Conference paper · 2022 · Open access
Selective Differential Privacy for Language Modeling
Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), 2848–2859 · Published 1 January 2022 · DOI 10.18653/v1/2022.naacl-main.205
In brief
“Selective Differential Privacy for Language Modeling” is a 2022 conference paper by Weiyan Shi and 4 co-authors, published in Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT). OpenAlex counts 107 citations as of 19 November 2026, which puts it in the top 2% of works published in 2022 by citations. It is open access under a CC BY licence: the full text is free to read at Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT).
- Citations
- 107
- Most cited in
- 2024 (32)
- Field-weighted impact
- 6.29
- References
- 42
- Authors
- 5
Abstract
With the increasing applications of language models, it has become crucial to protect these models from leaking private information.Previous work has attempted to tackle this challenge by training RNN-based language models with differential privacy guarantees.However, applying classical differential privacy to language models leads to poor model performance as the underlying privacy notion is over-pessimistic and provides undifferentiated protection for all tokens in the data.Given that the private information in natural language is sparse (for example, the bulk of an email might not carry personally identifiable information), we propose a new privacy notion, selective differential privacy, to provide rigorous privacy guarantees on the sensitive portion of the data to improve model utility.To realize such a new notion, we develop a corresponding privacy mechanism, Selective-DPSGD, for RNN-based language models.Besides language modeling, we also apply the method to a more concrete application -dialog systems.Experiments on both language modeling and dialog system building show that the proposed privacy-preserving mechanism achieves better utilities while remaining safe under various privacy attacks compared to the baselines.The data and code are released to facilitate future research 1 .
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Shi, W., Cui, A., Li, E., Jia, R., & Zhou, Y. (2022). Selective Differential Privacy for Language Modeling. Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), 2848–2859. https://doi.org/10.18653/v1/2022.naacl-main.205
MLA 9
Shi, Weiyan, et al. “Selective Differential Privacy for Language Modeling.” Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), 2022, pp. 2848–2859. https://doi.org/10.18653/v1/2022.naacl-main.205.
Chicago
Shi, Weiyan, Aiqi Cui, Evan Li, Ruoxi Jia, and Yu Zhou. 2022. “Selective Differential Privacy for Language Modeling.” Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT): 2848–2859. https://doi.org/10.18653/v1/2022.naacl-main.205.
Harvard
Shi, W. et al. (2022) ‘Selective Differential Privacy for Language Modeling’, Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pp. 2848–2859. doi:10.18653/v1/2022.naacl-main.205.
IEEE
W. Shi, A. Cui, E. Li, R. Jia, and Y. Zhou, “Selective Differential Privacy for Language Modeling,” Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pp. 2848–2859, 2022, doi: 10.18653/v1/2022.naacl-main.205.
BibTeX
@inproceedings{shi2022selective,
title = {{Selective Differential Privacy for Language Modeling}},
author = {Shi, Weiyan and Cui, Aiqi and Li, Evan and Jia, Ruoxi and Zhou, Yu},
booktitle = {Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT)},
year = {2022},
pages = {2848--2859},
doi = {10.18653/v1/2022.naacl-main.205},
url = {https://doi.org/10.18653/v1/2022.naacl-main.205}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Citations by year
| Year | Citations | Share of the best year |
|---|---|---|
| 2022 | 8 | |
| 2023 | 22 | |
| 2024 | 32 | |
| 2025 | 24 | |
| 2026 | 21 |
Its field-weighted citation impact is 6.29: it has been cited 6.29 times as often as the average work of the same field, type and year (1.00 is the world average).
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- language models
- privacy-preserving machine learning
- sensitive data protection
- dialogue systems
- privacy-utility trade-off
- privacy attacks
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Shi, W., Cui, A., Li, E., Jia, R., & Zhou, Y. (2022). Selective Differential Privacy for Language Modeling. Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), 2848–2859. https://doi.org/10.18653/v1/2022.naacl-main.205
How many times has this paper been cited?
107 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT): sign in and open the PDF from this page.
Who wrote this paper?
Weiyan Shi, Aiqi Cui, Evan Li, Ruoxi Jia, Yu Zhou.
What is the DOI of this paper?
10.18653/v1/2022.naacl-main.205, permanently at https://doi.org/10.18653/v1/2022.naacl-main.205. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.