mlsanityData to 7 Oct 2026

SoK: Privacy-Preserving Data Synthesis

arXiv 2307.02106 ↗First posted 5 Jul 2023cs.CR cs.DB cs.LGcited by 37

Not on the map. Why?

Abstract (arXiv)

As the prevalence of data analysis grows, safeguarding data privacy has become a paramount concern. Consequently, there has been an upsurge in the development of mechanisms aimed at privacy-preserving data analyses. However, these approaches are task-specific; designing algorithms for new tasks is a cumbersome process. As an alternative, one can create synthetic data that is (ideally) devoid of private information. This paper focuses on privacy-preserving data synthesis (PPDS) by providing a comprehensive overview, analysis, and discussion of the field. Specifically, we put forth a master recipe that unifies two prominent strands of research in PPDS: statistical methods and deep learning (DL)-based methods. Under the master recipe, we further dissect the statistical methods into choices of modeling and representation, and investigate the DL-based methods by different generative modeling principles. To consolidate our findings, we provide comprehensive reference tables, distill key takeaways, and identify open problems in the existing literature. In doing so, we aim to answer the following questions: What are the design principles behind different PPDS methods? How can we categorize these methods, and what are the advantages and disadvantages associated with each category? Can we provide guidelines for method selection in different real-world scenarios? We proceed to benchmark several prominent DL-based methods on the task of private image synthesis and conclude that DP-MERF is an all-purpose approach. Finally, upon systematizing the work over the past decade, we identify future directions and call for actions from researchers.

Key facts. Cited by 37 papers in the map (data through 7 Oct 2026). Builds on 72 of its 254 references that are in the map. First posted on arXiv 5 Jul 2023.

Builds on

72 of 254 references.

References · 72

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  3. 2015Distilling the Knowledge in a Neural Networkcited by 10,076
  4. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  5. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,283
  6. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  7. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  8. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 2,994
  9. 2016Deep Learning with Differential Privacycited by 2,682
  10. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,221
  11. 2015LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loopcited by 1,350
  12. 2022Diffusion Models: A Comprehensive Survey of Methods and Applicationscited by 1,107
38 more
  1. 2020Improved Techniques for Training Score-Based Generative Modelscited by 1,036
  2. 2017A Closer Look at Memorization in Deep Networkscited by 936
  3. 2019Modeling Tabular data using Conditional GANcited by 695
  4. 2017Renyi Differential Privacycited by 643
  5. 2021Large Language Models Can Be Strong Differentially Private Learnerscited by 405
  6. 2018Video-to-Video Synthesiscited by 388
  7. 2016Semi-supervised Knowledge Transfer for Deep Learning from Private Training Datacited by 347
  8. 2017Learning Generative Models with Sinkhorn Divergencescited by 310
  9. 2018Deep Face Recognition: A Surveycited by 275
  10. 2018Scalable Private Learning with PATEcited by 266
  11. 2021Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Modelscited by 251
  12. 2019The Secret Revealer: Generative Model-Inversion Attacks Against Deep Neural Networkscited by 241
  13. 2018Differentially Private Generative Adversarial Networkcited by 235
  14. 2018Adversarial Audio Synthesiscited by 222
  15. 2021Adversary Instantiation: Lower Bounds for Differentially Private Machine Learningcited by 193
  16. 2018Protection Against Reconstruction and Its Applications in Private Federated Learningcited by 127
  17. 2018Improving GANs Using Optimal Transportcited by 122
  18. 2019Reconstruction and Membership Inference Attacks against Generative Modelscited by 122
  19. 2020DP-CGAN: Differentially Private Synthetic Data and Label Generationcited by 118
  20. 2022Differentially Private Diffusion Modelscited by 118
  21. 2020GS-WGAN: A Gradient-Sanitized Approach for Learning Differentially Private Generatorscited by 117
  22. 2020Synthetic Data -- Anonymisation Groundhog Daycited by 105
  23. 2021Winning the NIST Contest: A scalable and general approach to differentially private synthetic datacited by 101
  24. 2021Benchmarking Differentially Private Synthetic Data Generation Algorithmscited by 80
  25. 2018Privacy Preserving Machine Learning: Threats and Solutionscited by 78
  26. 2020DP-MERF: Differentially Private Mean Embeddings with Random Features for Practical Privacy-Preserving Data Generationcited by 78
  27. 2019Graphical-model based estimation and inference for differential privacycited by 73
  28. 2020When Machine Learning Meets Privacy: A Survey and Outlookcited by 67
  29. 2014Differential Privacy and Machine Learning: a Survey and Reviewcited by 66
  30. 2022Synthetic Data -- what, why and how?cited by 66
  31. 2021Don't Generate Me: Training Differentially Private Generative Models with Sinkhorn Divergencecited by 54
  32. 2019G-PATE: Scalable Differentially Private Data Generator via Private Aggregation of Teacher Discriminatorscited by 53
  33. 2020Differentially Private Synthetic Medical Data Generation using Convolutional GANscited by 52
  34. 2021Differentially Private Query Release Through Adaptive Projectioncited by 50
  35. 2021Robin Hood and Matthew Effects: Differential Privacy Has Disparate Impact on Synthetic Datacited by 48
  36. 2020New Oracle-Efficient Algorithms for Private Synthetic Data Releasecited by 43
  37. 2017Differentially Private Mixture of Generative Neural Networkscited by 39
  38. 2019ER-AE: Differentially Private Text Generation for Authorship Anonymizationcited by 35

50 most cited of 72 All 72 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 4; 2024: 8; 2025: 19; 2026: 6.

37 papers.

Sort:

By area:

  1. 2024Provably Unlearnable Data Examplescited by 18
  2. 2023Privately Aligning Language Models with Reinforcement Learningcited by 17
  3. 2025Differentially Private Synthetic Data via APIs 3: Using Simulators Instead of Foundation Modelcited by 13
  4. 2025DPImageBench: A Unified Benchmark for Differentially Private Image Synthesiscited by 11
  5. 2025The DCR Delusion: Measuring the Privacy Risk of Synthetic Datacited by 11
  6. 2024Privacy Vulnerabilities in Marginals-based Synthetic Datacited by 10
  7. 2024Privacy-preserving data release leveraging optimal transport and particle gradient descentcited by 8
  8. 2024LLM-based Privacy Data Augmentation Guided by Knowledge Distillation with a Distribution Tutor for Medical Text Classificationcited by 8
  9. 2025Struct-Bench: A Benchmark for Differentially Private Structured Text Generationcited by 7
  10. 2024Synthetic Data: Revisiting the Privacy-Utility Trade-offcited by 6
  11. 2023A Unified View of Differentially Private Deep Generative Modelingcited by 4
  12. 2025Generating Synthetic Data with Formal Privacy Guarantees: State of the Art and the Road Aheadcited by 4
25 more
  1. 2025The Importance of Being Discrete: Measuring the Impact of Discretization in End-to-End Differentially Private Synthetic Datacited by 4
  2. 2024Combining Machine Learning Defenses without Conflictscited by 2
  3. 2025RAPID: Retrieval Augmented Training of Differentially Private Diffusion Modelscited by 2
  4. 2025Private Text Generation by Seeding Large Language Model Promptscited by 2
  5. 2025dpmm: Differentially Private Marginal Models, a Library for Synthetic Tabular Data Generationcited by 2
  6. 2025Privacy Amplification Through Synthetic Data: Insights from Linear Regressioncited by 2
  7. 2025Virus Infection Attack on LLMs: Your Poisoning Can Spread "VIA" Synthetic Datacited by 2
  8. 2025ACTG-ARL: Differentially Private Conditional Text Generation with RL-Boosted Controlcited by 2
  9. 2026Differentially Private Synthetic Data via APIs 4: Tabular Datacited by 2
  10. 2025SoK: Can Synthetic Images Replace Real Data? A Survey of Utility and Privacy of Synthetic Image Generationcited by 1
  11. 2025Democratizing Tabular Data Access with an Openunicodex2013Source Syntheticunicodex2013Data SDKcited by 1
  12. 2025BinaryShield: Cross-Service Threat Intelligence in LLM Services using Privacy-Preserving Fingerprintscited by 1
  13. 2025PrivORL: Differentially Private Synthetic Dataset for Offline Reinforcement Learningcited by 1
  14. 2026Rethinking Anonymity Claims in Synthetic Data Generation: A Model-Centric Privacy Attack Perspectivecited by 1
  15. 2026FHAIM: Fully Homomorphic AIM For Private Synthetic Data Generationcited by 1
  16. 2026Turning Black Box into White Box: Dataset Distillation Leakscited by 1
  17. 2023Scalable and Privacy-Preserving Synthetic Data Generation on Decentralised Webcited by 0
  18. 2023PAC Privacy Preserving Diffusion Modelscited by 0
  19. 2024SAFES: Sequential Privacy and Fairness Enhancing Data Synthesis for Responsible AIcited by 0
  20. 2024End to End Collaborative Synthetic Data Generationcited by 0
  21. 2025Understanding the Impact of Data Domain Extraction on Synthetic Data Privacycited by 0
  22. 2025Private Training & Data Generation by Clustering Embeddingscited by 0
  23. 2025How to Protect Models against Adversarial Unlearning?cited by 0
  24. 2026From Easy to Hard++: Promoting Differentially Private Image Synthesis Through Spatial-Frequency Curriculumcited by 0
  25. 2026MAPLE: Metadata Augmented Private Language Evolutioncited by 0