mlsanityData to 14 Oct 2026

Towards Reliable and Generalizable Differentially Private Machine Learning (Extended Version)

arXiv 2508.15141 ↗First posted 21 Aug 2025cs.LG cs.CRcited by 0

Not on the map. Why?

Abstract (arXiv)

There is a flurry of recent research papers proposing novel differentially private machine learning (DPML) techniques. These papers claim to achieve new state-of-the-art (SoTA) results and offer empirical results as validation. However, there is no consensus on which techniques are most effective or if they genuinely meet their stated claims. Complicating matters, heterogeneity in codebases, datasets, methodologies, and model architectures make direct comparisons of different approaches challenging. In this paper, we conduct a reproducibility and replicability (R+R) experiment on 11 different SoTA DPML techniques from the recent research literature. Results of our investigation are varied: while some methods stand up to scrutiny, others falter when tested outside their initial experimental conditions. We also discuss challenges unique to the reproducibility of DPML, including additional randomness due to DP noise, and how to address them. Finally, we derive insights and best practices to obtain scientifically valid and reliable results.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 118 of its 196 references that are in the map. First posted on arXiv 21 Aug 2025.

Builds on

118 of 196 references.

References · 118

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
  4. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,349
  5. 2016Deep Learning with Differential Privacycited by 2,728
  6. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,281
  7. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,496
  8. 2020Sharpness-Aware Minimization for Efficiently Improving Generalizationcited by 1,417
  9. 2021Membership Inference Attacks From First Principlescited by 878
  10. 2017Deep Reinforcement Learning that Matterscited by 754
  11. 2019Large Batch Optimization for Deep Learning: Training BERT in 76 minutescited by 674
  12. 2021Large Language Models Can Be Strong Differentially Private Learnerscited by 412
38 more
  1. 2021Differentially Private Fine-tuning of Language Modelscited by 400
  2. 2023A Survey on Large Language Model (LLM) Security and Privacy: The Good, the Bad, and the Uglycited by 387
  3. 2016Semi-supervised Knowledge Transfer for Deep Learning from Private Training Datacited by 348
  4. 2016Concentrated Differential Privacy: Simplifications, Extensions, and Lower Boundscited by 346
  5. 2020Improving Reproducibility in Machine Learning Research (A Report from the NeurIPS 2019 Reproducibility Program)cited by 322
  6. 2018Scalable Private Learning with PATEcited by 269
  7. 2021Enhanced Membership Inference Attacks against Machine Learning Modelscited by 248
  8. 2018Subsampled R'enyi Differential Privacy and Analytical Moments Accountantcited by 236
  9. 2020Differentially Private Learning Needs Better Features (or Much More Data)cited by 235
  10. 2019Differentially Private Learning with Adaptive Clippingcited by 220
  11. 2020Auditing Differentially Private Machine Learning: How Private is Private SGD?cited by 218
  12. 2019SGD: General Analysis and Improved Ratescited by 207
  13. 2019R'enyi Differential Privacy of the Sampled Gaussian Mechanismcited by 204
  14. 2021Adversary Instantiation: Lower Bounds for Differentially Private Machine Learningcited by 199
  15. 2024Do Membership Inference Attacks Work on Large Language Models?cited by 195
  16. 2021Numerical Composition of Differential Privacycited by 180
  17. 2023How to DP-fy ML: A Practical Guide to Machine Learning with Differential Privacycited by 153
  18. 2020Understanding Gradient Clipping in Private SGD: A Geometric Perspectivecited by 142
  19. 2023Privacy Auditing with One (1) Training Runcited by 123
  20. 2020Tempered Sigmoid Activations for Deep Learning with Differential Privacycited by 122
  21. 2021Hyperparameter Tuning with Renyi Differential Privacycited by 104
  22. 2021Learning to See by Looking at Noisecited by 92
  23. 2021Large Scale Private Learning via Low-rank Reparametrizationcited by 88
  24. 2022Automatic Clipping: Differentially Private Deep Learning Made Easier and Strongercited by 88
  25. 2022Position: Considerations for Differentially Private Learning with Large-Scale Public Pretrainingcited by 88
  26. 2022Toward Training at ImageNet Scale with Differential Privacycited by 87
  27. 2021Do Not Let Privacy Overbill Utility: Gradient Embedding Perturbation for Private Learningcited by 86
  28. 2021Private Stochastic Convex Optimization: Optimal Rates in ℓ1 Geometrycited by 83
  29. 2023Revisiting Gradient Clipping: Stochastic bias and tight convergence guaranteescited by 83
  30. 2021Torch.manual_seed(3407) is all you need: On the influence of random seeds in deep learning architectures for computer visioncited by 81
  31. 2019AdaCliP: Adaptive Clipping for Private SGDcited by 77
  32. 2019A Step Toward Quantifying Independently Reproducible Machine Learning Researchcited by 77
  33. 2021Differential Privacy Dynamics of Langevin Diffusion and Noisy Gradient Descentcited by 73
  34. 2022Differentially Private Optimization on Large Model at Small Costcited by 65
  35. 2024SoK: Membership Inference Attacks on LLMs are Rushing Nowhere (and How to Fix It)cited by 59
  36. 2020Enabling Fast Differentially Private SGD via Just-in-Time Compilation and Vectorizationcited by 58
  37. 2022Mixed Differential Privacy in Computer Visioncited by 54
  38. 2022TAN Without a Burn: Scaling Laws of DP-SGDcited by 52

50 most cited of 118 All 118 (JSON)

Built on by

Not cited yet.