sign in

arxiv: 2210.09929 · v3 · pith:L3PATTCWnew · submitted 2022-10-18 · 📊 stat.ML · cs.CR· cs.LG

Differentially Private Diffusion Models

Tim Dockhorn , Tianshi Cao , Arash Vahdat , Karsten Kreis This is my paper
classification 📊 stat.ML cs.CRcs.LG
keywords modelsdatadifferentiallydiffusiondpdmsprivatebenchmarksclassifiers
0
0 comments X
read the original abstract

While modern machine learning models rely on increasingly large training datasets, data is often limited in privacy-sensitive domains. Generative models trained with differential privacy (DP) on sensitive data can sidestep this challenge, providing access to synthetic data instead. We build on the recent success of diffusion models (DMs) and introduce Differentially Private Diffusion Models (DPDMs), which enforce privacy using differentially private stochastic gradient descent (DP-SGD). We investigate the DM parameterization and the sampling algorithm, which turn out to be crucial ingredients in DPDMs, and propose noise multiplicity, a powerful modification of DP-SGD tailored to the training of DMs. We validate our novel DPDMs on image generation benchmarks and achieve state-of-the-art performance in all experiments. Moreover, on standard benchmarks, classifiers trained on DPDM-generated synthetic data perform on par with task-specific DP-SGD-trained classifiers, which has not been demonstrated before for DP generative models. Project page and code: https://nv-tlabs.github.io/DPDM.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. PrivCode++: Latent-Conditioned Differentially Private Code Generation for Comprehensive Guarantees

    cs.CR 2026-06 unverdicted novelty 7.0

    PrivCode++ introduces the first DP code generation method protecting both prompts and code via latent-conditioned two-stage training, claiming higher utility and stronger privacy than prior baselines.

  2. Filtering Memorization from Parameter-Space in Diffusion Models

    cs.CV 2026-05 unverdicted novelty 6.0

    BAF reduces memorization in diffusion LoRAs by filtering spectral channels of the adaptation weights that show weak alignment with the base model's principal subspace.

  3. Fundamental Limitations of Favorable Privacy-Utility Guarantees for DP-SGD

    cs.LG 2026-01 unverdicted novelty 6.0

    Shuffled DP-SGD requires σ ≥ 1/√(2 ln M) or κ ≥ (1/√8)(1 - 1/√(4π ln M)) to limit adversarial advantage, preventing strong privacy and high utility simultaneously.

  4. eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers

    cs.CV 2022-11 unverdicted novelty 6.0

    An ensemble of stage-specialized text-to-image diffusion models improves prompt alignment over single shared-parameter models while preserving visual quality and inference speed.