Journal article · 2026 · Open access

Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation

Bochao Liu, Shiming Ge, Pengju Wang, Shikun Li, Tongliang Liu

IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(6), 6786–6798 · Published 29 January 2026 · DOI 10.1109/tpami.2026.3659110

Open PDFView at the publisherCite this paper

In brief

“Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation” is a 2026 journal article by Bochao Liu and 4 co-authors, published in IEEE Transactions on Pattern Analysis and Machine Intelligence. OpenAlex counts 0 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).

Citations
0
References
89
Authors
5

Abstract

While many deep learning models trained on private datasets have been deployed in various practical tasks, they may pose a privacy leakage risk as attackers could recover informative data or label knowledge from models. In this work, we present privacy-preserving model transcription, a data-free model-to-model conversion solution to facilitate model deployment with a privacy guarantee. To this end, we propose a cooperative-competitive learning approach termed differentially private synthetic distillation that learns to convert a pretrained model (teacher) into its privacy-preserving counterpart (student) via a trainable generator without access to private data. The learning collaborates with three players in a unified framework and performs alternate optimization: i) the generator is learned to generate synthetic data, ii) the teacher and student accept the synthetic data and compute differential private labels by flexible data or label noisy perturbation, and iii) the student is updated with noisy labels and the generator is updated by taking the student as a discriminator for adversarial training. We theoretically prove that our approach can guarantee differential privacy and convergence. The transcribed student has good performance and privacy protection, while the resulting generator can generate private synthetic data for downstream tasks. Extensive experiments clearly demonstrate that our approach outperforms 26 state-of-the-arts.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Liu, B., Ge, S., Wang, P., Li, S., & Liu, T. (2026). Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(6), 6786–6798. https://doi.org/10.1109/tpami.2026.3659110

MLA 9

Liu, Bochao, et al. “Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation.” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 48, no. 6, 2026, pp. 6786–6798. https://doi.org/10.1109/tpami.2026.3659110.

Chicago

Liu, Bochao, Shiming Ge, Pengju Wang, Shikun Li, and Tongliang Liu. 2026. “Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation.” IEEE Transactions on Pattern Analysis and Machine Intelligence 48 (6): 6786–6798. https://doi.org/10.1109/tpami.2026.3659110.

Harvard

Liu, B. et al. (2026) ‘Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation’, IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(6), pp. 6786–6798. doi:10.1109/tpami.2026.3659110.

IEEE

B. Liu, S. Ge, P. Wang, S. Li, and T. Liu, “Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 48, no. 6, pp. 6786–6798, 2026, doi: 10.1109/tpami.2026.3659110.

BibTeX

@article{liu2026privacy,
  title     = {{Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation}},
  author    = {Liu, Bochao and Ge, Shiming and Wang, Pengju and Li, Shikun and Liu, Tongliang},
  journal   = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year      = {2026},
  volume    = {48},
  number    = {6},
  pages     = {6786--6798},
  doi       = {10.1109/tpami.2026.3659110},
  url       = {https://doi.org/10.1109/tpami.2026.3659110}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data

  • differential privacy
  • synthetic data generation
  • knowledge distillation
  • label noise
  • adversarial training
  • privacy leakage
  • model deployment

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Liu, B., Ge, S., Wang, P., Li, S., & Liu, T. (2026). Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(6), 6786–6798. https://doi.org/10.1109/tpami.2026.3659110

How many times has this paper been cited?

0 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 arXiv (Cornell University): sign in and open the PDF from this page.

Who wrote this paper?

Bochao Liu, Shiming Ge, Pengju Wang, Shikun Li, Tongliang Liu.

What is the DOI of this paper?

10.1109/tpami.2026.3659110, permanently at https://doi.org/10.1109/tpami.2026.3659110. 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.