mlsanityData to 7 Oct 2026

Should I use Synthetic Data for That? An Analysis of the Suitability of Synthetic Data for Data Sharing and Augmentation

arXiv 2602.03791 ↗First posted 3 Feb 2026cs.LG cs.CYcited by 0

Not on the map. Why?

Abstract (arXiv)

Recent advances in generative modelling have led many to see synthetic data as the go-to solution for a range of problems around data access, scarcity, and under-representation. In this paper, we study three prominent use cases: (1) Sharing synthetic data as a proxy for proprietary datasets to enable statistical analyses while protecting privacy, (2) Augmenting machine learning training sets with synthetic data to improve model performance, and (3) Augmenting datasets with synthetic data to reduce variance in statistical estimation. For each use case, we formalise the problem setting and study, through formal analysis and case studies, under which conditions synthetic data can achieve its intended objectives. We identify fundamental and practical limits that constrain when synthetic data can serve as an effective solution for a particular problem. Our analysis reveals that due to these limits many existing or envisioned use cases of synthetic data are a poor problem fit. Our formalisations and classification of synthetic data use cases enable decision makers to assess whether synthetic data is a suitable approach for their specific data availability problem.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 36 of its 84 references that are in the map. First posted on arXiv 3 Feb 2026.

Builds on

36 of 84 references.

References · 36

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,592
  3. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,322
  4. 2018Confounding variables can degrade generalization performance of radiological deep learning modelscited by 374
  5. 2017Data Augmentation Generative Adversarial Networkscited by 278
  6. 2023Can Large Language Models Transform Computational Social Science?cited by 245
  7. 2023Prediction-Powered Inferencecited by 214
  8. 2021Simple data balancing achieves competitive worst-group-accuracycited by 177
  9. 2020Synthetic Data -- Anonymisation Groundhog Daycited by 105
  10. 2023PPI++: Efficient Prediction-Powered Inferencecited by 103
  11. 2021Winning the NIST Contest: A scalable and general approach to differentially private synthetic datacited by 101
  12. 2023Generative models improve fairness of medical classifiers under distribution shiftscited by 90
24 more
  1. 2022AIM: An Adaptive and Iterative Mechanism for Differentially Private Synthetic Datacited by 88
  2. 2023Differentially Private Diffusion Models Generate Useful Synthetic Imagescited by 85
  3. 2019Fair Decision Making using Privacy-Protected Datacited by 58
  4. 2024Fairness Feedback Loops: Training on Synthetic Data Amplifies Biascited by 55
  5. 2024Can Unconfident LLM Annotations Be Used for Confident Conclusions?cited by 44
  6. 2022Geographic Spines in the 2020 Census Disclosure Avoidance Systemcited by 42
  7. 2023Using Imperfect Surrogates for Downstream Inference: Design-based Supervised Learning for Social Science Applications of Large Language Modelscited by 37
  8. 2022A Comprehensive Empirical Study of Bias Mitigation Methods for Machine Learning Classifierscited by 30
  9. 2023Beyond Privacy: Navigating the Opportunities and Challenges of Synthetic Datacited by 29
  10. 2024Real Risks of Fake Data: Synthetic Data, Diversity-Washing and Consent Circumventioncited by 21
  11. 2021Conditional Synthetic Data Generation for Robust Machine Learning Applications with Limited Pandemic Datacited by 18
  12. 2024Differentially Private Release of Israel's National Registry of Live Birthscited by 16
  13. 2020Increasing the efficiency of randomized trial estimates via linear adjustment for a prognostic scorecited by 14
  14. 2022Conditional Generation of Medical Time Series for Extrapolation to Underrepresented Populationscited by 14
  15. 2023The Inadequacy of Similarity-based Privacy Metrics: Privacy Attacks against "Truly Anonymous" Synthetic Datasetscited by 14
  16. 2023On the Usefulness of Synthetic Tabular Data Generationcited by 12
  17. 2022Effect of Balancing Data Using Synthetic Data on the Performance of Machine Learning Classifiers for Intrusion Detection in Computer Networkscited by 11
  18. 2025Unifying Re-Identification, Attribute Inference, and Data Reconstruction Risks in Differential Privacycited by 11
  19. 2025How to DP-fy Your Data: A Practical Guide to Generating Synthetic Data With Differential Privacycited by 11
  20. 2025Examining the Expanding Role of Synthetic Data Throughout the AI Development Pipelinecited by 8
  21. 2025Valid Inference with Imperfect Synthetic Datacited by 7
  22. 2024Bounding the Excess Risk for Linear Models Trained on Marginal-Preserving, Differentially-Private, Synthetic Datacited by 3
  23. 2025Synthetic Data Generation for Augmenting Small Samplescited by 3
  24. 2024Lomas: A Platform for Confidential Analysis of Private Datacited by 1

Built on by

Not cited yet.