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 301–400 of 1,035
Alsamhi, S. H., Shvetsov, A. V., Kumar, S., Hassan, J. A., Alhartomi, M. A., Shvetsova, S. V., Sahal, R., & Hawbani, A. (2022) · Drones · doi:10.3390/drones6070177 · Cited by 149
Advanced Deep Learning Models for 6G: Overview, Opportunities, and Challenges
Jiao, L., Shao, Y., Sun, L., Liu, F., Yang, S., Ma, W., Li, L., Liu, X., Hou, B., Zhang, X., et al. (2024) · IEEE Access · doi:10.1109/access.2024.3418900 · Cited by 149
Mitigating Backdoor Attacks in Federated Learning
Wu, C., Xian, Y., Zhu, S., & Mitra, P. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.01767 · Cited by 55
“The server does not have access to the data of clients, and thus, cannot verify model updates from clients, especially when the system is augmented with secure aggregation protocols to further protect the privacy of clients [4].”
BackgroundEfficient and Privacy-Preserving Truth Discovery in Mobile Crowd Sensing Systems
Xu, G., Li, H., Liu, S., Wen, M., & Lu, R. (2019) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2019.2895834 · Cited by 142
“Here we omit the specific proof and interested readers can refer to the literature [32] for more relevant details.”
BackgroundMethodInfluentialVahidian, S., Morafah, M., Wang, W., Kungurtsev, V., Chen, C., Shah, M., & Lin, B. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i8.26197 · Cited by 150
A Survey of Graph-Based Deep Learning for Anomaly Detection in Distributed Systems
Pazho, A. D., Noghre, G. A., Purkayastha, A. A., Vempati, J., Martin, O., & Tabkhi, H. (2023) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2023.3282898 · Cited by 145
“It also greatly reduces communication overhead [129].”
FedV: Privacy-Preserving Federated Learning over Vertically Partitioned Data
Xu, R., Baracaldo, N., Zhou, Y., Anwar, A., Joshi, J. B. D., & Ludwig, H. (2021) · ACM Workshop on Artificial Intelligence and Security (AISec) · doi:10.1145/3474369.3486872 · Cited by 143
“while these papers address issues related to privacy and security [5, 7, 20, 22, 35, 36, 41, 43, 50], system architecture [30, 34], and new learning algorithms e.”
BackgroundMethodSatellite Computing: Vision and Challenges
Wang, S., & Li, Q. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3303346 · Cited by 144
No Peek: A Survey of private distributed deep learning
Vepakomma, P., Swedish, T. B., Raskar, R., Gupta, O., & Dubey, A. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1812.03288 · Cited by 42
From distributed machine learning to federated learning: In the view of data privacy and security
Shen, S., Zhu, T., Wu, D., Wang, W., & Zhou, W. (2020) · Concurrency and Computation Practice and Experience · doi:10.1002/cpe.6002 · Cited by 142
PVD-FL: A Privacy-Preserving and Verifiable Decentralized Federated Learning Framework
Zhao, J., Zhu, H., Wang, F., Lu, R., Liu, Z., & Li, H. (2022) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2022.3176191 · Cited by 142
“Based on the double mask method, Bonawitz et al. [9] designed a secure aggregation algorithm to aggregate the local updates without leaking any single local updates, which can protect the local updates from inference attacks, but the global model is still exposed to every participant in plaintext.”
MethodTian, Y., Wang, S., Xiong, J., Bi, R., Zhou, Z., & Bhuiyan, M. Z. A. (2024) · IEEE/ACM Transactions on Computational Biology and Bioinformatics · doi:10.1109/tcbb.2023.3243932 · Cited by 115
“Although edges follow the protocol correctly, they can collude with the remaining t − 2 edges [14], [32] to infer information about others [46], [47].”
BackgroundRingSFL: An Adaptive Split Federated Learning Towards Taming Client Heterogeneity
Shen, J., Cheng, N., Wang, X., Lyu, F., Xu, W., Liu, Z., Aldubaikhy, K., & Shen, X. (2023) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2023.3309633 · Cited by 145
“Cryptographic mechanisms, such as secure aggregation studied in [21], [22], that use an encryption mechanism to prevent any party, including the server, from accessing each client’s uploaded model, are becoming more mainstream.”
BackgroundProvably Secure Federated Learning against Malicious Clients
Cao, X., Jia, J., & Gong, N. Z. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i8.16849 · Cited by 141
ReFRS: Resource-efficient Federated Recommender System for Dynamic and Diversified User Preferences
Imran, M., Yin, H., Chen, T., Nguyen, Q. V. H., Zhou, A., & Zheng, K. (2023) · ACM Transactions on Information Systems · doi:10.1145/3560486 · Cited by 142
“Horizontal FL is suitable for a scenario where the items that are interacted by the users overlap but not the users themselves [9, 74].”
BackgroundDifferentially Private Federated Learning: A Systematic Review
Fu, J., Hong, Y., Ling, X., Wang, L., Ran, X., Sun, Z., Wang, W. H., Chen, Z., & Cao, Y. (2026) · ACM Computing Surveys · doi:10.1145/3801079 · Cited by 148
LightSecAgg: a Lightweight and Versatile Design for Secure Aggregation in Federated Learning
So, J., He, C., Yang, C.-S., Li, S., Yu, Q., Ali, R. E., Güler, B., & Avestimehr, S. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2109.14236 · Cited by 58
“We consider the worst-case scenario in [4] where the selected pN users artificially drop after uploading the masked model.”
BackgroundMethodInfluentialHybrid Blockchain-Based Resource Trading System for Federated Learning in Edge Computing
Fan, S., Zhang, H., Zeng, Y., & Cai, W. (2020) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.3028101 · Cited by 138
Decentralised Learning in Federated Deployment Environments: A System-Level Survey
Bellavista, P., Foschini, L., & Mora, A. (2022) · ACM Computing Surveys · doi:10.1145/3429252 · Cited by 99
“Furthermore, other aspects in conjugating privacy and communication efficiency emerge; the secure aggregation protocol [13] can be redesigned to account from the beginning for communication efficiency [14], while tailored DP-mechanisms can be more amenable to privacy analysis when quantisation of…”
BackgroundMethodInfluentialFederated learning in intrusion detection: advancements, applications, and future directions
Büyüktanır, B., Altınkaya, Ş., Baydoğmus, G. K., & Yıldız, K. (2025) · Cluster Computing · doi:10.1007/s10586-025-05325-w · Cited by 141
“Cryptographic precautions include privacy-preserving machine learning algorithms such as secure multi-party computation (SMC) [57, 58] and homomorphic encryption [59, 60].”
The Skellam Mechanism for Differentially Private Federated Learning
Agarwal, N., Kairouz, P., & Liu, Z. (2021) · neural information processing systems · doi:10.48550/arxiv.2110.04995 · Cited by 139
“Moreover, it cannot be used in distributed learning settings with cryptographic multi-party computation primitives involving modular arithmetic, such as secure aggregation [12, 11].”
BackgroundMethodCrowdFL: Privacy-Preserving Mobile Crowdsensing System Via Federated Learning
Zhao, B., Liu, X., Chen, W., & Deng, R. H. (2022) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2022.3157603 · Cited by 136
“A reasonable method is to encrypt training models and only aggregate encrypted training models [12], [34], [51].”
MethodA Survey of Trustworthy Federated Learning: Issues, Solutions, and Challenges
Zhang, Y., Zeng, D., Luo, J., Fu, X., Chen, G., Xu, Z., & King, I. (2024) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3678181 · Cited by 139
“Currently, various meticulously designed MM protocols are extensively employed in FL to ensure the secure aggregation of local gradients [14, 22, 148, 200, 201, 208].”
Privacy-Preserving Deep Learning via Weight Transmission
Phong, L. T., & Phuong, T. T. (2019) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2019.2911169 · Cited by 114
“[21] proposed a secure aggregation method and applied it to deep neural networks to aggregate user-provided model updates.”
BackgroundMethodRahman, A., Hasan, K., Kundu, D., Islam, M. J., Debnath, T., Band, S. S., & Kumar, N. (2022) · Future Generation Computer Systems · doi:10.1016/j.future.2022.08.004 · Cited by 135
Meta-HAR: Federated Representation Learning for Human Activity Recognition
Li, C., Niu, D., Jiang, B., Zuo, X., & Yang, J. (2021) · Proceedings of the Web Conference 2021 · doi:10.1145/3442381.3450006 · Cited by 133
“[6] allows a server to sum up large vectors from mobile devices in a secure manner through a communication-”
BackgroundShredder: Learning Noise Distributions to Protect Inference Privacy
Mireshghallah, F., Taram, M., Ramrakhyani, P., Jalali, A., Tullsen, D. M., & Esmaeilzadeh, H. (2020) · International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) · doi:10.1145/3373376.3378522 · Cited by 139
“[57], for instance, introduces a privacy-preserving protocol for federated learning.”
BackgroundDuan, Q., Hu, S., Deng, R., & Lu, Z. (2022) · Sensors · doi:10.3390/s22165983 · Cited by 108
“For example, an SMC-based protocol for privacy-preserving model aggregation in FL was proposed in [45].”
A Study of Face Obfuscation in ImageNet
Yang, K., Yau, J., Fei-Fei, L., Deng, J. H., & Russakovsky, O. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.06191 · Cited by 64
“…possible for multiple data owners to train a model collectively without sharing their private data using federated learning (McMahan et al., 2017; Bonawitz et al., 2017; Li et al., 2020) or secure multi-party computation (Shokri & Shmatikov, 2015; Melis et al., 2019; Hamm et al., 2016; Pathak et…”
BackgroundHUMAN‐CENTERED DESIGN OF ARTIFICIAL INTELLIGENCE
Margetis, G., Ntoa, S., Antona, M., & Stephanidis, C. (2021) · HANDBOOK OF HUMAN FACTORS AND ERGONOMICS · doi:10.1002/9781119636113.ch42 · Cited by 106
REFL: Resource-Efficient Federated Learning
Abdelmoniem, A. M., Sahu, A. N., Canini, M., & Fahmy, S. A. (2023) · European Conference on Computer Systems · doi:10.1145/3552326.3567485 · Cited by 133
“A key assumption in FL is a lack of trust, implying training data should not leave the data source [7], and any possible breach of private data during communication should be avoided [8].”
BackgroundA Survey on Securing Federated Learning: Analysis of Applications, Attacks, Challenges, and Trends
Neto, H. N. C., Hribar, J., Dusparić, I., Mattos, D. M. F., & Fernandes, N. C. (2023) · IEEE Access · doi:10.1109/access.2023.3269980 · Cited by 134
“The most used techniques to protect the parameters of the model are homomorphic cryptography [41] and SMC [42].”
MethodPrivacy-preserving in Blockchain-based Federated Learning systems
Sameera, K. M., Nicolazzo, S., Arazzi, M., Nocera, A., A., R. R. K., Vinod, P., & Conti, M. (2024) · Computer Communications · doi:10.1016/j.comcom.2024.04.024 · Cited by 133
“To tackle this issue, multiple studies, exemplified by [110] and [111], advocate for the adoption of the SecAgg protocol [112].”
BackgroundNguyen, T. D., Nguyen, T., Nguyen, P. L., Pham, H. H., Doan, K. D., & Wong, K.-S. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107166 · Cited by 130
Topology-aware Federated Learning in Edge Computing: A Comprehensive Survey
Wu, J., Dong, F., Leung, H., Zhu, Z., Zhou, J., & Drew, S. (2024) · ACM Computing Surveys · doi:10.1145/3659205 · Cited by 130
“The required resource of CCESA is reduced by a factor of at least 𝑂 ( √︁ 𝑛 / log 𝑛 ) compared to [5].”
BackgroundInfluentialBrauneck, A., Schmalhorst, L., Kazemi Majdabadi, M. M., Bakhtiari, M., Völker, U., Baumbach, J., Baumbach, L., & Buchholtz, G. (2023) · Journal of Medical Internet Research · doi:10.2196/41588 · Cited by 130
Learn to Forget: Machine Unlearning via Neuron Masking
Ma, Z., Liu, Y., Liu, X., Liu, J., Ma, J., & Ren, K. (2022) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2022.3194884 · Cited by 130
Decentral and Incentivized Federated Learning Frameworks: A Systematic Literature Review
Witt, L., Heyer, M., Toyoda, K., Samek, W., & Li, D. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3231363 · Cited by 101
“(iii) While Yao’s garbled circuits were developed for 2party secure computing, Homomorphic Encryption (HE) allows for higher numbers of participants [91].”
BackgroundMethodInfluentialBlockchain-Based Federated Learning for Intelligent Control in Heavy Haul Railway
Hua, G., Zhu, L., Wu, J., Shen, C., Zhou, L., & Lin, Q. (2020) · IEEE Access · doi:10.1109/access.2020.3021253 · Cited by 115
AI in healthcare: navigating opportunities and challenges in digital communication
Sun, G., & Zhou, Y.-H. (2023) · Frontiers in Digital Health · doi:10.3389/fdgth.2023.1291132 · Cited by 86
Federated Domain Generalization: A Survey
Li, Y., Wang, X., Zeng, R., Donta, P. K., Murturi, I., Huang, M., & Dustdar, S. (2025) · Proceedings of the IEEE · doi:10.1109/jproc.2025.3596173 · Cited by 131
“Here are some approaches related to privacy-preserving FDG: Secure aggregation [218], differential privacy [219], privacy-aware model selection [220], privacy-preserving data preprocessing [221], and privacy regulations [222].”
Trustworthy Distributed AI Systems: Robustness, Privacy, and Governance
Wei, W., & Liu, L. (2025) · ACM Computing Surveys · doi:10.1145/3645102 · Cited by 128
“[147] is a subfield of cryptography that allows creating methods to jointly compute a function using inputs from different parties without revealing those inputs neither to each other 143:21 nor to the central server [148].”
Challenges and future directions of secure federated learning: a survey
Zhang, K., Song, X., Zhang, C., & Yu, S. (2022) · Frontiers of Computer Science · doi:10.1007/s11704-021-0598-z · Cited by 127
Securing Secure Aggregation: Mitigating Multi-Round Privacy Leakage in Federated Learning
So, J., E. Ali, R., Güler, B., Jiao, J., & Avestimehr, A. S. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i8.26177 · Cited by 79
Liu, C., Tian, W., & Kan, C. (2022) · Journal of Manufacturing Systems · doi:10.1016/j.jmsy.2022.04.010 · Cited by 127
Preserving Privacy and Security in Federated Learning
Nguyen, T., & Thai, M. T. (2023) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2023.3302016 · Cited by 129
“Recent research on federated learning focuses on either improving the privacy of model updates [2], [7], [30] or preventing certain poisoning attacks from malicious users [3], [28], [29], but not both.”
BackgroundMethodResultInfluentialAccelerating Decentralized Federated Learning in Heterogeneous Edge Computing
Wang, L., Xu, Y., Xu, H., Chen, M., & Huang, L. (2022) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2022.3178378 · Cited by 126
“Each worker only needs to exchange models with its neighbors rather than sharing its original data, and some techniques, e.g., secure aggregation [28] and differential privacy [29], can be adopted to preserve the privacy of transmitted models or gradients better.”
BackgroundCommunication-Efficient Federated Learning with Adaptive Quantization
Mao, Y., Zhao, Z., Yan, G., Liu, Y., Lan, T., Song, L., & Ding, W. (2022) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3510587 · Cited by 127
“Thus, the robustness to client dropout is another practical requirement for FL systems [5].”
BackgroundInfluentialFederated Graph Neural Network for Fast Anomaly Detection in Controller Area Networks
Zhang, H., Zeng, K., & Lin, S. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3240291 · Cited by 128
“Privacy and security in federated learning have recently drawn increasing attention [51], [52], [53].”
BackgroundVision-Language Models for Edge Networks: A Comprehensive Survey
Sharshar, A., Khan, L. U., Ullah, W., & Guizani, M. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3579032 · Cited by 127
“Blockchain ensures tamper resistance and transparency [110].”
InfluentialRahmati, M., & Pagano, A. (2025) · Informatics · doi:10.3390/informatics12030062 · Cited by 125
“This reduction in resource consumption is consistent with recent findings on energy-efficient edge AI systems [3,4,29].”
InfluentialFedSpace: An Efficient Federated Learning Framework at Satellites and Ground StationsPreprint
So, J., Hsieh, K., Arzani, B., Noghabi, S. A., Avestimehr, S., & Chandra, R. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2202.01267 · Cited by 127
“The vast majority of existing FL algorithms assumes synchronous FL (e.g., McMahan et al. 2017; Bonawitz et al. 2017; 2019; Kairouz et al. 2021), where all local updates are based on the same global model.”
BackgroundMethodA Differentially Private Federated Learning Model against Poisoning Attacks in Edge Computing
Zhou, J., Wu, N., Wang, Y., Gu, S., Cao, Z., Dong, X., & Choo, K. R. (2022) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2022.3168556 · Cited by 124
“Cryptographic primitives [24], [25], [26], [27], [28], [29], [30], [31], [32], and Differential Privacy [33], [34], [35], [36], [37], [38], [39], can be used to encrypt data and parameters in order to protect the privacy of the model training process.”
MethodThe Evolution of Federated Learning-Based Intrusion Detection and Mitigation: A Survey
Lavaur, L., Pahl, M., Busnel, Y., & Autrel, F. (2022) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2022.3177512 · Cited by 125
Privacy-Preserving Federated Learning With Malicious Clients and Honest-but-Curious Servers
Le, J., Zhang, D., Lei, X., Jiao, L., Zeng, K., & Liao, X. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3295949 · Cited by 125
“In recent years, numerous researchers have studied the aforementioned security issues of FL and proposed various defense solutions such as the Secure Multiparty Computation (SMC)-based schemes [13], the Homomorphic Encryption (HE)-based schemes [14], the Differential Privacy (DP)-based schemes [15], [16], [17], the by…”
BackgroundMethodResultInfluentialMaskCrypt: Federated Learning With Selective Homomorphic Encryption
Hu, C., & Li, B. (2024) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2024.3392424 · Cited by 125
“Secure aggregation [23] is another choice that allows a server to compute the sum of data vectors from clients in a secure manner, i.e., the server learns nothing but the aggregation result.”
BackgroundA Framework for Collaborative Learning in Secure High-Dimensional Space
Imani, M., Kim, Y., Riazi, S., Messerly, J., Liu, P., Koushanfar, F., & Rosing, T. (2019) · 2019 IEEE 12th International Conference on Cloud Computing (CLOUD) · doi:10.1109/cloud.2019.00076 · Cited by 99
“Google has also proposed a federated learning approach [17] for collaborative learning.”
BackgroundInfluentialA Comprehensive Empirical Study of Heterogeneity in Federated Learning
Abdelmoniem, A. M., Ho, C.-Y., Papageorgiou, P., & Canini, M. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3250275 · Cited by 126
“A common assumption is that the learners are honest whereas the server is honest but curious [25].”
BackgroundMethodSecure and Trustworthy Artificial Intelligence-extended Reality (AI-XR) for Metaverses
Qayyum, A., Butt, M. A., Ali, H. M., Usman, M., Halabi, O., Al‐Fuqaha, A., Abbasi, Q. H., Imran, M. A., & Qadir, J. (2024) · ACM Computing Surveys · doi:10.1145/3614426 · Cited by 127
Fouda, M. M., Fadlullah, Z. M., Ibrahem, M. I., & Kato, N. (2024) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3486690 · Cited by 123
“To address this issue, customized SMPC protocols have been developed that incorporate certain sections of the ML and FL training algo-rithm to improve privacy, such as in the case of FL where safe aggregation of weight updates can mitigate the risk of sensitive data exposure [137], [138], [139].”
Zheng, H., Hu, H., & Han, Z. (2020) · IEEE Intelligent Systems · doi:10.1109/mis.2020.3010335 · Cited by 96
“…FML for a loose federation of multiple mobile clients with scalable design 5 and develops secure aggregation using encryption scheme such as MPC. 13 As for the system aspects, the architecture of federated learning discussed in this article is horizontal design, 19 which enables easy unification…”
MethodUser-Level Label Leakage from Gradients in Federated Learning
Wainakh, A., Ventola, F., Müßig, T., Keim, J., Cordero, C. G., Zimmer, E., Grube, T., Kersting, K., & Mühlhäuser, M. (2022) · Proceedings on Privacy Enhancing Technologies · doi:10.2478/popets-2022-0043 · Cited by 120
“The server also can access the gradients of an individual user, in case no secure aggregation [6] or other protection techniques are used.”
MethodGradient Inversion with Generative Image PriorPreprint
Jeon, J., Kim, J., Lee, K., Oh, S., & Ok, J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.14962 · Cited by 121
Smart Wearables for the Detection of Cardiovascular Diseases: A Systematic Literature Review
Moshawrab, M., Adda, M., Bouzouane, A., Ibrahim, H., & Raad, A. (2023) · Sensors · doi:10.3390/s23020828 · Cited by 120
Personalized Federated Learning by Structured and Unstructured Pruning under Data Heterogeneity
Vahidian, S., Morafah, M., & Lin, B. (2021) · IEEE International Conference on Distributed Computing Systems Workshops (ICDCSW) · doi:10.1109/icdcsw53096.2021.00012 · Cited by 120
GIANT: Globally Improved Approximate Newton Method for Distributed OptimizationPreprint
Wang, S., Roosta, F., Xu, P., & Mahoney, M. W. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1709.03528 · Cited by 123
“The desire to reduce communication costs is even more pronounced in the federated learning framework [8, 9, 1, 18, 37].”
BackgroundDeepPAR and DeepDPA: Privacy Preserving and Asynchronous Deep Learning for Industrial IoT
Zhang, X., Chen, X., Liu, J. K., & Xiang, Y. (2020) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2019.2941244 · Cited by 117
“[2] proposed a secure data aggregation method for privacy preserving machine learning by leveraging secret sharing, which supports any user revocation.”
MethodData Protection in AI Services: A Survey
Meurisch, C., & Mühlhäuser, M. (2021) · ACM Computing Surveys · doi:10.1145/3440754 · Cited by 120
“Secure aggregation for FL [23] 2017 / / 26”
BackgroundInfluentialA Secure Federated Learning framework using Homomorphic Encryption and Verifiable Computing
Madi, A., Stan, O., Mayoue, A., Sébert, A. G., Gouy‐Pailler, C., & Sirdey, R. (2021) · doi:10.1109/rdaaps48126.2021.9452005 · Cited by 120
“In this context, there have been recent proposals to improve data privacy and model confidentiality through emerging cryptographic techniques such as Homomorphic Encryption [7], [8] or Multi-Party Computation [9], [10].”
MethodPrivate Retrieval, Computing, and Learning: Recent Progress and Future Challenges
Ulukuş, Ş., Avestimehr, S., Gastpar, M., Jafar, S. A., Tandon, R., & Tian, C. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3142358 · Cited by 120
Mandal, K., & Gong, G. (2019) · Proceedings of the 2019 ACM SIGSAC Conference on Cloud Computing Security Workshop · doi:10.1145/3338466.3358926 · Cited by 89
“No complete training protocol for any machine learning algorithm is developed in [7, 44].”
BackgroundMethodResultInfluentialWang, C., Yang, Y., & Zhou, P. (2020) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2020.3023905 · Cited by 119
Active Learning Based Federated Learning for Waste and Natural Disaster Image Classification
Ahmed, L., Ahmad, K., Said, N., Qolomany, B., Qadir, J., & Al‐Fuqaha, A. (2020) · IEEE Access · doi:10.1109/access.2020.3038676 · Cited by 118
“In [25], a Secure Multiparty Computation (SMC) protocol is developed to secure local model updates from the server where the server can only aggregate the local models.”
MethodFLOP: Federated Learning on Medical Datasets using Partial Networks
Yang, Q., Zhang, J., Hao, W., Spell, G. P., & Carin, L. (2021) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3447548.3467185 · Cited by 120
“Furthermore, we can achieve guaranteed privacy by masking a selection of gradients with encryption [3], differential privacy [25], or secret sharing [4] techniques in step 1, which is out of the scope of this paper.”
BackgroundMethodHeteroSAg: Secure Aggregation With Heterogeneous Quantization in Federated Learning
Elkordy, A. R., & Avestimehr, A. S. (2022) · IEEE Transactions on Communications · doi:10.1109/tcomm.2022.3151126 · Cited by 118
“Hence, this expansion makes them ineffective with aggressive quantization, specially for large N [1,2,25].”
BackgroundMethodInfluentialPodschwadt, R., Takabi, D., Hu, P., Rafiei, M. H., & Cai, Z. (2022) · IEEE Access · doi:10.1109/access.2022.3219049 · Cited by 119
“A central authority can aggregate the gradients [16], or the involved parties can each collect the gradient updates [17].”
BackgroundFederated learning for 6G-enabled secure communication systems: a comprehensive survey
Sirohi, D., Kumar, N., Rana, P. S., Tanwar, S., Iqbal, R., & Hijji, M. (2023) · Artificial Intelligence Review · doi:10.1007/s10462-023-10417-3 · Cited by 118
A survey on federated learning: a perspective from multi-party computation
Liu, F., Zheng, Z., Shi, Y., Tong, Y., & Zhang, Y. (2023) · Frontiers of Computer Science · doi:10.1007/s11704-023-3282-7 · Cited by 118
Attack-Resistant Federated Learning with Residual-based Reweighting
Fu, S., Xie, C., Li, B., & Chen, Q. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.11464 · Cited by 54
“(Bonawitz et al. 2017) propose a privacy-preserving protocol for model aggregation in federated learning.”
BackgroundInfluentialThe distributed discrete gaussian mechanism for federated learning with secure aggregation
Kairouz, P., Liu, Z., & Steinke, T. (2021) · arXiv · doi:10.48550/arxiv.2102.06387 · Cited by 117
“…(McMahan et al., 2017; Kairouz et al., 2019), cryptographic secure aggregation (SecAgg) to prevent a server from inspecting individual user updates (Bonawitz et al., 2017; Bell et al., 2020), and differentially private stochastic gradient descent (DP-SGD) to train models with provably limited…”
BackgroundA Survey of Trustworthy Federated Learning with Perspectives on Security, Robustness and Privacy
Zhang, Y., Zeng, D., Luo, J., Xu, Z., & King, I. (2023) · International World Wide Web Conference (WWW) · doi:10.1145/3543873.3587681 · Cited by 117
“Attack [33, 103, 104, 113, 118] • Cryptography-based Methods [5, 8, 14, 22, 25, 58, 98, 105, 109, 115] Deployment Security”
BackgroundMethodApplications of Distributed Machine Learning for the Internet-of-Things: A Comprehensive Survey
Le, M., Huynh‐The, T., Do‐Duy, T., Vu, T.-H., Hwang, W., & Pham, Q. (2024) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3427324 · Cited by 122
“For securing the individual devices’ updates against threats within the centralized server, the secure Aggregation protocol [122] is implemented as a privacy preservation to the FL approach.”
Throughput-Optimal Topology Design for Cross-Silo Federated Learning
Marfoq, O., Xu, C., Neglia, G., & Vidal, R. (2020) · neural information processing systems · doi:10.48550/arxiv.2010.12229 · Cited by 118
“Synchronous optimization algorithms are often preferred for federated learning [9], because they enjoy stronger convergence guarantees than their asynchronous counterparts and can be easily combined with cryptographic secure aggregation protocols [8], differential privacy techniques [1], and model and update compressi…”
BackgroundSocInf: Membership Inference Attacks on Social Media Health Data With Machine Learning
Liu, G., Wang, C., Peng, K., Huang, H., Li, Y., & Cheng, W. (2019) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2019.2916086 · Cited by 95
“protocol over high-dimensional data was proposed in [46],”
BackgroundEaSTFLy: Efficient and secure ternary federated learning
Dong, Y., Chen, X., Shen, L., & Wang, D. (2020) · Computers & Security · doi:10.1016/j.cose.2020.101824 · Cited by 103
Tran, A.-T., Luong, T. D., Karnjana, J., & Huynh, V. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.10.014 · Cited by 117
Protect Privacy from Gradient Leakage Attack in Federated Learning
Wang, J., Guo, S., Xie, X. N., & Qi, H. (2022) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom48880.2022.9796841 · Cited by 116
“[17] present a secure aggregation and ensure that the server can only learn the aggregated data instead of the individual data from any clients.”
BackgroundPrivacy-Preserving and Traceable Federated Learning for data sharing in industrial IoT applications
Chen, J., Xue, J., Wang, Y., Huang, L., Baker, T., & Zhou, Z. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.119036 · Cited by 116
Artificial General Intelligence for Medical Imaging Analysis
Li, X., Zhao, L., Zhang, L., Wu, Z., Liu, Z., Jiang, H., Cao, C., Xu, S., Li, Y., Dai, H., et al. (2024) · IEEE Reviews in Biomedical Engineering · doi:10.1109/rbme.2024.3493775 · Cited by 116
“Federated learning [152] is an innovative machine learning approach that addresses privacy, data ownership, and scalability challenges by enabling the training of models across decentralized devices or data sources.”
Multiple Diseases and Pests Detection Based on Federated Learning and Improved Faster R-CNN
Deng, F., Mao, W., Zeng, Z., Zeng, H., & Wei, B. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3201937 · Cited by 119
“2) Improving data privacy protection, in 2017, Bonawitz et al. [23] proposed a secure multiparty computation (MPC) scheme in FL, where the server can only see the gradient after the aggregation is completed and cannot know the private true gradient of each user, but can still observe the exact…”
BackgroundMalik, H., Naeem, A., Naqvi, R. A., & Loh, W.-K. (2023) · Sensors · doi:10.3390/s23020743 · Cited by 115
“The problem of inconsistent data must be solved.”
BackgroundPan, Y., Chao, Z., He, W., Yang, J., Li, H., & Liming, W. (2024) · Cybersecurity · doi:10.1186/s42400-024-00232-w · Cited by 115
“Bonawitz et al. (2017) designed a secure aggregation protocol for FL that can tolerate client outages using blinding with random values, Shamir’s Secret Sharing, and symmetric encryption.”
MethodReviewing Federated Machine Learning and Its Use in Diseases Prediction
Moshawrab, M., Adda, M., Bouzouane, A., Ibrahim, H., & Raad, A. (2023) · Sensors · doi:10.3390/s23042112 · Cited by 114
“To this end, they created three models, including standard, FedAVG, and FedProx, and evaluated their models with five publicly available diabetic retinopathy datasets, including EyePACS [123], Messidor [124], IDRID [125], APTOS [126], and College of Auckland (UoA) [127].”
BackgroundMemory-aware curriculum federated learning for breast cancer classification
Jiménez-Sánchez, A., Tardy, M., González Ballester, M. A., Mateus, D., & Piella, G. (2023) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2022.107318 · Cited by 114
“To address the second challenge, data protection, cryptographic techniques [4] or differential privacy [5], [6] are employed.”
MethodBonawitz, K., Kairouz, P., McMahan, B., & Ramage, D. (2021) · Queue · doi:10.1145/3494834.3500240 · Cited by 69
Self-citationFedCVT: Semi-supervised Vertical Federated Learning with Cross-view Training
Kang, Y., Liu, Y., & Liang, X. (2022) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3510031 · Cited by 113
Quantum Federated Learning With Decentralized Data
Huang, R., Tan, X., & Xu, Q. (2022) · IEEE Journal of Selected Topics in Quantum Electronics · doi:10.1109/jstqe.2022.3170150 · Cited by 116
Privacy Inference Attack and Defense in Centralized and Federated Learning: A Comprehensive Survey
Rao, B., Zhang, J., Wu, D., Zhu, C., Sun, X., & Chen, B. (2024) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2024.3363670 · Cited by 115
Defending Model Inversion and Membership Inference Attacks via Prediction PurificationPreprint
Yang, Z., Shao, B., Xuan, B., Chang, E., & Zhang, F. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2005.03915 · Cited by 115
PEFL: A Privacy-Enhanced Federated Learning Scheme for Big Data Analytics
Zhang, J., Chen, B., Yu, S., & Deng, H. (2019) · 2019 IEEE Global Communications Conference (GLOBECOM) · doi:10.1109/globecom38437.2019.9014272 · Cited by 79
“…FL can provide a basic privacy guarantee for each participant’s raw data, a user’s private data is possibly leaked to an untrusted server even with a small portion of the uploaded parameters [7], such as model inversion attack and gradient inference attack as described in [8] and [9], respectively.”
Background
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.