Conference paper · 2026 · Open access
Optimization and Robustness-Informed Membership Inference Attacks for LLMs
Proceedings of the AAAI Conference on Artificial Intelligence, 40(39), 33047–33055 · Published 14 March 2026 · DOI 10.1609/aaai.v40i39.40587
In brief
“Optimization and Robustness-Informed Membership Inference Attacks for LLMs” is a 2026 conference paper by Zichen Song and 3 co-authors, published in Proceedings of the AAAI Conference on Artificial Intelligence. OpenAlex counts 1 citations as of 19 November 2026. It is open access: the full text is free to read at Proceedings of the AAAI Conference on Artificial Intelligence.
- Citations
- 1
- Most cited in
- 2026 (1)
- Field-weighted impact
- 0.88
- References
- 15
- Authors
- 4
Abstract
The proliferation of Large Language Models (LLMs) has raised concerns over training data privacy. Membership Inference Attacks (MIA), aiming to identify whether specific data was used for training, pose significant privacy risks. However, existing MIA methods struggle to address the scale and complexity of modern LLMs. This paper introduces OR-MIA, a novel MIA framework inspired by model optimization and input robustness. First, training data points are expected to exhibit smaller gradient norms due to optimization dynamics. Second, member samples show greater stability, with gradient norms being less sensitive to controlled input perturbations. OR-MIA leverages these principles by perturbing inputs, computing gradient norms, and using them as features for a robust classifier to distinguish members from non-members. Evaluations on LLMs (70M to 6B parameters) and various datasets demonstrate that OR-MIA outperforms existing methods, achieving over 90% accuracy. Our findings highlight a critical vulnerability in LLMs and underscore the need for improved privacy-preserving training paradigms.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Song, Z., Zhang, Q., Li, M., & Shu, Y. (2026). Optimization and Robustness-Informed Membership Inference Attacks for LLMs. Proceedings of the AAAI Conference on Artificial Intelligence, 40(39), 33047–33055. https://doi.org/10.1609/aaai.v40i39.40587
MLA 9
Song, Zichen, et al. “Optimization and Robustness-Informed Membership Inference Attacks for LLMs.” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 40, no. 39, 2026, pp. 33047–33055. https://doi.org/10.1609/aaai.v40i39.40587.
Chicago
Song, Zichen, Qixin Zhang, Ming Li, and Yao Shu. 2026. “Optimization and Robustness-Informed Membership Inference Attacks for LLMs.” Proceedings of the AAAI Conference on Artificial Intelligence 40 (39): 33047–33055. https://doi.org/10.1609/aaai.v40i39.40587.
Harvard
Song, Z. et al. (2026) ‘Optimization and Robustness-Informed Membership Inference Attacks for LLMs’, Proceedings of the AAAI Conference on Artificial Intelligence, 40(39), pp. 33047–33055. doi:10.1609/aaai.v40i39.40587.
IEEE
Z. Song, Q. Zhang, M. Li, and Y. Shu, “Optimization and Robustness-Informed Membership Inference Attacks for LLMs,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 40, no. 39, pp. 33047–33055, 2026, doi: 10.1609/aaai.v40i39.40587.
BibTeX
@inproceedings{song2026optimization,
title = {{Optimization and Robustness-Informed Membership Inference Attacks for LLMs}},
author = {Song, Zichen and Zhang, Qixin and Li, Ming and Shu, Yao},
booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {39},
pages = {33047--33055},
doi = {10.1609/aaai.v40i39.40587},
url = {https://doi.org/10.1609/aaai.v40i39.40587}
}
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
- membership inference attacks
- large language models
- training data privacy
- gradient norm
- input perturbation
- robustness
- privacy-preserving training
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?
Song, Z., Zhang, Q., Li, M., & Shu, Y. (2026). Optimization and Robustness-Informed Membership Inference Attacks for LLMs. Proceedings of the AAAI Conference on Artificial Intelligence, 40(39), 33047–33055. https://doi.org/10.1609/aaai.v40i39.40587
How many times has this paper been cited?
1 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 Proceedings of the AAAI Conference on Artificial Intelligence: sign in and open the PDF from this page.
Who wrote this paper?
Zichen Song, Qixin Zhang, Ming Li, Yao Shu.
What is the DOI of this paper?
10.1609/aaai.v40i39.40587, permanently at https://doi.org/10.1609/aaai.v40i39.40587. 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.