mlsanityData to 7 Oct 2026

Differentially Private Natural Language Models: Recent Advances and Future Directions

arXiv 2301.09112 ↗First posted 22 Jan 2023cs.CL cs.AI cs.LGcited by 20

Not on the map. Why?

Abstract (arXiv)

Recent developments in deep learning have led to great success in various natural language processing (NLP) tasks. However, these applications may involve data that contain sensitive information. Therefore, how to achieve good performance while also protecting the privacy of sensitive data is a crucial challenge in NLP. To preserve privacy, Differential Privacy (DP), which can prevent reconstruction attacks and protect against potential side knowledge, is becoming a de facto technique for private data analysis. In recent years, NLP in DP models (DP-NLP) has been studied from different perspectives, which deserves a comprehensive review. In this paper, we provide the first systematic review of recent advances in DP deep learning models in NLP. In particular, we first discuss some differences and additional challenges of DP-NLP compared with the standard DP deep learning. Then, we investigate some existing work on DP-NLP and present its recent developments from three aspects: gradient perturbation based methods, embedding vector perturbation based methods, and ensemble model based methods. We also discuss some challenges and future directions.

Key facts. Cited by 20 papers in the map (data through 7 Oct 2026). Builds on 92 of its 140 references that are in the map. First posted on arXiv 22 Jan 2023.

Builds on

92 of 140 references.

References · 92

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  4. 2013Efficient Estimation of Word Representations in Vector Spacecited by 4,346
  5. 2019Parameter-Efficient Transfer Learning for NLPcited by 4,195
  6. 2016Deep Learning with Differential Privacycited by 2,682
  7. 2018Deep contextualized word representationscited by 2,285
  8. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,221
  9. 2020Extracting Training Data from Large Language Modelscited by 2,144
  10. 2021BitFit: Simple Parameter-efficient Fine-tuning for Transformer-based Masked Language-modelscited by 1,234
  11. 2018The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networkscited by 847
  12. 2017Renyi Differential Privacycited by 643
38 more
  1. 2018Federated Learning for Mobile Keyboard Predictioncited by 565
  2. 2021Compacter: Efficient Low-Rank Hypercomplex Adapter Layerscited by 458
  3. 2021Large Language Models Can Be Strong Differentially Private Learnerscited by 405
  4. 2021Differentially Private Fine-tuning of Language Modelscited by 392
  5. 2016Semi-supervised Knowledge Transfer for Deep Learning from Private Training Datacited by 347
  6. 2016Concentrated Differential Privacy: Simplifications, Extensions, and Lower Boundscited by 339
  7. 2019Differential Privacy Has Disparate Impact on Model Accuracycited by 316
  8. 2013The Composition Theorem for Differential Privacycited by 290
  9. 2018Privacy Amplification by Subsampling: Tight Analyses via Couplings and Divergencescited by 237
  10. 2018Subsampled R'enyi Differential Privacy and Analytical Moments Accountantcited by 234
  11. 2019Differentially Private Learning with Adaptive Clippingcited by 218
  12. 2019R'enyi Differential Privacy of the Sampled Gaussian Mechanismcited by 199
  13. 2022What Does it Mean for a Language Model to Preserve Privacy?cited by 194
  14. 2021Numerical Composition of Differential Privacycited by 174
  15. 2021Practical and Private (Deep) Learning without Sampling or Shufflingcited by 169
  16. 2016Concentrated Differential Privacycited by 139
  17. 2021Large-Scale Differentially Private BERTcited by 117
  18. 2018Distributed Differential Privacy via Shufflingcited by 115
  19. 2019Privacy- and Utility-Preserving Textual Analysis via Calibrated Multivariate Perturbationscited by 115
  20. 2019Deep Learning with Gaussian Differential Privacycited by 112
  21. 2022Synthetic Text Generation with Differential Privacy: A Simple and Practical Recipecited by 108
  22. 2022The Text Anonymization Benchmark (TAB): A Dedicated Corpus and Evaluation Framework for Text Anonymizationcited by 94
  23. 2021Large Scale Private Learning via Low-rank Reparametrizationcited by 87
  24. 2021Differential Privacy for Text Analytics via Natural Text Sanitizationcited by 74
  25. 2019Differentially Private Model Publishing for Deep Learningcited by 69
  26. 2020Training Production Language Models without Memorizing User Datacited by 68
  27. 2014Differential Privacy and Machine Learning: a Survey and Reviewcited by 66
  28. 2020Differentially Private Representation for NLP: Formal Guarantee and An Empirical Study on Privacy and Fairnesscited by 66
  29. 2021The Skellam Mechanism for Differentially Private Federated Learningcited by 66
  30. 2022Differentially Private Optimization on Large Model at Small Costcited by 65
  31. 2022The Limits of Word Level Differential Privacycited by 63
  32. 2021Selective Differential Privacy for Language Modelingcited by 61
  33. 2022Improved Differential Privacy for SGD via Optimal Private Linear Operators on Adaptive Streamscited by 59
  34. 2021Natural Language Understanding with Privacy-Preserving BERTcited by 54
  35. 2020On Differentially Private Stochastic Convex Optimization with Heavy-tailed Datacited by 51
  36. 2022A Customized Text Sanitization Mechanism with Differential Privacycited by 51
  37. 2019Leveraging Hierarchical Representations for Preserving Privacy and Utility in Textcited by 49
  38. 2021Public Data-Assisted Mirror Descent for Private Model Trainingcited by 49

50 most cited of 92 All 92 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 4; 2024: 9; 2025: 5; 2026: 2.

20 papers.

Sort:

By area:

  1. 2023Diminished Diversity-of-Thought in a Standard Large Language Modelcited by 43
  2. 2024Multi-hop Question Answering under Temporal Knowledge Editingcited by 36
  3. 2023Integration of Large Language Models and Federated Learningcited by 28
  4. 2024Privacy Preserving Prompt Engineering: A Surveycited by 24
  5. 2024Prompt-SAW: Leveraging Relation-Aware Graphs for Textual Prompt Compressioncited by 18
  6. 2024A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoningcited by 15
  7. 2024Dialectical Alignment: Resolving the Tension of 3H and Security Threats of LLMscited by 11
  8. 2024Leveraging Logical Rules in Knowledge Editing: A Cherry on the Topcited by 11
  9. 2024Understanding Reasoning in Chain-of-Thought from the Hopfieldian Viewcited by 10
  10. 2025PIXEL: Adaptive Steering Via Position-wise Injection with eXact Estimated Levels under Subspace Calibrationcited by 10
  11. 2024MONAL: Model Autophagy Analysis for Modeling Human-AI Interactionscited by 7
  12. 2023To share or not to share: What risks would laypeople accept to give sensitive data to differentially-private NLP systems?cited by 3
8 more
  1. 2023PrIeD-KIE: Towards Privacy Preserved Document Key Information Extractioncited by 2
  2. 2025Attributing Data for Sharpness-Aware Minimizationcited by 2
  3. 2025Empirical Calibration and Metric Differential Privacy in Language Modelscited by 1
  4. 2025Improved Algorithms for Differentially Private Language Model Alignmentcited by 1
  5. 2025DP-DocLDM: Differentially Private Document Image Generation using Latent Diffusion Modelscited by 1
  6. 2024Regulating Large Language Models: A Roundtable Reportcited by 0
  7. 2026Beyond Theoretical Bounds: Empirical Privacy Loss Calibration for Text Rewriting Under Local Differential Privacycited by 0
  8. 2026The Privacy-Hallucination Tradeoff in Differentially Private Language Modelscited by 0