Our Data, Ourselves: Privacy Via Distributed Noise Generation
The DOI matches its Crossref record; 29 of the 31 references it cites were found in a registry or an index.
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
“Differential privacy (DP) [26] is an alternative theoretical model for protecting the privacy of individual data, which has been widely applied to many areas, not only traditional algorithms, e.g., boosting [27], principal component analysis [15], support vector machine [86], but also deep learning research [1, 70]. It ensures that the addition or removal does not substantially affect the outcome”
From Federated Learning for Healthcare Informatics (2021) · cites this paper as [26]
In the citing paper
“The simplest way to create a DP-mechanism is by adding Gaussian noise to the outcomes of a deterministic function with bounded sensitivity 21 .”
From Federated Learning and Differential Privacy for Medical Image Analysis (2021) · cites this paper as [21]
Citations
Cites
Calibrating Noise to Sensitivity in Private Data Analysis
Dwork, C., McSherry, F., Nissim, K., & Smith, A. (2006) · Lecture Notes in Computer Science · doi:10.1007/11681878_14 · Cited by 7,272
Self-citationThe Byzantine generals problem
Lamport, L., Shostak, R., & Pease, M. (1982) · ACM Transactions on Programming Languages and Systems · doi:10.1145/357172.357176 · Cited by 6,023 · cited 1×
“We will refer to several well-known primitive building blocks for constructing distributed protocols: Byzantine Agreement [20], Distributed Coin Flipping [22], Verifiable Secret Sharing (VSS) [8], Non-Malleable VSS, and Secure Function Evaluation (SFE) [18].”
Goldreich, O., Micali, S., & Wigderson, A. (1987) · Proceedings of the nineteenth annual ACM conference on Theory of computing - STOC '87 · doi:10.1145/28395.28420 · Cited by 3,590
Protocols for secure computations
Yao, A. C. (1982) · 23rd Annual Symposium on Foundations of Computer Science (sfcs 1982) · doi:10.1109/sfcs.1982.38 · Cited by 3,306
Privacy-preserving data mining
Agrawal, R., & Srikant, R. (2000) · ACM SIGMOD Record · doi:10.1145/335191.335438 · Cited by 2,961
Protocols for secure computations
Yao, A. C.-C. (1982) · doi:10.1109/sfcs.1982.88 · Cited by 2,661
Goldreich, O. (2001) · Cambridge University Press eBooks · doi:10.1017/cbo9780511546891 · Cited by 2,181
Goldreich, O. (2004) · Cambridge University Press eBooks · doi:10.1017/cbo9780511721656 · Cited by 1,805
Foundations of Cryptography: Volume 2, Basic Applications
Goldreich, O. (2004) · Cited by 1,592
On the design and quantification of privacy preserving data mining algorithms
Agrawal, D., & Aggarwal, C. C. (2001) · Proceedings of the twentieth ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems · doi:10.1145/375551.375602 · Cited by 1,006 · cited 1×
“accurate statistics about a population while preserving the privacy of individuals [1, 2, 15, 11, 14, 5, 6, 4, 13].”
Cited by
Items 301–400 of 1,008
Secure Two-Party Differentially Private Data Release for Vertically Partitioned Data
Mohammed, N., Alhadidi, D., Fung, B. C. M., & Debbabi, M. (2013) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2013.22 · Cited by 81
“Distributed interactive approach.”
Benchmarking Differential Privacy and Federated Learning for BERT ModelsPreprint
Basu, P., Roy, T. S., Naidu, R., Müftüoğlu, Z., Singh, S., & Mireshghallah, F. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.13973 · Cited by 78
“Differential privacy (Dwork, 2011b; Dwork et al., 2006) which will be explored further in the next section, uses random noise to provide that the publicly visible information doesn’t change much if one individual in the dataset is re- ar X iv :2 10 6.”
BackgroundHard to Forget: Poisoning Attacks on Certified Machine Unlearning
Marchant, N. G., Rubinstein, B. I. P., & Alfeld, S. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i7.20736 · Cited by 64
“They identified a connection to differential privacy (Dwork et al. 2006) and suggested adversarial robustness as a direction for future work.”
InfluentialSimultaneous Private Learning of Multiple Concepts
Bun, M., Kobbi Nissim, B., & Stemmer, U. (2016) · Conference on Innovations in Theoretical Computer Science · doi:10.1145/2840728.2840747 · Cited by 92
“Differential privacy was defined in [14] and the relaxation to approximate differential privacy is from [13].”
BackgroundSAQE: practical privacy-preserving approximate query processing for data federations
Bater, J., Park, Y., He, X. Q., Wang, X. S., & Rogers, J. (2020) · Proceedings of the VLDB Endowment · doi:10.14778/3407790.3407854 · Cited by 87
“Finally, after query processing, the system adds carefully calibrated noise to the query results using differential privacy [22] such that an adversary viewing this output will get essentially the same answer whether or not an individual chose to make his or her data available for querying.”
BackgroundMethodRmind: A Tool for Cryptographically Secure Statistical Analysis
Bogdanov, D., Kamm, L., Laur, S., & Sokk, V. (2016) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2016.2587623 · Cited by 75
“Differential privacy guarantees security against attackers with unbounded background knowledge, provided that the result parties are willing to accept a potentially unbounded amount of noise added to results [44].”
BackgroundOptimal Task Recommendation for Mobile Crowdsourcing With Privacy Control
Gong, Y., Wei, L., Guo, Y., Zhang, C., & Fang, Y. M. (2015) · IEEE Internet of Things Journal · doi:10.1109/jiot.2015.2512282 · Cited by 64
“Hence, we should protect every worker from an adversary (with arbitrary background knowledge) who tries to trace or de-anonymize a user between several runs of the statistics collection approach.”
BackgroundDifferentially Private Consensus With Quantized Communication
Gao, L., Deng, S., Ren, W., & Hu, C. (2019) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2018.2890645 · Cited by 64
Binarizing Split Learning for Data Privacy Enhancement and Computation Reduction
Pham, N. D., Abuadbba, A., Gao, Y., Phan, K. T., & Chilamkurti, N. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3274391 · Cited by 78
“Regarding the definition of approximate DP, (ε, δ)-DP [27], deterministic binarization could provide ε-DP with probability (1 − exp(−ε/2)).”
BackgroundCloud-Enabled Differentially Private Multiagent Optimization With Constraints
Hale, M., & Egerstedt, M. (2017) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2017.2751458 · Cited by 67
“For this reason, ( , δ)-differential privacy can be regarded as a δ-approximate form of -differential privacy [38].”
BackgroundDo Not Let Privacy Overbill Utility: Gradient Embedding Perturbation for Private LearningPreprint
Yu, D., Zhang, H., Chen, W., & Liu, T. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.12677 · Cited by 114
“Definition 1 ( ( (cid:15), δ ) -DP (Dwork et al., 2006a;b)) .”
Differentially Private Fair Learning
Jagielski, M., Kearns, M., Mao, J., Oprea, A., Roth, A., Sharifi-Malvajerdi, S., & Ullman, J. (2018) · arXiv · doi:10.48550/arxiv.1812.02696 · Cited by 62
The structure of optimal private tests for simple hypotheses
Canonne, C. L., Kamath, G., McMillan, A., Smith, A., & Ullman, J. (2019) · ACM Symposium on Theory of Computing (STOC) · doi:10.1145/3313276.3316336 · Cited by 99
“For hypothesis tests with constant error probabilities, sample complexity bounds for differential privacy are equivalent, up to constant factors, to sample complexity bounds for other notions of distributional algorithmic stability, such as (ε,δ )-DP [31], concentrated DP [14, 34], KL- and TV-stability [9, 77] (see [3…”
BackgroundPrivacy-Preserving Efficient Federated-Learning Model Debugging
Li, A., Zhang, L., Wang, J., Han, F., & Li, X. (2021) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2021.3137321 · Cited by 68
“…domain X and range R satisfies ( (cid:2); d )-DP if for any subset of outputs S and for any two adjacent databases Instead of the original definition of (cid:2) -DP, we use the variant [32], which allows for the possibility that plain (cid:2) -differ- ential privacy is broken with probability d .”
MethodZheleva, E., Terzi, E., & Getoor, L. C. (2012) · Synthesis lectures on data mining and knowledge discovery · doi:10.2200/s00408ed1v01y201203dmk004 · Cited by 61
Privacy-Preserving Crowdsourced Spectrum Sensing
Jin, X., & Zhang, Y. (2018) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2018.2823272 · Cited by 60
Secure Federated Averaging Algorithm with Differential Privacy
Li, Y., Chang, T., & Chi, C. (2020) · Machine learning for signal processing ... · doi:10.1109/mlsp49062.2020.9231531 · Cited by 75
“One of the approaches to prevent differential attacks from breaching privacy is differential privacy (DP) [6], which can protect the privacy even when the PS/adversaries have full knowledge of the training mechanism and access the model parameters [7].”
BackgroundThe Skellam Mechanism for Differentially Private Federated LearningPreprint
Agarwal, N., Kairouz, P., & Liu, Z. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.04995 · Cited by 138
“The closest mechanisms to Skellam are the binomial [2, 19] and the discrete Gaussian mechanisms [15, 25].”
BackgroundAn Axiomatic View of Statistical Privacy and Utility
Kifer, D., & Lin, B.-R. (2012) · Journal of Privacy and Confidentiality · doi:10.29012/jpc.v4i1.610 · Cited by 59
“The corresponding curves for ( , δ)-differential privacy [16, 38] are shown in Figure 1.”
BackgroundInfluentialPrivacy Loss Classes: The Central Limit Theorem in Differential Privacy
Sommer, D., Meiser, S., & Mohammadi, E. (2019) · Proceedings on Privacy Enhancing Technologies · doi:10.2478/popets-2019-0029 · Cited by 104
Not All Features Are Equal: Discovering Essential Features for Preserving Prediction Privacy
Mireshghallah, F., Taram, M., Jalali, A., Elthakeb, A. T., Tullsen, D. M., & Esmaeilzadeh, H. (2021) · International World Wide Web Conference (WWW) · doi:10.1145/3442381.3449965 · Cited by 74
Deep Learning with Label Differential PrivacyPreprint
Ghazi, B., Golowich, N., Kumar, R., Manurangsi, P., & Zhang, C. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.06062 · Cited by 111
“Differential privacy (DP) (Dwork et al., 2006b,a) is a popular privacy notion that has been the basis of several practical deployments including in tech (Erlingsson et al., 2014; Shankland, 2014; Greenberg, 2016; Apple Differential Privacy Team, 2017; Ding et al., 2017) and the U.S. Census (Abowd,…”
BackgroundMethodInfluentialDifferentially Private Federated Stochastic Primal-Dual Learning for Internet of Vehicles
Li, Y., Wang, S., & Chang, T. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3534982 · Cited by 68
“A common approach to preventing differential attacks is known as differential privacy (DP), which can guarantee with high confidence that an adversary cannot deduce any private information from databases or released models [20].”
Akkuş, İ. E., Chen, R., Hardt, M., Francis, P., & Gehrke, J. E. (2012) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/2382196.2382268 · Cited by 58
“Although attempts have been made to provide differential privacy in a distributed environment, these attempts either incur high overhead [36] or suffer from client churn [47,50], making them impractical in a large-scale environment.”
BackgroundInfluentialPrivacy-preserving Data Mining in Industry
Kenthapadi, K., Mironov, I., & Thakurta, A. (2019) · ACM International Conference on Web Search and Data Mining (WSDM) · doi:10.1145/3289600.3291384 · Cited by 63
“We will motivate the need for a rigorous privacy guarantee, and present the notion of differential privacy [8–10].”
BackgroundSelf-citationStability of Stochastic Gradient Descent on Nonsmooth Convex LossesPreprint
Bassily, R., Feldman, V., Guzmán, C., & Talwar, K. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.06914 · Cited by 58
“A randomized algorithm A is (α, β)-differentially private if, for any pair of datasets S ' S′, and for all events O in the output range of A, we have P [A(S) ∈ O] ≤ e · P [A(S′) ∈ O] + β, where the probability is taken over the random coins ofA [14, 13].”
BackgroundVisual privacy attacks and defenses in deep learning: a survey
Zhang, G., Liu, B., Zhu, T., Zhou, A., & Zhou, W. (2022) · Artificial Intelligence Review · doi:10.1007/s10462-021-10123-y · Cited by 60
P4P: practical large-scale privacy-preserving distributed computation robust against malicious users
Duan, Y., Canny, J., & Zhan, J. (2010) · USENIX Security Symposium · Cited by 66
Differential Privacy for Data and Model Publishing of Medical Data
Sun, Z., Wang, Y., Shu, M., Liu, R., & Zhao, H. (2019) · IEEE Access · doi:10.1109/access.2019.2947295 · Cited by 56
“In this paper, we mainly use the basic composition theorem [16], [17] and advanced composition theorems [18]–[21] to repeatedly apply additive-noise mechanisms.”
MethodSecurity and Privacy Issues in Deep Learning
Bae, H., Jang, J., Jung, D., Jang, H., Ha, H., Lee, H., & Yoon, S. (2018) · arXiv · doi:10.48550/arxiv.1807.11655 · Cited by 56
“Against such privacy threats, privacy-preserving techniques including fully homomorphic encryption (FHE) [16, 17, 18, 19, 20], differential privacy [21, 22, 23, 24, 25, 26, 27, 28], and secure multi-party computation (SMC) [29, 30], have been combined with the DL frameworks.”
BackgroundMethodInfluentialShortest Paths and Distances with Differential Privacy
Sealfon, A. (2016) · ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems (PODS) · doi:10.1145/2902251.2902291 · Cited by 79
Privacy Odometers and Filters: Pay-as-you-Go CompositionPreprint
Rogers, R., Roth, A., Ullman, J., & Vadhan, S. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1605.08294 · Cited by 56
UnLynx: A Decentralized System for Privacy-Conscious Data Sharing
Froelicher, D., Egger, P., Sousa, J. S., Raisaro, J. L., Huang, Z., Mouchet, C., Ford, B., & Hubaux, J. (2017) · Proceedings on Privacy Enhancing Technologies · doi:10.1515/popets-2017-0047 · Cited by 55
“[22] remove the need for a trusted party by using verifiable secret sharing of noise values that are then combined in order to generate the noise.”
BackgroundLocal Differential Privacy for data collection and analysis
Wang, T., Zhao, J., Hu, Z. J., Yang, X., Ren, X., & Lam, K. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.09.073 · Cited by 61
TextHide: Tackling Data Privacy in Language Understanding Tasks
Huang, Y., Song, Z., Chen, D., Li, K., & Arora, S. (2020) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2020.findings-emnlp.123 · Cited by 77
Extractors for Low-Weight Affine Sources
Rao, A. (2009) · IEEE Conference on Computational Complexity (CCC) · doi:10.1109/ccc.2009.36 · Cited by 55
Privacy-Preserving Retrieval-Augmented Generation with Differential PrivacyPreprint
Koga, T., Wu, R., Zhang, Z., & Chaudhuri, K. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2412.04697 · Cited by 72
“Definition 1 ( ( ϵ, δ ) -Differential Privacy [10]) .”
Differentially Private LQ Control
Yazdani, K., Jones, A. M., Leahy, K., & Hale, M. (2022) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2022.3148710 · Cited by 70
Max-Information, Differential Privacy, and Post-selection Hypothesis Testing
Rogers, R., Roth, A., Smith, A., & Thakkar, O. D. (2016) · Annual Symposium on Foundations of Computer Science · doi:10.1109/focs.2016.59 · Cited by 90
“3 (Differential Privacy [23, 28]).”
EdgeSanitizer: Locally Differentially Private Deep Inference at the Edge for Mobile Data Analytics
Xu, C., Ren, J., She, L., Zhang, Y., Qin, Z., & Ren, K. (2019) · IEEE Internet of Things Journal · doi:10.1109/jiot.2019.2897005 · Cited by 53
“The Laplace Perturbation Mechanism (LPM) [15] applies noise drawn from a suitable Laplace distribution to perturb the query results.”
BackgroundMethodInfluentialChallenges and Open Problems of Legal Document Anonymization
Csányi, G. M., Nagy, D., Vági, R., Vadász, J. P., & Orosz, T. (2021) · Symmetry · doi:10.3390/sym13081490 · Cited by 53
“, when the information is shared in multiple places or with multiple individuals [30,45].”
Practical and private (deep) learning without sampling or shuffling
Kairouz, P., McMahan, B., Song, S., Thakkar, O., Thakurta, A., & Xu, Z. (2021) · arXiv · doi:10.48550/arxiv.2103.00039 · Cited by 52
“Assuming the data set D is sampled i.i.d. from a distribution τ , and the algorithm A outputs (cid:98) θ ∈ C , we want to minimize All the algorithms in this paper guarantee differential privacy [20, 21] and R´enyi differential privacy [53] (See Section 3 for details).”
MethodTowards Privacy-Aware and Trustworthy Data Sharing Using Blockchain for Edge Intelligence
Qu, Y., Ma, L., Ye, W., Zhai, X., Yu, S., Li, Y., & Smith, D. B. (2023) · Big Data Mining and Analytics · doi:10.26599/bdma.2023.9020012 · Cited by 54
Privacy-Preserving Distributed Processing: Metrics, Bounds and Algorithms
Li, Q., Gundersen, J. S., Heusdens, R. R., & Christensen, M. G. (2021) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2021.3050064 · Cited by 52
Privacy-Preserving Federated Learning in Healthcare
Moon, S., & Lee, W. H. (2023) · International Conference on Electronics, Information, and Communication (ICEIC) · doi:10.1109/iceic57457.2023.10049966 · Cited by 62
“Several privacy-enhancing technologies, differential privacy to communicate noisy data sketches [28], homomorphic encryption to operate on encrypted data [29], and secure multiparty computation for sensitive datasets distributed across different data owners [30], can be used to increase the data privacy capabilities o…”
DP-BART for Privatized Text Rewriting under Local Differential Privacy
Igamberdiev, T., & Habernal, I. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.874 · Cited by 108
“While the original definition of DP only included this term, the additional additive term δ was later introduced in Dwork et al. (2006a) and represents the ‘crypto-graphically small’ probability that pure ε -DP is broken.”
BackgroundFederated Learning in Smart Healthcare: A Survey of Applications, Challenges, and Future Directions
Nasajpour, M., Pouriyeh, S., Parizi, R. M., Han, M., Mosaiyebzadeh, F., Liu, L., Xie, Y., & Batista, D. M. (2025) · Electronics · doi:10.3390/electronics14091750 · Cited by 57
Generalized Bisimulation Metrics
Konstantinos Chatzikokolakis, A., Gebler, D., Catuscia Palamidessi, A., & Xu, L. (2014) · Lecture notes in computer science · doi:10.1007/978-3-662-44584-6_4 · Cited by 51
“An approximate, also known as (ε, δ) version of differential privacy is also widely used [20], relaxing the definition by an additive factor δ, i.”
BackgroundDifferentially private subspace clustering
Wang, Y., Wang, Y.-X., & Singh, A. (2015) · Singapore Management University Institutional Knowledge (InK) (Singapore Management University) · Cited by 67
On Syntactic Anonymity and Differential Privacy
Clifton, C., & Tassa, T. (2013) · Transactions on data privacy · doi:10.5555/2612167.2612170 · Cited by 51
Differentially Private Graphical Degree Sequences and Synthetic Graphs
Karwa, V., & Slavković, A. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33627-0_21 · Cited by 51
Average-Case Averages: Private Algorithms for Smooth Sensitivity and Mean EstimationPreprint
Bun, M., & Steinke, T. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1906.02830 · Cited by 51
“Definition 1 (Differential Privacy (DP) [14, 12]).”
Clustering Algorithms for the Centralized and Local ModelsPreprint
Kobbi Nissim, B., & Stemmer, U. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1707.04766 · Cited by 75
“y j ). Let > 0, and assume f : U ! R d has L 1 -sensitivity k . The mechanism that on input S 2 U outputs f (S )+ Lap( k ) d is (; 0) -dierentially private. Theorem 2.5 (Gaussian Mechanism [7]) . Let ; 2 (0 ;1) , and assume f : U ! R d has L 2 - sensitivity k . Let k p 2ln(1 :25 = ). The mechanism that on input S 2…”
BackgroundMethodInfluentialTaming the Devil: Techniques for Evaluating Anonymized Network Data
Coull, S. E., Wright, C. V., Keromytis, A. D., Monrose, F., & Reiter, M. K. (2008) · Columbia Academic Commons (Columbia University) · doi:10.7916/d8bc47w0 · Cited by 50
Differentially Private and Fair Classification via Calibrated Functional Mechanism
Ding, J., Zhang, X., Li, X., Wang, J., Yu, R., & Pan, M. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i01.5402 · Cited by 67
Scaling up Differentially Private Deep Learning with Fast Per-Example Gradient Clipping
Lee, J. W., & Kifer, D. (2020) · Proceedings on Privacy Enhancing Technologies · doi:10.2478/popets-2021-0008 · Cited by 81
Differentially Private Language Models for Secure Data Sharing
Mattern, J., Jin, Z., Weggenmann, B., Schölkopf, B., & Sachan, M. (2022) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2022.emnlp-main.323 · Cited by 99
Differentially private state estimation in distribution networks with smart meters
Sandberg, H., Dán, G., & Thobaben, R. (2015) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2015.7402921 · Cited by 49
“That there is not a dramatic loss of privacy under compositions of differentially private measurements is in fact one of the main advantages of differential privacy, see [14], [9] for more general treatments.”
BackgroundDifferentially Private Ordinary Least Squares
Sheffet, O. (2019) · Journal of Privacy and Confidentiality · doi:10.29012/jpc.654 · Cited by 63
“It is known (Dwork et al., 2006b) that if ALG outputs a vector in R d such that for any A and A (cid:48) it holds that (cid:107) ALG ( A ) − ALG ( A (cid:48) ) (cid:107) 1 ≤ B , then adding Laplace noise Lap (1 /(cid:15) ) to each coordinate of the output of ALG ( A ) satisfies (cid:15)…”
BackgroundInfluentialZhai, L., & Vamvoudakis, K. G. (2020) · International Journal of Robust and Nonlinear Control · doi:10.1002/rnc.5040 · Cited by 49
“31,32 The parameters 𝜖 and 𝛿 will be used to describe the privacy-preserving requirements.”
MethodLDP-Fed: Federated Learning with Local Differential PrivacyPreprint
Truex, S., Liu, L., Chow, K.-H., Gürsoy, M. E., & Wei, W. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.03637 · Cited by 70
“Additional protocols allow users to leverage such cryptographic techniques to decrease the scale of noise [7, 10, 23].”
BackgroundEvaluation of the Precision-Privacy Tradeoff of Data Perturbation for Smart Metering
Savi, M., Rottondi, C., & Verticale, G. (2015) · IEEE Transactions on Smart Grid · doi:10.1109/tsg.2014.2387848 · Cited by 48
“[18] proposed a protocol for the distributed generation of random noise, aimed at the distributed implementation of privacy-preserving statistical databases by means of a verifiable secret sharing scheme.”
BackgroundOn-Average KL-Privacy and Its Equivalence to Generalization for Max-Entropy Mechanisms
Wang, Y.-X., Lei, J., & Fienberg, S. E. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-45381-1_10 · Cited by 61
“The (certainly incomplete) list includes ( , δ)-approximate differential privacy (Approx-DP) [13], random differential privacy (Rand-DP) [17], Personalized Differential Privacy (Personal-DP) [15, 22] and Total-Variation-Privacy (TV-Privacy) [4, 5].”
BackgroundPrivacy-Preserving Power System Obfuscation: A Bilevel Optimization Approach
Mak, T. W. K., Fioretto, F., Shi, L., & Hentenryck, P. V. (2019) · IEEE Transactions on Power Systems · doi:10.1109/tpwrs.2019.2945069 · Cited by 56
“In particular, Differential Privacy (DP) [2], [3] captures a desirable privacy property of computations over datasets.”
MethodInfluentialHarder, F., Adamczewski, K., & Park, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.11603 · Cited by 48
“We can add noise to h for privacy, where the level of noise is calibrated to the global sensitivity (Dwork et al., 2006), ∆h, defined by the maximum difference in terms of L2-norm ||h(D)−h(D)||2, for neighboring D and D′ (i.”
BackgroundInfluentialDifferential Privacy in the Shuffle Model: A Survey of SeparationsPreprint
Cheu, A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.11839 · Cited by 48
“Lemma 2 (Binomial Mechanism [29, 36]) .”
Federated Learning with Differential Privacy for Resilient Vehicular Cyber Physical Systems
Olowononi, F. O., Rawat, D. B., & Liu, C. (2021) · 2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC) · doi:10.1109/ccnc49032.2021.9369480 · Cited by 47
“As highlighted in [7], the formal definition of differential privacy on the basis of two neighboring databases D and D that differ from each other in only one single member is stated as follows: Definition 1: A randomized algorithmic function < satisfies -differential privacy condition PR if for any two adjacent datas…”
BackgroundRPPTD: Robust Privacy-Preserving Truth Discovery Scheme
Chen, J., Liu, Y., Xiang, Y., & Sood, K. (2021) · IEEE Systems Journal · doi:10.1109/jsyst.2021.3099103 · Cited by 47
“To resist differential attacks, we add random independent and identically distributed noise v k to the unencrypted data UD k (i.e., A k ) [23].”
MethodFederated learning to comply with data protection regulations
Chalamala, S. R., Kummari, N. K., Singh, A. K., Saibewar, A., & Chalavadi, K. M. (2022) · CSI Transactions on ICT · doi:10.1007/s40012-022-00351-0 · Cited by 56
“Differential privacy [45–48] can be used to provide quantifiable privacy in a database.”
BackgroundA Blockchain-Assisted Trusted Federated Learning for Smart Agriculture
Manoj, T., Makkithaya, K., & Narendra, V. G. (2025) · SN Computer Science · doi:10.1007/s42979-025-03672-4 · Cited by 57
“It can be attained through a technique called DP, which was initially proposed by [25].”
A rate-disortion perspective on local differential privacy
Sarwate, A. D., & Sankar, L. (2014) · Allerton Conference on Communication, Control, and Computing · doi:10.1109/allerton.2014.7028550 · Cited by 60
“(9) The worst case Hamming distortion for an M -ary alphabet is defined as ∆ (P, Q) = max P∈P EP×Q[d(X, X̂)].”
BackgroundImproving utility and security of the shuffler-based differential privacy
Wang, T., Ding, B., Xu, M., Huang, Z., Hong, C., Zhou, J., Li, N., & Jha, S. (2020) · Proceedings of the VLDB Endowment · doi:10.14778/3424573.3424576 · Cited by 47
“In [27], a distributed noise generation protocol was proposed to prevent some party from adding malicious noise.”
BackgroundHow Does Data Augmentation Affect Privacy in Machine Learning?
Yu, D., Zhang, H., Chen, W., Yin, J., & Liu, T. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i12.17284 · Cited by 69
“This is connected with the differential privacy (Dwork et al. 2006b,a), which measures how a single data point affects the parameter posterior in the worst case.”
BackgroundXu, J. J., Zhang, W., & Fei, W. (2021) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2021.3107796 · Cited by 66
“To alleviate this issue, differential privacy (DP), as an alternative theoretical model to provide mathematical privacy guarantees, has caught people’s attention [13].”
BackgroundInfluentialAn Ad Omnia Approach to Defining and Achieving Private Data Analysis
Dwork, C. (2008) · Springer eBooks · doi:10.1007/978-3-540-78478-4_1 · Cited by 63
“It is also possible to combine techniques of secure function evaluation with the techniques described above, permitting a collection of data holders to cooperatively simulate K; see [17] for details.”
MethodSelf-citationA framework for adaptive differential privacy
Winograd-Cort, D., Haeberlen, A., Roth, A., & Pierce, B. C. (2017) · Proceedings of the ACM on Programming Languages · doi:10.1145/3110254 · Cited by 51
“Among the numerous proposals for negotiating this inherent tension, differential privacy [Dwork et al. 2006b], with its firm statistical foundation and minimal assumptions on attacker knowledge, is appealingly resilient to the vulnerabilities that have plagued other mechanisms [Mills 2006;…”
BackgroundInfluentialSecure Computation with Differentially Private Access Patterns
Mazloom, S., & Gordon, S. (2018) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3243734.3243851 · Cited by 64
“[12] suggest using the geometric distribution with p = 2, precisely because it is easy to sample in a garbled circuit.”
MethodRobustness Implies Privacy in Statistical Estimation
Hopkins, S. B., Kamath, G., Majid, M., & Narayanan, S. (2023) · ACM Symposium on Theory of Computing (STOC) · doi:10.1145/3564246.3585115 · Cited by 46
“Definition 1.1 (Differential Privacy (DP) [18, 20]).”
InfluentialDynamic Differential-Privacy Preserving SGDPreprint
Du, J. Q., Li, S., Chen, X., Chen, S., & Hong, M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.00173 · Cited by 60
“Simple composition (Dwork et al., 2006; Dwork & Lei, 2009) with (cid:15) = O ( T ) and advance composition (Dwork et al., 2010) √ Murtagh & Vadhan (2016) that computing the exact DP guarantees for the composition Abadi et al. (2016) propose the moment accountant of DP-SGD.”
Privacy Principles for Sharing Cyber Security Data
Fisk, G., Ardi, C., Pickett, N., Heidemann, J., Fisk, M., & Papadopoulos, C. (2015) · IEEE Security and Privacy Workshops (SPW) · doi:10.1109/spw.2015.23 · Cited by 45
“Various tools and solutions that one could use to protect privacy in queries include (but are not limited to) salted hashes [26], homomorphic encryption [27], [28], privacy integrated queries [24], privacy preserving queries [29], location cloaking [30], distributed noise generation [31], and secure queries [23].”
BackgroundFeldman, D., Xiang, C., Zhu, R., & Rus, D. L. (2017) · ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN) · doi:10.1145/3055031.3055090 · Cited by 67
“In this paper we use the relaxed version of pure dierential privacy, known as (ε,δ )-dierential privacy [8].”
MethodXi, B. (2020) · Wiley Interdisciplinary Reviews Computational Statistics · doi:10.1002/wics.1511 · Cited by 55
“…preserving mechanisms, k − anonymity (Samarati & Sweeney, 1998; Samarati, 2001; Sweeney, 2002) and differential privacy (Dwork & Smith, 2010a; Dwork, Kenthapadi, McSherry, Mironov, & Naor, 2006; Dwork, McSherry, Nissim, & Smith, 2006). k − anonymity directly perturbs the individual data…”
MVG Mechanism: Differential Privacy under Matrix-Valued Query
Chanyaswad, T., Dytso, A., Poor, H. V., & Mittal, P. (2018) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3243734.3243750 · Cited by 60
“The Gaussian mechanism [23, 25, 62] to the L 2 -sensitivity.”
BackgroundEfficient Algorithms for Privately Releasing Marginals via Convex Relaxations
Dwork, C., Nikolov, A., & Talwar, K. (2015) · Discrete & Computational Geometry · doi:10.1007/s00454-015-9678-x · Cited by 65
Self-citationCollaborative Machine Learning: Schemes, Robustness, and Privacy
Wang, J., Pal, A., Yang, Q., Kant, K., Zhu, K., & Guo, S. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3169347 · Cited by 54
“Currently, the solutions are mainly based on differential privacy [132] [133], secure multiparty computation [134] [135] [136], and homomorphic encryption [137], [138].”
BackgroundAccurate and efficient private release of datacubes and contingency tables
Yaroslavtsev, G., Cormode, G., Procopiuc, C. M., & Srivastava, D. (2013) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde.2013.6544871 · Cited by 44
Proactive fault-tolerant aggregation protocol for privacy-assured smart metering
Won, J., Ma, C. Y. T., Yau, D. K. Y., & Rao, N. S. V. (2014) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom.2014.6848230 · Cited by 44
“( -differential privacy) A randomized algorithm A is -differentially private if for any two datasets D1 and D2 that differ on a single element, and for all S ⊆ Range(A), Pr(A(D1) ∈ S) ≤ exp( ) · Pr(A(D2) ∈ S).”
BackgroundKarwa, V., Krivitsky, P. N., & Slavković, A. (2016) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12185 · Cited by 44
“The framework of differential privacy (DP) (Dwork et al., 2006) has emerged from the theoretical computer science community as a principled way to provably limit a worst case disclosure risk in presence of any arbitrary external information.”
BackgroundMethodInfluentialDifferentially Private Nonparametric Hypothesis Testing
Couch, S. P., Kazan, Z., Shi, K., Bray, A. P., & Groce, A. (2019) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3319535.3339821 · Cited by 64
“[10] Cynthia Dwork, Krishnaram Kenthapadi, Frank McSherry, Ilya Mironov, and Moni Naor.”
BackgroundProving differential privacy with shadow execution
Wang, Y., Ding, Z., Wang, G., Kifer, D., & Zhang, D. (2019) · ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI) · doi:10.1145/3314221.3314619 · Cited by 58
“There are currently several popular variants of differential privacy [13, 18, 19, 32].”
BackgroundCryptϵ: Crypto-Assisted Differential Privacy on Untrusted Servers
Chowdhury, A. R., Wang, C., He, X., Machanavajjhala, A., & Jha, S. (2020) · ACM SIGMOD International Conference on Management of Data (SIGMOD) · doi:10.1145/3318464.3380596 · Cited by 92
“with the trust assumption of the CDP model is using cryptographic primitives [9, 14, 17, 20, 29, 31, 37, 41, 92, 93].”
MethodNon-interactive differential privacy: a survey
Leoni, D. (2012) · doi:10.1145/2422604.2422611 · Cited by 44
Not All Features Are Equal: Discovering Essential Features for Preserving Prediction Privacy
Mireshghallah, F., Taram, M., Jalali, A., Elthakeb, A. T., Tullsen, D., & Esmaeilzadeh, H. (2020) · arXiv · doi:10.48550/arxiv.2003.12154 · Cited by 44
“he effect of an individual opting out of contributing to the dataset, by ensuring that any effect due to the inclusion of one’s data is small. 2.3. Laplace Mechanism Definition 2.2.Laplace Mechanism. (Dwork et al., 2006b) Given a target function fand a fixed 0, the randomizing algorithm A f(D) = f(D) + xwhere xis a pert…”
BackgroundInfluentialLabel-Only Membership Inference Attacks and Defenses in Semantic Segmentation Models
Zhang, G., Liu, B., Zhu, T., Ding, M., & Zhou, W. (2022) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2022.3154029 · Cited by 43
Ruan, W., Xu, M., Fang, W., Wang, L., Wang, L., & Han, W. (2023) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp46215.2023.10179422 · Cited by 56
“[22] to securely generate the Gaussian noise with mean 0 and standard deviation σ.”
BackgroundInfluentialSplitX: high-performance private analytics
Chen, R., Akkuş, İ. E., & Francis, P. (2013) · ACM SIGCOMM Conference (SIGCOMM) · doi:10.1145/2486001.2486013 · Cited by 43
“of which scale substantially better than any of [13, 26, 28], nevertheless require a public-key operation per single bit of client answer, and require roughly one kilobit of message data per single bit of client answer.”
BackgroundInfluentialGradient Perturbation is Underrated for Differentially Private Convex Optimization
Yu, D., Zhang, H., Chen, W., Yin, J., & Liu, T. (2020) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2020/431 · Cited by 48
Kamino: constraint-aware differentially private data synthesis
Ge, C., Mohapatra, S., He, X., & Ilyas, I. F. (2021) · Proceedings of the VLDB Endowment · doi:10.14778/3467861.3467876 · Cited by 60
“Complex DP algorithms can be built from the basic algorithms following two important properties of differential privacy: 1) Post-processing [25] states that for any function g defined over the output of the mechanismM, ifM satisfies (ε, δ)-DP, so does g(M); 2) Composability [24] states that ifM1,M2, · · · ,Mk satisfy…”
BackgroundMethodDP-SGD vs PATE: Which Has Less Disparate Impact on Model Accuracy?Preprint
Uniyal, A., Naidu, R., Kotti, S., Singh, S., Kenfack, P. J., Mireshghallah, F., & Trask, A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.12576 · Cited by 54
“For (cid:15) ≥ 0 , an algorithm A is understood to be Differentially Private (Dwork et al., 2006a;b) if and only if for any pair of datasets that differ in only one element, the following statement holds true.”
Federated and Privacy-Preserving Learning of Accounting Data in Financial Statement Audits
Schreyer, M., Sattarov, T., & Borth, D. (2022) · ACM International Conference on AI in Finance (ICAIF) · doi:10.1145/3533271.3561674 · Cited by 53
“[16] proposed the concept of Differential Privacy (DP) to address the challenge of inferring confidential training data information through ML models.”
Influential
Metadata
- Authors
- Dwork, C., Kenthapadi, K., McSherry, F., Mironov, I., & Naor, M.
- Affiliations
- Microsoft (United States) (US)
- Stanford University (US)
- Weizmann Institute of Science (IL)
- Venue
- Lecture Notes in Computer Science
- Publisher
- Springer Berlin Heidelberg
- Published
- 2006
- Pages
- 486-503
- ISSN
- 0302-9743 (ISSN-L)
- 1611-3349
- OpenAlex
- W1557833142
- Cited by
- 1,835
- References
- 31
- Citations by year
- 2026: 65 · 2025: 138 · 2024: 184 · 2023: 179 · 2022: 175 · 2021: 188 · 2020: 216 · 2019: 123 · 2018: 80 · 2017: 81 · 2016: 61 · 2015: 69 · 2014: 78 · 2013: 70 · 2012: 52
- FWCI
- 7.51
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- Lecture notes in computer science (publishedVersion)
- 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
- malicious security, verifiable secret sharing, Gaussian noise, exponential distribution, random bit generation, distributed protocols
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.