mlsanityData to 7 Oct 2026

PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning

arXiv 2106.04590 ↗First posted 8 Jun 2021cs.LG cs.CRcited by 23

Not on the map. Why?

Abstract (arXiv)

We propose a new framework of synthesizing data using deep generative models in a differentially private manner. Within our framework, sensitive data are sanitized with rigorous privacy guarantees in a one-shot fashion, such that training deep generative models is possible without re-using the original data. Hence, no extra privacy costs or model constraints are incurred, in contrast to popular approaches such as Differentially Private Stochastic Gradient Descent (DP-SGD), which, among other issues, causes degradation in privacy guarantees as the training iteration increases. We demonstrate a realization of our framework by making use of the characteristic function and an adversarial re-weighting objective, which are of independent interest as well. Our proposal has theoretical guarantees of performance, and empirical evaluations on multiple datasets show that our approach outperforms other methods at reasonable levels of privacy.

Key facts. Cited by 23 papers in the map (data through 7 Oct 2026). Builds on 25 of its 58 references that are in the map. First posted on arXiv 8 Jun 2021.

Builds on

25 of 58 references.

References · 25

  1. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,283
  2. 2016Deep Learning with Differential Privacycited by 2,682
  3. 2014Conditional Generative Adversarial Netscited by 2,152
  4. 2018Demystifying MMD GANscited by 1,268
  5. 2019Modeling Tabular data using Conditional GANcited by 695
  6. 2017Renyi Differential Privacycited by 643
  7. 2016Semi-supervised Knowledge Transfer for Deep Learning from Private Training Datacited by 347
  8. 2019Differential Privacy Has Disparate Impact on Model Accuracycited by 316
  9. 2017MMD GAN: Towards Deeper Understanding of Moment Matching Networkcited by 251
  10. 2018Differentially Private Generative Adversarial Networkcited by 235
  11. 2018Subsampled R'enyi Differential Privacy and Analytical Moments Accountantcited by 234
  12. 2020Auditing Differentially Private Machine Learning: How Private is Private SGD?cited by 213
13 more
  1. 2021Adversary Instantiation: Lower Bounds for Differentially Private Machine Learningcited by 193
  2. 2020DP-CGAN: Differentially Private Synthetic Data and Label Generationcited by 118
  3. 2020GS-WGAN: A Gradient-Sanitized Approach for Learning Differentially Private Generatorscited by 117
  4. 2017The Cramer Distance as a Solution to Biased Wasserstein Gradientscited by 116
  5. 2020DP-MERF: Differentially Private Mean Embeddings with Random Features for Practical Privacy-Preserving Data Generationcited by 78
  6. 2016Interpretable Distribution Features with Maximum Testing Powercited by 63
  7. 2015Fast Two-Sample Testing with Analytic Representations of Probability Measurescited by 53
  8. 2019Differentially Private Generative Adversarial Networks for Time Series, Continuous, and Discrete Open Datacited by 31
  9. 2019A Characteristic Function Approach to Deep Implicit Generative Modelingcited by 25
  10. 2017McGan: Mean and Covariance Feature Matching GANcited by 21
  11. 2020P3GM: Private High-Dimensional Data Release via Privacy Preserving Phased Generative Modelcited by 21
  12. 2020Reciprocal Adversarial Learning via Characteristic Functionscited by 7
  13. 2020Differentially Private Variational Autoencoders with Term-wise Gradient Aggregationcited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 6; 2023: 7; 2024: 4; 2025: 5; 2026: 1.

23 papers.

Sort:

By area:

  1. 2022Differentially Private Diffusion Modelscited by 118
  2. 2023Differentially Private Diffusion Models Generate Useful Synthetic Imagescited by 85
  3. 2022Pre-trained Perceptual Features Improve Differentially Private Image Generationcited by 28
  4. 2023DP-LDMs: Differentially Private Latent Diffusion Modelscited by 28
  5. 2023Differentially Private Neural Tangent Kernels for Privacy-Preserving Data Generationcited by 15
  6. 2023PrivImage: Differentially Private Synthetic Image Generation using Diffusion Models with Semantic-Aware Pretrainingcited by 12
  7. 2022Noise-Aware Statistical Inference with Differentially Private Synthetic Datacited by 9
  8. 2022On the Utility Recovery Incapability of Neural Net-based Differential Private Tabular Training Data Synthesizer under Privacy Deregulationcited by 8
  9. 2024Tabular Data Synthesis with Differential Privacy: A Surveycited by 6
  10. 2025From Easy to Hard: Building a Shortcut for Differentially Private Image Synthesiscited by 6
  11. 2022Transfer Learning with Pre-trained Conditional Generative Modelscited by 4
  12. 2023A Unified View of Differentially Private Deep Generative Modelingcited by 4
11 more
  1. 2024PFGuard: A Generative Framework with Privacy and Fairness Safeguardscited by 3
  2. 2022NeuroMixGDP: A Neural Collapse-Inspired Random Mixup for Private Data Releasecited by 2
  3. 2025RAPID: Retrieval Augmented Training of Differentially Private Diffusion Modelscited by 2
  4. 2023A self-attention-based differentially private tabular GAN with high data utilitycited by 1
  5. 2024Efficient Differentially Private Fine-Tuning of Diffusion Modelscited by 1
  6. 2024Leveraging Programmatically Generated Synthetic Data for Differentially Private Diffusion Trainingcited by 1
  7. 2025DP-DocLDM: Differentially Private Document Image Generation using Latent Diffusion Modelscited by 1
  8. 2025PrivORL: Differentially Private Synthetic Dataset for Offline Reinforcement Learningcited by 1
  9. 2023DPAF: Image Synthesis via Differentially Private Aggregation in Forward Phasecited by 0
  10. 2025VP-NTK: Exploring the Benefits of Visual Prompting in Differentially Private Data Synthesiscited by 0
  11. 2026From Easy to Hard++: Promoting Differentially Private Image Synthesis Through Spatial-Frequency Curriculumcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Learning Differentially Private Diffusion Models via Stochastic Adversarial Distillation
  2. 2026Protecting Sensitive Data in Image Synthesis via PAC-Private Adaptation for Diffusion Models
  3. 2024DP-TLDM: Differentially Private Tabular Latent Diffusion Model
  4. 2026StraightDP: Geometry-Aware Differential Privacy for Rectified-Flow Transformers
  5. 2025Differentially Private Synthetic Data via APIs 3: Using Simulators Instead of Foundation Model
  6. 2026Escaping Iterative Parameter-Space Noise: Differentially Private Learning with a Hypernetwork
  7. 2021Causally Constrained Data Synthesis for Private Data Release
  8. 2025On the MIA Vulnerability Gap Between Private GANs and Diffusion Models