Practical Secure Aggregation for Privacy-Preserving Machine Learning
The DOI matches its Crossref record; 56 of the 56 references it cites were found in a registry or an index.
Abstract
We design a novel, communication-efficient, failure-robust protocol for secure aggregation of high-dimensional data. Our protocol allows a server to compute the sum of large, user-held data vectors from mobile devices in a secure manner (i.e. without learning each user's individual contribution), and can be used, for example, in a federated learning setting, to aggregate user-provided model updates for a deep neural network. We prove the security of our protocol in the honest-but-curious and active adversary settings, and show that security is maintained even if an arbitrarily chosen subset of users drop out at any time. We evaluate the efficiency of our protocol and show, by complexity analysis and a concrete implementation, that its runtime and communication overhead remain low even on large data sets and client pools. For 16-bit input values, our protocol offers $1.73 x communication expansion for 210 users and 220-dimensional vectors, and 1.98 x expansion for 214 users and 224-dimensional vectors over sending data in the clear.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Although SMC guarantees that none of the parties shares anything with each other or with any third party, it can not prevent an adversary from learning some individual information”
From Federated Learning for Healthcare Informatics (2021) · cites this paper as [8]
In the citing paper
“By combining local inference with trusted aggregation, TDAI can enhance both privacy and robustness of anomaly detection pipelines [239].”
From Anomaly detection and root-cause identification in microservices: a survey (2026) · cites this paper as [239]
In the citing paper
“FL is still in its infancy and is still vulnerable to various challenges [148,149].”
From Smart Wearables for the Detection of Cardiovascular Diseases: A Systematic Literature Review (2023) · cites this paper as [148]
In the citing paper
“Therefore, many FL frameworks use secure aggregation [81], [82], [83], [84] algorithms in the aggregation process, at which point most of the poisoning attack defense methods mentioned above will be invalid.”
From Poisoning attacks in federated learning: A survey (2023) · cites this paper as [81]
In the citing paper
“Bonawitz et al. [62] introduce a secure aggregation protocol for protecting individual model updates.”
From Handling privacy-sensitive medical data with federated learning: Challenges and future directions (2023) · cites this paper as [62]
6 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
New directions in cryptography
Diffie, W., & Hellman, M. (1976) · IEEE Transactions on Information Theory · doi:10.1109/tit.1976.1055638 · Cited by 14,713
Communication-Efficient Learning of Deep Networks from Decentralized Data
McMahan, H. B., Moore, E., Ramage, D., Hampson, S., & Arcas, B. A. Y. (2016) · arXiv · doi:10.48550/arxiv.1602.05629 · Cited by 14,215 · cited 1×
“maintains a private database of her text messages securely on her own mobile device, and a shared global model is trained under the coordination of a central server based upon highly processed, minimally scoped, ephemeral updates from users [20], [10].”
BackgroundSelf-citationShamir, A. (1979) · Communications of the ACM · doi:10.1145/359168.359176 · Cited by 13,756 · cited 1×
“We rely on Shamir’s t -out-of- n Secret Sharing [38], which allows a user to split a secret s into n shares, such that any t shares can be used to reconstruct s , but any set of at most t − 1 shares gives no information about s .”
Method(2016) · Cited by 11,679
Deep Learning with Differential Privacy
Abadi, M., Chu, A., Goodfellow, I., McMahan, H. B., Mironov, I., Talwar, K., & Zhang, L. (2016) · Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/2976749.2978318 · Cited by 6,578 · cited 3×
“This is particular advantageous in the local privacy setting [13], which offers provable guarantees for the protection of individual training examples [14], [15] even when the data aggregator is not assumed to be trusted [16], [17].”
BackgroundMethodSelf-citationThe Algorithmic Foundations of Differential Privacy
Dwork, C., & Roth, A. (2014) · Foundations and Trends® in Theoretical Computer Science · doi:10.1561/0400000042 · Cited by 5,557
Dwork, C. (2006) · Lecture Notes in Computer Science · doi:10.1007/11787006_1 · Cited by 5,291 · cited 1×
“In such cases, secure aggregation composes well with differential privacy [12].”
BackgroundMembership Inference Attacks Against Machine Learning Models
Shokri, R., Stronati, M., Song, C., & Shmatikov, V. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.41 · Cited by 4,754
The Algorithmic Foundations of Differential Privacy
Dwork, C., & Roth, A. (2013) · doi:10.1561/9781601988195 · Cited by 3,856
Federated Learning: Strategies for Improving Communication Efficiency
Konečný, J., McMahan, H. B., Yu, F. X., Richtárik, P., Suresh, A. T., & Bacon, D. (2016) · arXiv · doi:10.48550/arxiv.1610.05492 · Cited by 3,766
Self-citation
Research works printed but not linked: 3
- Stephen Advokat. 1987. Publication Of Bork's Video Rentals Raises Privacy Issue. Chicago Tribune (1987).
- Michael Barbaro, Tom Zeller, and Saul Hansell. 2006. A face is exposed for AOL searcher no. 4417749. New York Times 9, 2008 (2006).
- Elaine Shi, HTH Chan, Eleanor Rieffel, Richard Chow, and Dawn Song. 2011. Privacy-preserving aggregation of time-series data. In Annual Network & Distributed System Security Symposium (NDSS). Internet Society.
Cited by
Items 401–500 of 1,035
Continual Local Training For Better Initialization Of Federated Models
Yao, X., & Sun, L. (2020) · Proceedings - International Conference on Image Processing · doi:10.1109/icip40778.2020.9190968 · Cited by 113
“, differential privacy guarantees [5], homomorphic encryption [6], and secure frameworks [7].”
BackgroundLoki: Large-scale Data Reconstruction Attack against Federated Learning through Model Manipulation
Zhao, J. C., Sharma, A., Elkordy, A. R., Ezzeldin, Y. H., Avestimehr, S., & Bagchi, S. (2024) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp54263.2024.00030 · Cited by 116
“However,secure aggregation has been shown to be an effective defense against all optimization-based privacy attacks in FL. Secure aggregation guarantees that the server will not gain access to individual model updates of other clients, only the aggregate of all client updates [14], [15], [16].”
BackgroundMethodInfluentialGrataloup, A., Jonas, S., & Meyer, A. (2024) · Energy and AI · doi:10.1016/j.egyai.2024.100375 · Cited by 116
“Privacy enhancements include: Secure multi-party computation [52] enables the secure computation of functions (e.g., sums) on data from different clients without actually sharing the full data samples.”
Inference Attacks Against Collaborative Learning.Preprint
Melis, L., Song, C., Cristofaro, E. D., & Shmatikov, V. (2018) · arXiv (Cornell University) · Cited by 112
Federated Learning with Autotuned Communication-Efficient Secure Aggregation
Bonawitz, K., Salehi, F., Konečný, J., McMahan, B., & Gruteser, M. (2019) · 2019 53rd Asilomar Conference on Signals, Systems, and Computers · doi:10.1109/ieeeconf44664.2019.9049066 · Cited by 76
Self-citationA Survey on federated learning
Li, L., Fan, Y., & Lin, K. (2020) · International Conference on Control and Automation · doi:10.1109/icca51439.2020.9264412 · Cited by 112
“Specially, TFF integrates FedAvg for model update and Secure Aggregation for privacy concern [9].”
BackgroundMethod${\sf FederBoost}$: Private Federated Learning for GBDT
Tian, Z., Zhang, R., Hou, X., Lyu, L., Zhang, T., Liu, J., & Ren, K. (2023) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2023.3276365 · Cited by 111
“t they cannot train models locally. Furthermore, the FL research committee is focusing on neural networks and less attention has been paid to decision trees. 3 2.3 Secure aggregation Bonawitz, et al. [6] propose a secure aggregation protocol to protect the local gradients in Google’s horizontal FL. Specifically, they u…”
InfluentialDifferentially Private Federated Learning With an Adaptive Noise Mechanism
Xue, R., Xue, K., Zhu, B., Luo, X., Zhang, T., Sun, Q., & Lü, J. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3318944 · Cited by 111
“Another methodology is the use of data masking [9], [10], [11].”
MethodIPLS: A Framework for Decentralized Federated Learning
Pappas, C., Chatzopoulos, D., Lalis, S., & Vavalis, M. (2021) · IFIP Networking Conference (IFIP Networking) · doi:10.23919/ifipnetworking52078.2021.9472790 · Cited by 113
Federated Quantum-Based Privacy-Preserving Threat Detection Model for Consumer Internet of Things
Namakshenas, D., Yazdinejad, A., Dehghantanha, A., & Srivastava, G. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3377550 · Cited by 113
Saha, S., Hota, A., Chattopadhyay, A. K., Nag, A., & Nandi, S. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-024-10766-7 · Cited by 111
Blika, A., Palmos, S., Doukas, G. S., Lamprou, V., Pelekis, S., Kontoulis, M., Ntanos, C., & Askounis, D. (2024) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2024.3449563 · Cited by 118
“The study in [274] presented a protocol designed to securely combine individual gradients using SMC focusing on creating a secure, efficient communication process that could withstand failures.”
Secure Machine Learning Scenario from Big Data in Cloud Computing via Internet of Things Network
Stergiou, C. L., Plageras, A. P., Psannis, K. E., & Gupta, B. B. (2019) · Springer eBooks · doi:10.1007/978-3-030-22277-2_21 · Cited by 109
NPMML: A Framework for Non-interactive Privacy-preserving Multi-party Machine Learning
Li, T., Li, J., Chen, X., Liu, Z., Lou, W., & Hou, T. (2020) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2020.2971598 · Cited by 109
Enabling Execution Assurance of Federated Learning at Untrusted Participants
Zhang, X., Li, F., Zhang, Z., Li, Q., Wang, C., & Wu, J. (2020) · IEEE INFOCOM 2020 - IEEE Conference on Computer Communications · doi:10.1109/infocom41043.2020.9155414 · Cited by 87
“They can be well addressed by differential privacy [29], [30] and secure multi-party computation based aggregation techniques [31], which are complementary to our work.”
BackgroundDifferentially Private Byzantine-Robust Federated Learning
Ma, X., Sun, X., Wu, Y., Liu, Z., Chen, X., & Dong, C. (2022) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2022.3167434 · Cited by 110
“HE Practical Secure Aggregation [40] SMC – Hybrid [41] DP+SMC LDP-Fed [26] DP However, DPBFL does not require computation-intensive cryptographic protocols and can defend against attacks because the data exchanged throughout the entire training phase is differentially private.”
BackgroundBlockFLA: Accountable Federated Learning via Hybrid Blockchain Architecture
Desai, H., Özdayi, M. S., & Kantarcıoğlu, M. (2021) · ACM Conference on Data and Application Security and Privacy (CODASPY) · doi:10.1145/3422337.3447837 · Cited by 109
“For example, original FL paper [21] and many subsequent papers on FL [3, 5, 7, 12, 32] consider weighted averaging to aggregate updates.”
BackgroundEncode, Shuffle, Analyze Privacy Revisited: Formalizations and Empirical Evaluation
Erlingsson, Ú., Feldman, V., Mironov, I., Raghunathan, A., Song, S., Talwar, K., & Thakurta, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2001.03618 · Cited by 33
“…constrained by the specifics of their construction and mechanisms (e.g., built-in sampling rates and means of multi-party cryptographic computation, as in [11]); some systems are more specific still, and focus only on certain applications, such as the maintenance of (statistical) models [12], [22].”
BackgroundInfluentialPrivacy-Preserving Federated Learning via Functional Encryption, Revisited
Chang, Y., Zhang, K., Gong, J., & Qian, H. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3255171 · Cited by 108
“2) Using Multi-Party Computation (MPC) [11], [12], [13], [14], [15], [16]: Most work in this category employ secret sharing [17].”
MethodIoannou, I. I., Nagaradjane, P., Angın, P., Balasubramanian, P., Kavitha, K. J., Murugan, P., & Vassiliou, V. (2024) · Computer Communications · doi:10.1016/j.comcom.2024.02.023 · Cited by 106
Privacy in Deep Learning: A Survey
Mireshghallah, F., Taram, M., Vepakomma, P., Singh, A., Raskar, R., & Esmaeilzadeh, H. (2020) · arXiv · doi:10.48550/arxiv.2004.12254 · Cited by 106
“-Differential Privacy ( -DP).”
MethodInfluentialFederated Self-Supervised Learning of Multisensor Representations for Embedded Intelligence
Saeed, A., Salim, F. D., Ozcelebi, T., & Lukkien, J. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.3009358 · Cited by 106
“The research interest in federated learning revolves around improving communication efficiency [39], personalization [40], fault tolerance [41], privacy preservation [42], as well as looking into the theoretical underpinning of the federated optimization [37].”
BackgroundFederated Learning With Taskonomy for Non-IID Data
Jamali‐Rad, H., Abdizadeh, M. T., & Singh, A. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3152581 · Cited by 107
“A possible solution to address this is adding encryption and client ID verification processes, which are outside the scope of our work [6], [46], [47].”
BackgroundSPDL: A Blockchain-Enabled Secure and Privacy-Preserving Decentralized Learning System
Xu, M., Zou, Z., Cheng, Y., Hu, Q., Yu, D., & Cheng, X. (2022) · IEEE Transactions on Computers · doi:10.1109/tc.2022.3169436 · Cited by 106
“Federated learning [1] utilizes an efficient centralized network illustrated in Fig.”
MethodCommunication-Computation Efficient Secure Aggregation for Federated LearningPreprint
Choi, B., Sohn, J.-Y., Han, D.-J., & Moon, J. J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.05433 · Cited by 106
“For every n, q total setup, the proposed CCESA( n, p ) requires p times less running time compared with the conventional SA of (Bonawitz et al., 2017).”
BackgroundMethodInfluentialAI on the edge: a comprehensive review
Su, W., Li, L., Liu, F., He, M., & Liang, X. (2022) · Artificial Intelligence Review · doi:10.1007/s10462-022-10141-4 · Cited by 106
“Many existing works on FL suffer from several problems: (1) device resources are inevitably consumed when users participate in collaborative training, which may lower the willingness of users to engage in training; (2) users, as providers of the data, may refuse to participate in training due to privacy concerns; (3)…”
BackgroundJahani-Nezhad, T., Maddah-Ali, M. A., Li, S., & Caire, G. (2023) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2023.3242702 · Cited by 106
“4) Inter-group Communication and Aggregation: In this phase, user (1, t), t ∈ [6], calculates message S(1,t) = Q(1,t), and it to user (2, t).”
BackgroundInfluentialXing, Z., Zhang, Z., Zhang, Z., Li, Z., Li, M., Liu, J., Zhang, Z., Zhao, Y., Sun, Q., Zhu, L., et al. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3561657 · Cited by 109
“To provide further privacy protection, a mask-based secure aggregation protocol [173] is also introduced.”
DP-FL: a novel differentially private federated learning framework for the unbalanced data
Huang, X., Ding, Y., Jiang, Z. L., Qi, S., Wang, X., & Liao, Q. (2020) · World Wide Web · doi:10.1007/s11280-020-00780-4 · Cited by 104
“Bonawitz et al. [4] propose a practical secure aggregation protocol in a federated learning setting for high dimensional data such as text messages.”
BackgroundEfficient-FedRec: Efficient Federated Learning Framework for Privacy-Preserving News Recommendation
Yi, J., Wu, F., Wu, C., Liu, R., Sun, G., & Xie, X. (2021) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2021.emnlp-main.223 · Cited by 104
“In this subsection, we first introduce secure aggregation proposed by Bonawitz et al. (2017), and then introduce how we apply it to our framework for secure gradients aggregation and news representations distributing.”
Industry-Scale Orchestrated Federated Learning for Drug Discovery
Oldenhof, M., Ács, G., Pejó, B., Schuffenhauer, A., Holway, N., Sturm, N., Dieckmann, A., Fortmeier, O., Boniface, E., Mayer, C., et al. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i13.26847 · Cited by 105
FedDisco: Federated Learning with Discrepancy-Aware CollaborationPreprint
Ye, R., Xu, M., Wang, J., Xu, C., Chen, S., & Wang, Y. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2305.19229 · Cited by 106
“Specifically, we can send each client’s category distribution Dk (a vector) to the server using Secure Aggregation (Bonawitz et al., 2017) technique such that the server knows the actual global category distribution without knowing each client’s actual category distribution.”
MethodAuditing Data Provenance in Text-Generation Models
Song, C., & Shmatikov, V. (2019) · Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3292500.3330885 · Cited by 103
“designed a secure aggregation algorithm for distributed training that hides users’ contributions and can be used in conjunction with a differentially private mechanism to provide stronger guarantees [5]. Auditing target models trained with user-level differential privacy is expected to fail. However, given the current…”
BackgroundPrivacy-preserving big data analytics a comprehensive survey
Tran, H.-Y., & Hu, J. (2019) · Journal of Parallel and Distributed Computing · doi:10.1016/j.jpdc.2019.08.007 · Cited by 89
SecureNLP: A System for Multi-Party Privacy-Preserving Natural Language Processing
Feng, Q., He, D., Liu, Z., Wang, H., & Choo, K. R. (2020) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2020.2997134 · Cited by 103
“There are, however security and privacy concerns associated with NLP-enabled data analytics applications [2]–[5].”
BackgroundMatching-Theory-Based Low-Latency Scheme for Multitask Federated Learning in MEC Networks
Chen, D., Hong, C. S., Wang, L., Zha, Y., Zhang, Y., Liu, X., & Han, Z. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3053283 · Cited by 104
“At the same time, this manner can also decrease the probability of eavesdropping to a certain extend such that the privacy can be guaranteed [8].”
BackgroundA Survey on Gradient Inversion: Attacks, Defenses and Future Directions
Zhang, R., Guo, S., Wang, J., Xie, X. N., & Tao, D. (2022) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2022/791 · Cited by 105
FederatedTrust: A solution for trustworthy federated learning
Sánchez, P. M. S., Celdrán, A. H., Xie, N., Bovet, G., Pérez, G. M., & Stiller, B. (2023) · Future Generation Computer Systems · doi:10.1016/j.future.2023.10.013 · Cited by 104
ACORN: Input Validation for Secure Aggregation
Bell, J., Gascón, A., Lepoint, T., Li, B., Meiklejohn, S., Raykova, M., & Yun, C. (2023) · UCL Discovery (University College London) · Cited by 104
Privacy-Preserving Aggregation for Federated Learning-Based Navigation in Vehicular Fog
Kong, Q., Yin, F., Lu, R., Li, B., Wang, X., Cui, S., & Zhang, P. (2021) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2021.3075683 · Cited by 102
“[8] proposed a privacy-preserving scheme with the secure multiparty computation technique.”
BackgroundMethodResultInfluentialScalable edge-based hyperdimensional learning system with brain-like neural adaptation
Zou, Z., Kim, Y., Imani, F., Alimohamadi, H., Cammarota, R., & Imani, M. (2021) · International Conference for High Performance Computing, Networking, Storage and Analysis (SC) · doi:10.1145/3458817.3480958 · Cited by 102
Artificial Intelligence‐Based Medical Sensors for Healthcare System
Chen, M., Cui, D., Haick, H., & Tang, N. (2023) · Advanced Sensor Research · doi:10.1002/adsr.202300009 · Cited by 102
Anonymizing Data for Privacy-Preserving Federated LearningPreprint
Choudhury, O., Gkoulalas-Divanis, A., Salonidis, T., Sylla, I., Park, Y., Hsu, G., & Das, A. K. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.09096 · Cited by 102
“SMC for FL has been proposed to compute sums of model parameter updates from individual user’s devices in a secure manner [6].”
BackgroundTalwar, K. (2022) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.48550/arxiv.2202.10618 · Cited by 102
“This approach goes back to the early days of differential privacy [16], and has been heavily investigated [11, 8].”
A Review of Privacy-Preserving Federated Learning for the Internet-of-Things
Briggs, C., Fan, Z., & Andras, P. (2021) · Studies in Computational Intelligence · doi:10.1007/978-3-030-70604-3_2 · Cited by 101
“[106] 2019 Description of a full-scale production-ready federated learning system (focusing on mobile devices) SMC Optionally makes use of the Secure aggregation protocol in [103]”
BackgroundMethodInfluentialFedSAE: A Novel Self-Adaptive Federated Learning Framework in Heterogeneous Systems
Li, L., Duan, M., Liu, D., Zhang, Y., Ren, A., Chen, X., Tan, Y., & Wang, C. (2021) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn52387.2021.9533876 · Cited by 101
“For another aspect, some researchers try building a better FL system instead of optimizing the FL algorithm [9], [19], [24] where [9] preliminarily explores a scalable production FL system for Android while [19], [24] provide FL more secure”
BackgroundFederated Multi-Target Domain Adaptation
Yao, C.-H., Gong, B., Qi, H., Cui, Y., Zhu, Y., & Yang, M. (2022) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv51458.2022.00115 · Cited by 101
On-Device Recommender Systems: A Comprehensive Survey
Yin, H., Qu, L., Chen, T., Yuan, W., Zheng, R., Long, J., Xia, X., Shi, Y., & Zhang, C. (2025) · Data Science and Engineering · doi:10.1007/s41019-025-00308-8 · Cited by 101
“…global parameters as Θ ( t + 1 ) • Other Methods: Federated neural collaborative filtering (FedNCF) [105] introduces a novel aggregation method called MF-SecAvg, which extends the traditional FedAvg [129] by integrating the secure aggregation (SecAvg) protocol [138] with a low computational cost.”
Federated Learning Systems: Vision, Hype and Reality for Data Privacy and ProtectionPreprint
Li, Q., Wen, Z., & He, B. (2019) · arXiv (Cornell University) · Cited by 101
CS-MIA: Membership inference attack based on prediction confidence series in federated learning
Gu, Y., Bai, Y., & Xu, S. (2022) · Journal of Information Security and Applications · doi:10.1016/j.jisa.2022.103201 · Cited by 102
Privacy-Preserving and Byzantine-Robust Federated Learning
Dong, C., Weng, J., Li, M., Liu, J.-N., Liu, Z., Cheng, Y., & Yu, S. (2023) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2023.3264697 · Cited by 100
“Bonawitz et al. [9] used multi-server computation to preserve the privacy of parameters and proposed a secure aggregation rule which can tolerate users’ dropouts without sacrificing model privacy.”
BackgroundMethodTrading Off Privacy, Utility, and Efficiency in Federated Learning
Zhang, X., Kang, Y., Chen, K., Fan, L., & Yang, Q. (2023) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3595185 · Cited by 104
“[5, 6, 38] were developed to distribute a secret among a group of participants.”
BackgroundFedTracker: Furnishing Ownership Verification and Traceability for Federated Learning Model
Shao, S., Yang, W., Gu, H., Qin, Z., Fan, L., Yang, Q., & Ren, K. (2024) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2024.3390761 · Cited by 101
“The local gradients of the Clients are noised or encrypted [46].”
BackgroundNo Free Lunch Theorem for Security and Utility in Federated Learning
Zhang, X., Gu, H., Fan, L., Chen, K., & Yang, Q. (2022) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3563219 · Cited by 98
“3 Applications of Privacy-Utility Trade-of In this section, we apply the no free lunch theorem between the Bayesian privacy leakage and utility loss on four privacy-preserving mechanisms: Randomization [1, 34, 75], Sparsity [38, 69, 70], Homomorphic Encryption [31, 84] and Secret Sharing [4, 6, 68](9).”
BackgroundMethodInfluentialPatel, V. A., Bhattacharya, P., Tanwar, S., Gupta, R., Sharma, G., Bokoro, P. N., & Sharma, R. (2022) · IEEE Access · doi:10.1109/access.2022.3201876 · Cited by 79
“The authors in [133] proposed a trusted aggregation sys-tem for FL, but it possesses high communication and computation overhead.”
BackgroundFederated Learning Versus Classical Machine Learning: A Convergence ComparisonPreprint
Asad, M., Moustafa, A., & Itō, T. (2021) · Authorea · doi:10.22541/au.162074596.66890690/v1 · Cited by 99
Federated Learning of N-Gram Language Models
Chen, M., Suresh, A. T., Mathews, R., Wong, A., Allauzen, C., Beaufays, F., & Riley, M. (2019) · Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL) · doi:10.18653/v1/k19-1012 · Cited by 57
“…proposed approach has several advantages: Additional privacy-preserving techniques: FL can be further combined with privacy-preserving techniques such as secure multi-party computation (Bonawitz et al., 2017) and differential privacy (McMahan et al., 2018; Agarwal et al., 2018; Abadi et al., 2016).”
BackgroundInformation Theoretic Secure Aggregation With User Dropouts
Zhao, Y., & Sun, H. (2022) · IEEE Transactions on Information Theory · doi:10.1109/tit.2022.3192874 · Cited by 98
“Therefore, regarding security, the central technical problem is secure sum computation or secure aggregation [4], [5], i.e., how to compute, with as little communication as possible, the sum of the inputs of a number of users without exposing any information beyond the sum.”
BackgroundMethodMime: Mimicking Centralized Stochastic Algorithms in Federated Learning
Karimireddy, S. P., Jaggi, M., Kale, S., Mohri, M., Reddi, S. J., Stich, S. U., & Suresh, A. T. (2020) · arXiv · doi:10.48550/arxiv.2008.03606 · Cited by 96
Trusted AI in Multiagent Systems: An Overview of Privacy and Security for Distributed Learning
Ma, C., Li, J., Wei, K., Liu, B., Ding, M., Yuan, L., Han, Z., & Vincent Poor, H. (2023) · Proceedings of the IEEE · doi:10.1109/jproc.2023.3306773 · Cited by 96
“For instance, in the supervised ML task, secure multi-party computation (SMPC) and homomorphic encryption (HE) based privacy-enhancing tools have been proposed to enable secure computing.”
BackgroundMethodInfluentialA Simple and Efficient Federated Recommender System
Jalalirad, A., Scavuzzo, M., Capota, C., & Sprague, M. (2019) · Proceedings of the 6th IEEE/ACM International Conference on Big Data Computing, Applications and Technologies · doi:10.1145/3365109.3368788 · Cited by 69
“Encryption techniques are applied in [3, 6] to prevent an honest-but-curious parameter server from reconstructing users’ gradients or parameters updates.”
MethodGarrido, G. M., Sedlmeir, J., Uludağ, Ö., Alaoui, I. S., Luckow, A., & Matthes, F. (2022) · Journal of Network and Computer Applications · doi:10.1016/j.jnca.2022.103465 · Cited by 99
Toward Quantum Federated Learning
Ren, C., Yan, R., Zhu, H., Yu, H., Xu, M., Shen, Y., Xu, Y. H., Xiao, M., Dong, Z. Y., Skoglund, M., et al. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3552643 · Cited by 97
Forecasting influenza activity using machine-learned mobility map
Venkatramanan, S., Sadilek, A., Fadikar, A., Barrett, C. L., Biggerstaff, M., Chen, J., Dotiwalla, X., Eastham, P., Gipson, B. R., Higdon, D., et al. (2021) · Nature Communications · doi:10.1038/s41467-021-21018-5 · Cited by 95
“With sophisticated agent-based models 39 and on-device learning 40,41 , such techniques could also lead the way towards individual-level forecasting without involving the same privacy and anonymity constraints.”
BackgroundApplications of Federated Learning in Mobile Health: Scoping Review
Wang, T., Du, Y. N., Gong, Y., Choo, K. R., & Guo, Y. (2023) · Journal of Medical Internet Research · doi:10.2196/43006 · Cited by 95
“There are 3 main ways to protect data privacy in the FL framework [40], which are as follows: SMC [42], DP [41,43], and homomorphic encryption [47].”
Towards Utilizing Unlabeled Data in Federated Learning: A Survey and ProspectivePreprint
Jin, Y., Wei, X., Liu, Y., & Yang, Q. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.11545 · Cited by 104
“…, 2020; Wang et al. , 2020 ] , federated learning algorithms [ Cheng et al. , 2019; Li et al. , 2019; Shokri and Shmatikov, 2015 ] , privacy mechanisms and attacks [ Hitaj et al. , 2017; Mohassel and Zhang, 2017; Bonawitz et al. , 2017 ] , systems and communication [ Bonawitz et al. , 2019 ] , etc.”
Distributed Deep Learning in Open CollaborationsPreprint
Diskin, M., Bukhtiyarov, A., Ryabinin, M., Saulnier, L., Lhoest, Q., Sinitsin, A., Popov, D., Pyrkin, D., Kashirin, M., Borzunov, A., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.10207 · Cited by 99
FedPacket: A Federated Learning Approach to Mobile Packet Classification
Bakopoulou, E., Tillman, B., & Markopoulou, A. P. (2021) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2021.3058627 · Cited by 95
“SA [23] was proposed early as a defense mechanism added on top of FL.”
BackgroundInfluentialA survey: Distributed Machine Learning for 5G and beyond
Nassef, O., Sun, W., Purmehdi, H., Tatipamula, M., & Mahmoodi, T. (2022) · Computer Networks · doi:10.1016/j.comnet.2022.108820 · Cited by 95
Advances in Robust Federated Learning: A Survey With Heterogeneity Considerations
Chen, C., Liao, T., Deng, X., Wu, Z., Huang, S., & Zheng, Z. (2025) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2025.3527202 · Cited by 96
“Model perturbation-based methods enable relatively quick protection but reduce model accuracy, whereas cryptographic approaches maintain model performance nearly unaffected [66].”
Boosting Privately: Federated Extreme Gradient Boosting for Mobile Crowdsensing
Liu, Y., Ma, Z., Liu, X., Ma, S., Nepal, S., Deng, R. H., & Ren, K. (2020) · 2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS) · doi:10.1109/icdcs47774.2020.00017 · Cited by 93
“In F ED XGB, our security model is based on the curious-but-honest model, a standard security model in federated learning [4], [8], [11].”
BackgroundMethodInfluentialPDGAN: A Novel Poisoning Defense Method in Federated Learning Using Generative Adversarial Network
Zhao, Y., Chen, J., Zhang, J., Wu, D., Teng, J., & Yu, S. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-38991-8_39 · Cited by 93
A survey on federated learning in data mining
Yu, B., Mao, W., Lv, Y., Zhang, C. B. C., & Xie, Y. (2021) · Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery · doi:10.1002/widm.1443 · Cited by 94
DisBezant: Secure and Robust Federated Learning Against Byzantine Attack in IoT-Enabled MTS
Ma, X., Jiang, Q. G., Shojafar, M., Alazab, M., Kumar, S., & Kumari, S. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2022.3152156 · Cited by 93
“Based on the secret sharing, Bonawitz et al. [36] proposed a failure-robust protocol for secure data aggregation in machine learning model training.”
FedPOIRec: Privacy-preserving federated poi recommendation with social influence
Perifanis, V., Drosatos, G., Stamatelatos, G., & Efraimidis, P. S. (2023) · Information Sciences · doi:10.1016/j.ins.2022.12.024 · Cited by 79
Habibullah, K. M., Gay, G., & Horkoff, J. (2023) · Requirements Engineering · doi:10.1007/s00766-022-00395-3 · Cited by 93
“Therefore, despite the advances allowed by ML, much recent attention has been paid to certain qualities of ML systems—particularly regarding fairness [37], transparency [14], privacy [15], security [42], and safety [19].”
BackgroundTran, H.-Y., Hu, J., Yin, X., & Pota, H. R. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3267892 · Cited by 93
“Addressing the drawbacks of [18] and [19], the work in [20] proposed a secure aggregation scheme where the input is a vector and can deal with dropouts.”
BackgroundMethodInfluentialPractical Privacy Preserving POI Recommendation
Chen, C., Zhou, J., Wu, B., Fang, W., Wang, L., Qi, Y., & Zheng, X. (2020) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3394138 · Cited by 76
“The existing works propose a secure aggregation strategy to train a neural network model in federated learning settings [5, 29].”
BackgroundMethodA Privacy-Preserving Federated Learning System for Android Malware Detection Based on Edge Computing
Hsu, R., Wang, Y., Fan, C., Sun, B., Ban, T., Takahashi, T., Wu, T.-W., & Kao, S.-W. (2020) · Asia Joint Conference on Information Security (AsiaJCIS) · doi:10.1109/asiajcis50894.2020.00031 · Cited by 94
“Finally, we are also interested in the key agreement of the PPFL which mentioned in [26] because the encryption method in our framework maybe is not enough, and we want to design a stronger key agreement method to protect user data.”
BackgroundMethodFedAR: Activity and Resource-Aware Federated Learning Model for Distributed Mobile Robots
Imteaj, A., & Hadi Amini, M. (2020) · 2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA) · doi:10.1109/icmla51294.2020.00185 · Cited by 66
“Another important factor of FL is that it works with nonindependent and identically distributed (non-IID) data samples observed in real-world settings [5], [6].”
Dong, Y., Chen, X., Li, K., Wang, D., & Zeng, S. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-88418-5_24 · Cited by 92
“designed a secure aggregation scheme for sum function by exploiting secret sharing and key agreement protocol [8].”
BackgroundMethodFederated learning for intrusion detection in IoT environments: a privacy-preserving strategy
Khraisat, A., Alazab, A., Alazab, M., Obeidat, A., Singh, S., & Jan, T. (2025) · Discover Internet of Things · doi:10.1007/s43926-025-00169-7 · Cited by 98
Machine Learning Threatens 5G Security
Suomalainen, J., Juhola, A., Shahabuddin, S., Mämmelä, A., & Ahmad, I. (2020) · IEEE Access · doi:10.1109/access.2020.3031966 · Cited by 92
“The privacy advantage comes from secure aggregation [164].”
BackgroundA review on federated learning towards image processing
Khokhar, F. A., Shah, J. H., Khan, M. A., Sharif, M. I., Tariq, U., & Kadry, S. (2022) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2022.107818 · Cited by 92
Presotto, R., Civitarese, G., & Bettini, C. (2022) · Personal and Ubiquitous Computing · doi:10.1007/s00779-022-01688-8 · Cited by 91
Владимировна, Ф. Е., Novikova, E. S., & Shulepov, A. (2022) · Algorithms · doi:10.3390/a15070247 · Cited by 91
Liu, J., Huang, G., Hyyppa, J. M., Li, J., Gong, X., & Jiang, X. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.120492 · Cited by 91
PFLM: Privacy-preserving federated learning with membership proof
Jiang, C., Xu, C., & Zhang, Y. (2021) · Information Sciences · doi:10.1016/j.ins.2021.05.077 · Cited by 92
Non-IID data re-balancing at IoT edge with peer-to-peer federated learning for anomaly detection
Wang, H., Muñoz-González, L., Eklund, D., & Raza, S. (2021) · ACM Conference on Security and Privacy in Wireless and Mobile Networks (WiSec) · doi:10.1145/3448300.3467827 · Cited by 91
“Once the central node requests model aggregation, for example using Federated Averaging (FedAvg) [21], selected clients send back their model update in a specific form via assigned protocols as proposed in [5].”
MethodAnomaly detection and defense techniques in federated learning: a comprehensive review
Zhang, C., Yang, S., Mao, L., & Ning, H. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-024-10796-1 · Cited by 92
Shofiullah, S., Dhanekula, A., & Spectra Prime Services, B. A. (2021) · doi:10.63125/k18cbz55 · Cited by 90
Protecting Privacy of Users in Brain-Computer Interface Applications
Agarwal, A., Dowsley, R., McKinney, N. D., Wu, D., Lin, C., Cock, M. D., & Nascimento, A. C. A. (2019) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2019.2926965 · Cited by 92
“We do not cover other approaches for PPML that leak some information, such as: differential privacy [13]; trusted enclaves [14]; federated learning [15], [16].”
BackgroundBatool, H., Anjum, A., Khan, A., Izzo, S., Mazzocca, C., & Jeon, G. (2024) · Information Sciences · doi:10.1016/j.ins.2023.119717 · Cited by 89
Leveraging asynchronous federated learning to predict customers financial distress
Imteaj, A., & Amini, M. H. (2022) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2022.200064 · Cited by 91
Local-Global Knowledge Distillation in Heterogeneous Federated Learning with Non-IID DataPreprint
Yao, D., Pan, W., Dai, Y., Wan, Y., Ding, X., Jin, H., Xu, Z., & Sun, L. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.00051 · Cited by 90
“, 2017), which can train a machine learning model without sharing clients’ private data (Bonawitz et al., 2017), therefore the data privacy will be preserved.”
BackgroundPreserving integrity in online social networks
Halevy, A. Y., Canton-Ferrer, C., Ma, H., Özertem, U., Pantel, P., Saeidi, M., Silvestri, F., & Stoyanov, V. (2022) · Communications of the ACM · doi:10.1145/3462671 · Cited by 89
“In the context of integrity, certain kinds of violations are very rare, so even if we use techniques like secure aggregation [18] to aggregate training data from multiple devices, we may still see entire cohorts of users without positive examples of certain kinds of violations.”
MethodPrivacy-Preserving federated learning in medical diagnosis with homomorphic re-Encryption
Ku, H., Susilo, W., Zhang, Y., Liu, W., & Zhang, M. (2021) · Computer Standards & Interfaces · doi:10.1016/j.csi.2021.103583 · Cited by 89
CodedPaddedFL and CodedSecAgg: Straggler Mitigation and Secure Aggregation in Federated Learning
Schlegel, R., Kumar, S., Rosnes, E., & Amat, A. G. I. (2023) · IEEE Transactions on Communications · doi:10.1109/tcomm.2023.3244243 · Cited by 90
“To prevent such attacks and preserve users’ data privacy, secure aggregation protocols have been proposed [17]–[29] in which the central server only obtains the sum of all the local model updates instead of the local updates directly.”
BackgroundLiu, H., He, F., & Cao, G. (2023) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom53939.2023.10228970 · Cited by 89
“AdaGQ follows the system design of the state-of-the-art FL system [6], [22] by adopting the adaptive and heterogeneous quantization.”
MethodDhakal, S., Prakash, S., Yona, Y., Talwar, S., & Himayat, N. (2019) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps45667.2019.9024521 · Cited by 87
“For example, in [1], [2] federated learning is performed to improve the automatic text prediction capability of Gboard.”
Metadata
- Authors
- Bonawitz, K., Ivanov, V., Kreuter, B., Marcedone, A., McMahan, H. B., Patel, S., Ramage, D., Segal, A., & Seth, K.
- Affiliations
- Google (United States) (US)
- Cornell Tech (US)
- Venue
- Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security
- Publisher
- ACM
- Published
- 2017-10-30
- Volume
- 2017
- Pages
- 1175-1191
- ISSN
- 1543-7221 (ISSN-L)
- OpenAlex
- W2767079719
- Cited by
- 3,956
- References
- 55
- Citations by year
- 2026: 701 · 2025: 712 · 2024: 643 · 2023: 564 · 2022: 400 · 2021: 436 · 2020: 299 · 2019: 152 · 2018: 42 · 2017: 7
- FWCI
- 62.37
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- ACM SIGSAC Conference on Computer and Communications Security (CCS) (publishedVersion)
- IACR Cryptology ePrint Archive (submittedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Cryptography and Data Security
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Computer Science › Information Systems › Cloud Data Security Solutions
- Keywords
- secure aggregation, privacy-preserving machine learning, federated learning, high-dimensional data, communication efficiency, user dropout, honest-but-curious adversary, active adversaries, secure multi-party computation
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
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.