mlsanityData to 14 Oct 2026

DP-BART for Privatized Text Rewriting under Local Differential Privacy

arXiv 2302.07636 ↗First posted 15 Feb 2023cs.CR cs.CLcited by 28

Not on the map. Why?

Abstract (arXiv)

Privatized text rewriting with local differential privacy (LDP) is a recent approach that enables sharing of sensitive textual documents while formally guaranteeing privacy protection to individuals. However, existing systems face several issues, such as formal mathematical flaws, unrealistic privacy guarantees, privatization of only individual words, as well as a lack of transparency and reproducibility. In this paper, we propose a new system 'DP-BART' that largely outperforms existing LDP systems. Our approach uses a novel clipping method, iterative pruning, and further training of internal representations which drastically reduces the amount of noise required for DP guarantees. We run experiments on five textual datasets of varying sizes, rewriting them at different privacy guarantees and evaluating the rewritten texts on downstream text classification tasks. Finally, we thoroughly discuss the privatized text rewriting approach and its limitations, including the problem of the strict text adjacency constraint in the LDP paradigm that leads to the high noise requirement.

Key facts. Cited by 28 papers in the map (data through 14 Oct 2026). Builds on 23 of its 51 references that are in the map. First posted on arXiv 15 Feb 2023.

Builds on

23 of 51 references.

References · 23

  1. 2017Attention Is All You Needcited by 60,361
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,893
  4. 2016Deep Learning with Differential Privacycited by 2,728
  5. 2018The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networkscited by 2,262
  6. 2015Learning both Weights and Connections for Efficient Neural Networkscited by 2,031
  7. 2020What is the State of Neural Network Pruning?cited by 583
  8. 2016Ups and Downs: Modeling the Visual Evolution of Fashion Trends with One-Class Collaborative Filteringcited by 520
  9. 2018Snips Voice Platform: an embedded Spoken Language Understanding system for private-by-design voice interfacescited by 377
  10. 2018Improving the Gaussian Mechanism for Differential Privacy: Analytical Calibration and Optimal Denoisingcited by 218
  11. 2022Differentially Private Language Models for Secure Data Sharingcited by 68
  12. 2022The Limits of Word Level Differential Privacycited by 63
11 more
  1. 2019Leveraging Hierarchical Representations for Preserving Privacy and Utility in Textcited by 50
  2. 2020A Differentially Private Text Perturbation Method Using a Regularized Mahalanobis Metriccited by 42
  3. 2014Differential Privacy: An Economic Method for Choosing Epsiloncited by 35
  4. 2019ER-AE: Differentially Private Text Generation for Authorship Anonymizationcited by 35
  5. 2021ADePT: Auto-encoder based Differentially Private Text Transformationcited by 31
  6. 2022DP-Rewrite: Towards Reproducibility and Transparency in Differentially Private Text Rewritingcited by 28
  7. 2020A Comprehensive Survey on Local Differential Privacy Toward Data Statistics and Analysiscited by 26
  8. 2021When differential privacy meets NLP: The devil is in the detailcited by 25
  9. 2021Privacy-Preserving Graph Convolutional Networks for Text Classificationcited by 19
  10. 2022Differential Privacy in Natural Language Processing: The Story So Farcited by 12
  11. 2022How reparametrization trick broke differentially-private text representation learningcited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 3; 2024: 10; 2025: 9; 2026: 6.

28 papers.

Sort:

By area:

  1. 2024Safeguarding Large Language Models: A Surveycited by 78
  2. 2024Building Guardrails for Large Language Modelscited by 65
  3. 2023Privacy in Large Language Models: Attacks, Defenses and Future Directionscited by 51
  4. 2024IncogniText: Privacy-enhancing Conditional Text Anonymization via LLM-based Private Attribute Randomizationcited by 23
  5. 2025A False Sense of Privacy: Evaluating Textual Data Sanitization Beyond Surface-level Privacy Leakagecited by 16
  6. 2024Privacy in Fine-tuning Large Language Models: Attacks, Defenses, and Future Directionscited by 14
  7. 2024Just Rewrite It Again: A Post-Processing Method for Enhanced Semantic Similarity and Privacy Preservation of Differentially Private Rewritten Textcited by 9
  8. 2023Neural Text Sanitization with Privacy Risk Indicators: An Empirical Analysiscited by 5
  9. 2025Spend Your Budget Wisely: Towards an Intelligent Distribution of the Privacy Budget in Differentially Private Text Rewritingcited by 5
  10. 2025A Survey on Current Trends and Recent Advances in Text Anonymizationcited by 5
  11. 2024ObfuscaTune: Obfuscated Offsite Fine-tuning and Inference of Proprietary LLMs on Private Datasetscited by 4
  12. 2024Reconstruction of Differentially Private Text Sanitization via Large Language Modelscited by 4
16 more
  1. 2024ExpShield: Safeguarding Web Text from Unauthorized Crawling and LLM Exploitationcited by 3
  2. 2025MaskSQL: Safeguarding Privacy for LLM-Based Text-to-SQL via Abstractioncited by 3
  3. 2024Federated Domain-Specific Knowledge Transfer on Large Language Models Using Synthetic Datacited by 2
  4. 2026Mitigating Gradient Inversion Risks in Language Models via Token Obfuscationcited by 2
  5. 2023Defending Against Authorship Identification Attackscited by 1
  6. 2025Empirical Calibration and Metric Differential Privacy in Language Modelscited by 1
  7. 2025The Double-edged Sword of LLM-based Data Reconstruction: Understanding and Mitigating Contextual Vulnerability in Word-level Differential Privacy Text Sanitizationcited by 1
  8. 2026Towards Privacy-Preserving Machine Translation at the Inference Stage: A New Task and Benchmarkcited by 1
  9. 2024Preempting Text Sanitization Utility in Resource-Constrained Privacy-Preserving LLM Interactionscited by 0
  10. 2025Stereo-LiDAR Fusion by Semi-Global Matching With Discrete Disparity-Matching Cost and Semidensificationcited by 0
  11. 2025RL-Finetuned LLMs for Privacy-Preserving Synthetic Rewritingcited by 0
  12. 2025CTIGuardian: A Few-Shot Framework for Mitigating Privacy Leakage in Fine-Tuned LLMscited by 0
  13. 2026Learning Demographic-Conditioned Mobility Trajectories with Aggregate Supervisioncited by 0
  14. 2026LLM Anonymization Against Agentic Re-Identificationcited by 0
  15. 2026Differentially Private Datastore Generation for Retrieval-Augmented Inferencecited by 0
  16. 2026SpliTEE: Fast and Private LLM Inference by Coupling GPU-Assisted Trusted Execution Environments with Differential Privacycited by 0