When the Curious Abandon Honesty: Federated Learning Is Not Private
Open accessThe DOI matches its Crossref record; 17 of the 17 references it cites were found in a registry or an index.
Abstract
In federated learning (FL), data does not leave personal devices when they are jointly training a machine learning model. Instead, these devices share gradients, parameters, or other model updates, with a central party (e.g., a company) coordinating the training. Because data never "leaves" personal devices, FL is often presented as privacy-preserving. Yet, recently it was shown that this protection is but a thin facade, as even a passive, honest-but-curious attacker observing gradients can reconstruct data of individual users contributing to the protocol.In this work, we show a novel data reconstruction attack which allows an active and dishonest central party to efficiently extract user data from the received gradients. While prior work on data reconstruction in FL relies on solving computationally expensive optimization problems or on making easily detectable modifications to the shared model’s architecture or parameters, in our attack the central party makes inconspicuous changes to the shared model’s weights before sending them out to the users. We call the modified weights of our attack trap weights.Our active attacker is able to recover user data perfectly, i.e., with zero error, even when this data stems from the same class. Recovery comes with near-zero costs: the attack requires no complex optimization objectives. Instead, our attacker exploits inherent data leakage from model gradients and simply amplifies this effect by maliciously altering the weights of the shared model through the trap weights. These specificities enable our attack to scale to fully-connected and convolutional deep neural networks trained with large mini-batches of data. For example, for the high-dimensional vision dataset ImageNet, we perfectly reconstruct more than 50% of the training data points from mini-batches as large as 100 data points. In textual tasks, such as IMDB sentiment analysis, more than 65% of data points from mini-batches containing 100 data points can be perfectly reconstructed.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333
TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G. S., Davis, A., Dean, J., Devin, M., et al. (2016) · arXiv · doi:10.48550/arxiv.1603.04467 · Cited by 9,408
Practical Secure Aggregation for Privacy-Preserving Machine Learning
Bonawitz, K., Ivanov, V., Kreuter, B., Marcedone, A., McMahan, H. B., Patel, S., Ramage, D., Segal, A., & Seth, K. (2017) · Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/3133956.3133982 · Cited by 3,956 · cited 2×
“Note that [37]’s main goal is not to extract large amounts individual user data points but user updates, by circumventing the secure aggregation [6] used to protect the FL protocol.”
BackgroundKuznetsova, A., Rom, H., Alldrin, N., Uijlings, J., Krasin, I., Pont-Tuset, J., Kamali, S., Popov, S., Malloci, M., Kolesnikov, A., et al. (2018) · arXiv · doi:10.48550/arxiv.1811.00982 · Cited by 3,653
Federated Learning With Differential Privacy: Algorithms and Performance Analysis
Wei, K., Li, J., Ding, M., Ma, C., Yang, H. H., Farokhi, F., Jin, S., Quek, T. Q. S., & Vincent Poor, H. (2020) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2020.2988575 · Cited by 2,400 · cited 1×
“However, previous work has shown that this setup leads to poor privacy-utility trade-offs, such that LDP is not popular in practical applications [51].”
BackgroundZhu, L., & Han, S. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-63076-8_2 · Cited by 1,906
Exploiting Unintended Feature Leakage in Collaborative Learning
Melis, L., Song, C., De Cristofaro, E., & Shmatikov, V. (2019) · 2019 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2019.00029 · Cited by 1,734
Federated Learning for Internet of Things: A Comprehensive Survey
Nguyen, D. C., Ding, M., Pathirana, P. N., Seneviratne, A., Li, J., & Vincent Poor, H. (2021) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2021.3075439 · Cited by 1,503
Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning
Hitaj, B., Ateniese, G., & Perez-Cruz, F. (2017) · Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/3133956.3134012 · Cited by 1,377 · cited 2×
“Recent work [15], [23], [38], [45], [50], [53], [55], [56] proposed that the central party or users involved in FL training launch data reconstruction attacks based on either training a Generative Adversarial Network [19] (GAN) or solving a second order optimization problem.”
BackgroundA Hybrid Approach to Privacy-Preserving Federated Learning
Truex, S., Baracaldo, N., Anwar, A., Steinke, T., Ludwig, H., Zhang, R., & Zhou, Y. (2019) · Proceedings of the 12th ACM Workshop on Artificial Intelligence and Security · doi:10.1145/3338501.3357370 · Cited by 1,029
Cited by
Privacy-preserving artificial intelligence in healthcare: Techniques and applications
Khalid, N., Qayyum, A., Bilal, M., Al-Fuqaha, A., & Qadir, J. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.106848 · Cited by 581
Collaborative Federated Learning for Healthcare: Multi-Modal COVID-19 Diagnosis at the Edge
Qayyum, A., Ahmad, K., Ahsan, M. A., Al-Fuqaha, A., & Qadir, J. (2022) · IEEE Open Journal of the Computer Society · doi:10.1109/ojcs.2022.3206407 · Cited by 345
“shared between the client and server in FL can be used to reconstruct the input data [60].”
MethodZhao, J., Bagchi, S., Avestimehr, S., Chan, K., Chaterji, S., Dimitriadis, D., Li, J., Li, N., Nourian, A., & Roth, H. (2025) · ACM Computing Surveys · doi:10.1145/3724113 · Cited by 63
Metadata
- Authors
- Boenisch, F., Dziedzic, A., Schuster, R., Shamsabadi, A. S., Shumailov, I., & Papernot, N.
- Venue
- 2023 IEEE 8th European Symposium on Security and Privacy (EuroS&P)
- Publisher
- IEEE
- Published
- 2023-07
- Pages
- 175-199
- OpenAlex
- W4385412495
- Cited by
- 145
- References
- 17
- FWCI
- 26.46
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: When the Curious Abandon Honesty: Federated Learning Is Not Private. https://app.cochat.ai/papers/10.1109/eurosp57164.2023.00020
- The paper
- Boenisch, F., Dziedzic, A., Schuster, R., Shamsabadi, A. S., Shumailov, I., & Papernot, N. (2023). When the Curious Abandon Honesty: Federated Learning Is Not Private. 2023 IEEE 8th European Symposium on Security and Privacy (EuroS&P). https://doi.org/10.1109/eurosp57164.2023.00020
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.