REVIEW 9 cited by
Differentially Private Diffusion Models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
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.
Forward citations
Cited by 9 Pith papers
-
PrivCode++: Latent-Conditioned Differentially Private Code Generation for Comprehensive Guarantees
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.
-
Filtering Memorization from Parameter-Space in Diffusion Models
Base-Anchored Filtering suppresses weakly backbone-aligned LoRA spectral channels to cut memorization while preserving or improving generation quality, without data or re-training.
-
Generalization and Memorization in Rectified Flow
Rectified Flow models peak in membership-inference vulnerability at the flow midpoint under uniform training; U-shaped timestep sampling suppresses memorization without harming FID.
-
IDDM: Identity-Decoupled Personalized Diffusion Models with a Tunable Privacy-Utility Trade-off
IDDM immunizes authorized personalized diffusion models so public generations remain high-quality while identity linkability to face recognizers is reduced with a tunable privacy-utility knob.
-
Filtering Memorization from Parameter-Space in Diffusion Models
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.
-
Fundamental Limitations of Favorable Privacy-Utility Guarantees for DP-SGD
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.
-
Privacy-Utility Trade-off in Data Publication: A Bilevel Optimization Framework with Curvature-Guided Perturbation
A bilevel optimization framework with curvature-guided geodesic perturbation reduces membership inference attack success while preserving downstream classification accuracy and sample quality.
-
Improving Noise Efficiency in Privacy-preserving Dataset Distillation
Dosser improves differentially private dataset distillation by decoupling sampling from optimization and projecting signals into a learned subspace.
-
eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers
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.
Discussion (0). Sign in to comment.