Differentially Private Distributed Optimization
Open accessThe DOI matches its Crossref record; 16 of the 17 references it cites were found in a registry or an index, and 1 are not checked yet.
Abstract
In distributed optimization and iterative consensus literature, a standard problem is for N agents to minimize a function f over a subset of Euclidean space, where the cost function is expressed as a sum Σ fi. In this paper, we study the private distributed optimization problem (PDOP) with the additional requirement that the cost function of the individual agents should remain differentially private. The adversary attempts to infer information about the private cost functions from the messages that the agents exchange. Achieving differential privacy requires that any change of an individual's cost function only results in unsubstantial changes in the statistics of the messages. We propose a class of iterative algorithms for solving PDOP, which achieves differential privacy and convergence to a common value. Our analysis reveals the dependence of the achieved accuracy and the privacy levels on the the parameters of the algorithm. We observe that to achieve ϵ-differential privacy the accuracy of the algorithm has the order of O(1/ϵ2).
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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?
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“current work on the hybrid model typically assumes that regardless of the user trust preference, their data comes from the same distribution [40, 39, 57].”
From Advances and Open Problems in Federated Learning (2021) · cites this paper as [57]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Consensus and Cooperation in Networked Multi-Agent Systems
Olfati-Saber, R., Fax, J. A., & Murray, R. M. (2007) · Proceedings of the IEEE · doi:10.1109/jproc.2006.887293 · Cited by 10,522 · cited 1×
“For instance, such algorithms have been designed for the smart grids [3] and sensor networks [13].”
MethodDwork, C. (2006) · Lecture Notes in Computer Science · doi:10.1007/11787006_1 · Cited by 5,291 · cited 1×
“Categories and Subject Descriptors D.1.3 [Software]: Concurrent Programming—Distributed programming General Terms Algorithm Keywords Distributed Optimization, Differential Privacy, Iterative Consensus”
BackgroundInfluentialDistributed Subgradient Methods for Multi-Agent Optimization
Nedic, A., & Ozdaglar, A. (2009) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2008.2009515 · Cited by 3,719 · cited 2×
“Categories and Subject Descriptors D.1.3 [Software]: Concurrent Programming—Distributed programming General Terms Algorithm Keywords Distributed Optimization, Differential Privacy, Iterative Consensus”
BackgroundInfluentialDifferential Privacy: A Survey of Results
Dwork, C. (2008) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-79228-4_1 · Cited by 3,542 · cited 1×
“Categories and Subject Descriptors D.1.3 [Software]: Concurrent Programming—Distributed programming General Terms Algorithm Keywords Distributed Optimization, Differential Privacy, Iterative Consensus”
BackgroundInfluentialDistributed asynchronous deterministic and stochastic gradient optimization algorithms
Tsitsiklis, J., Bertsekas, D., & Athans, M. (1986) · IEEE Transactions on Automatic Control · doi:10.1109/tac.1986.1104412 · Cited by 1,967
Our Data, Ourselves: Privacy Via Distributed Noise Generation
Dwork, C., Kenthapadi, K., McSherry, F., Mironov, I., & Naor, M. (2006) · Lecture Notes in Computer Science · doi:10.1007/11761679_29 · Cited by 1,835
Differentially Private Empirical Risk Minimization
Chaudhuri, K., Monteleoni, C., & Sarwate, A. (2009) · Journal of machine learning research · doi:10.5555/1953048.2021036 · Cited by 1,634 · cited 1×
“In [1, 9] the authors solve a privacy preserving optimization problem with two methods: output perturbation and objective perturbation.”
Method(2011) · PubMed · Cited by 990 · cited 1×
“In [1, 9] the authors solve a privacy preserving optimization problem with two methods: output perturbation and objective perturbation.”
MethodGradient methods for minimizing composite objective function
Nesterov, Y. E. (2007) · Cited by 867
Distributed Stochastic Subgradient Projection Algorithms for Convex Optimization
Sundhar Ram, S., Nedić, A., & Veeravalli, V. V. (2010) · Journal of Optimization Theory and Applications · doi:10.1007/s10957-010-9737-7 · Cited by 814 · cited 1×
“Categories and Subject Descriptors D.1.3 [Software]: Concurrent Programming—Distributed programming General Terms Algorithm Keywords Distributed Optimization, Differential Privacy, Iterative Consensus”
BackgroundInfluential
Research works printed but not linked: 1
- 10.5555/1791834.1791836 · doi:10.5555/1791834.1791836
Cited by
Items 101–200 of 274
Protecting privacy of topology in consensus networks
Katewa, V., Chakrabortty, A., & Gupta, V. (2015) · Proceedings of the ... American Control Conference/Proceedings of the American Control Conference · doi:10.1109/acc.2015.7171103 · Cited by 22
“In [8], the authors present an algorithm to ensure differentially privacy in a distributed optimization problem with separable cost function.”
BackgroundDistributed dynamic online learning with differential privacy via path-length measurement
Chen, L., Ding, X., Zhou, P., & Jin, H. (2023) · Information Sciences · doi:10.1016/j.ins.2023.02.037 · Cited by 16
Distributed empirical risk minimization with differential privacy
Liu, C., Johansson, K. H., & Shi, Y. (2024) · Automatica · doi:10.1016/j.automatica.2024.111514 · Cited by 23
“, h ≡ 0, the authors in [14] developed a differentially private distributed gradient descent (DGD) algorithm by perturbing the local output with Laplace noise.”
BackgroundPrivacy-Preserving Channel Estimation in Cell-Free Hybrid Massive MIMO Systems
Xu, J., Wang, X., Zhu, P., & You, X. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3053770 · Cited by 16
Hu, J., Chen, G., Li, H., Huang, T., & Ran, L. (2023) · Information Sciences · doi:10.1016/j.ins.2023.119199 · Cited by 15
Wang, Y., & Başar, T. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3366037 · Cited by 15
“Definition 2 ( (cid:2) -DP, [20]): For a given (cid:2) > 0 , an algorithm A is -differentially private if for any two adjacent networked games P and P (cid:13) , any set of observation sequences O s ⊆ O ( O is the set of all possible observation sequences), and any initial state ϑ 0 , we always have…”
BackgroundMethodInfluentialLi, Q., Heusdens, R. R., & Christensen, M. G. (2022) · Signal Processing · doi:10.1016/j.sigpro.2022.108456 · Cited by 15
Programmable Crypto-Control for IoT Networks: An Application in Networked Microgrids
Wang, L., Zhang, P., Tang, Z., & Qin, Y. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3194838 · Cited by 15
“The traditional differential-privacy algorithm is adopted from [31].”
MethodDistributed Sketching Methods for Privacy Preserving RegressionPreprint
Bartan, B., & Pilancı, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.06538 · Cited by 15
“The approximation error of the algorithm in [6] depends on parameters that Algorithm 1 does not have such as the initial step size, the step size decay rate, and noise decay rate.”
BackgroundMethodInfluentialCustomized Local Differential Privacy for Multi-Agent Distributed OptimizationPreprint
Dobbe, R., Pu, Y., Zhu, J., Ramchandran, K., & Tomlin, C. J. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1806.06035 · Cited by 14
“Furthermore, most previous works consider privacy protection for either individual objective function [13] or the individual constraint [10], our more general formulation enables us to provide privacy guarantees on both local objective function parameters and local constraint parameters.”
BackgroundMethodResultInfluentialDifferentially Private Distributed Optimization Over Diagraphs With Application to Image Deblurring
Yang, Z., He, W., & Yang, Y. (2024) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2024.3395851 · Cited by 14
“Nevertheless, these strategies impose strict requirements on local cost functions, such as assuming bounded gradients [22] or requiring square-integrable objective functions [23], which significantly limit their applicability.”
InfluentialYang, Z., He, W., & Yang, S. (2025) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2025.3544517 · Cited by 13
“Following [16], [18], we describe the distributed optimization problem (1) using parameters ( S , Z , F, G ) .”
InfluentialDifferentially private distributed protocol for electric vehicle charging
Han, S., Topcu, U., & Pappas, G. J. (2014) · Allerton Conference on Communication, Control, and Computing · doi:10.1109/allerton.2014.7028462 · Cited by 13
“[10], it is assumed that the coordination signals do not change with changes in the individual cost functions, which may contain private information and needs to be protected.”
BackgroundDifferential Privacy for Network Identification
Katewa, V., Chakrabortty, A., & Gupta, V. (2019) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2019.2922169 · Cited by 20
“In [7], [8] and [17], [18], the authors present noisy update algorithms to protect sensitive constraints, states and cost functions, respectively.”
BackgroundLiu, E., & Cheng, P. (2018) · IEEE Access · doi:10.1109/access.2018.2802456 · Cited by 13
Differential Private Discrete Noise-Adding Mechanism: Conditions, Properties, and Optimization
Qin, S., He, J., Fang, C., & Lam, J. (2023) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2023.3317644 · Cited by 12
“Many studies have adopted the adjacency-based definition to protect the exact value of the record, as the generalized neighborhood property enables the privacy protection of specific data attributes [6], [7], [8], [9], [10].”
BackgroundMethodChen, Z., & Wang, Y. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3488504 · Cited by 17
“Asthedefactostandardfor privacyprotection,differentialprivacy(DP)isgainingincreased traction,andhasbeenemployedinmanydistributedlearningand optimizationalgorithms[35],[36],[37],[38],[39],[40],[41], [42],[43],[44],[45],[46],[47].”
InfluentialDistributed Mechanism Design With Learning Guarantees for Private and Public Goods Problems
Sinha, A., & Anastasopoulos, A. (2019) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2019.2955999 · Cited by 12
“Some recent works in the literature of distributed optimization attempt to resolve “privacy” issues by means of dithering (i.e., adding noise to) the exchanged messages as in [16] or by changing the objective function [17].”
BackgroundRen, X., Yu, C.-M., Yu, W., Yang, X., Zhao, J., & Yang, S. (2020) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.3020089 · Cited by 12
“[26], [27] proposed a class of iterative algorithms for solving the private distributed optimization problem.”
BackgroundPrivate Learning on Networks: Part IIPreprint
Gade, S., & Vaidya, N. H. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.09185 · Cited by 13
“differential privacy is a probabilistic technique that involves use of randomized perturbations [26]–[30] to minimize the probability of uncovering of specific records from databases.”
BackgroundSelf-citationPrivacy-preserving Decentralized Optimization via DecompositionPreprint
Zhang, C., Gao, H. B., & Wang, Y. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1808.09566 · Cited by 14
“One widely used approach to enabling privacypreservation in decentralized optimization is differential privacy [18]–[21] which protects sensitive information by adding carefully-designed noise to exchanged states or objective functions.”
BackgroundMethodResultInfluentialDistributed Differentially Private Computation of Functions with Correlated NoisePreprint
Imtiaz, H., Mohammadi, J. A., & Sarwate, A. D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.10059 · Cited by 12
Chen, X., Jiang, W., Charalambous, T., & Shi, L. (2023) · IEEE Control Systems Letters · doi:10.1109/lcsys.2023.3292463 · Cited by 12
“For example, in distributed-optimization-based localization and rendezvous, exchanging the gradient of an agent leads to disclosing its position [13].”
BackgroundDistributed Privacy-Aware Fast Selection Algorithm for Large-Scale Data
Liu, H., & Chen, J. (2017) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2017.2761344 · Cited by 12
Lou, J., & Cheung, Y. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11311 · Cited by 11
“…2011; Kifer, Smith, and Thakurta 2012; Bassily, Smith, and Thakurta 2014; Talwar, Thakurta, and Zhang 2014; Jain and Thakurta 2013; Kasiviswanathan and Jin 2016) and sample-wise distributed datasets (Huang, Mitra, and Vaidya 2015; Nozari, Tallapragada, and Cortés 2016; Han, Topcu, and Pappas 2017).”
BackgroundPreserving privacy between features in distributed estimation
Heinze‐Deml, C., McWilliams, B., & Meinshausen, N. (2018) · Stat · doi:10.1002/sta4.189 · Cited by 11
“In distributed supervised learning, private approaches have been much less studied and focus mainly on the setting where data are horizontally partitioned (Huang et al., 2015; Zhang & Zhu, 2016).”
BackgroundOn the Privacy of Optimization
Weeraddana, P. C., & Fischione, C. (2017) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2017.08.1590 · Cited by 11
Coalition Formation for Multi-agent Pursuit Based on Neural Network
Pei, Z., Piao, S., Souidi, M. E. H., Qadir, M. Z., & Li, G. (2018) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-018-0893-6 · Cited by 11
“…designed a noise-adding mechanism according to the classic differential privacy framework which is implemented by agents [17].literature [16] propose a class of iterative algorithms for solving private distributed optimization problem in order to achieves differential privacy and…”
MethodOn the Privacy of Optimization ApproachesPreprint
Weeraddana, P. C., Athanasiou, G., Martin, J., Fischione, C., & Baras, J. S. (2012) · arXiv (Cornell University) · doi:10.48550/arxiv.1210.3283 · Cited by 11
Preserving Statistical Privacy in Distributed Optimization
Gupta, N., Gade, S., Chopra, N., & Vaidya, N. H. (2020) · IEEE Control Systems Letters · doi:10.1109/lcsys.2020.3005766 · Cited by 10
“However, these differetially private protocols suffer inevitably from privacy-accuracy trade-offs [7], [8].”
BackgroundSelf-citationCommunication-Efficient and Privacy-Aware Distributed Learning
Gogineni, V. C., Moradi, A., Venkategowda, N. K. D., & Werner, S. (2023) · IEEE Transactions on Signal and Information Processing over Networks · doi:10.1109/tsipn.2023.3322783 · Cited by 10
“In distributed information processing scenarios, such as consensus [5], [6], optimization [7], [8], filtering [9], and state estimation [10], [11], [12] privacy can be ensured through the implementation of cryptography-based mechanisms.”
BackgroundDistributed Privacy-Preserving Optimization With Accumulated Noise in ADMM
Liu, Z., Wang, W., Guo, F., & Gao, Q. Z. (2024) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2024.3424221 · Cited by 10
“DP-based algorithms typically employ a noise-adding mechanism, usually adding Laplacian or Gaussian noise, to achieve privacy preservation [10], [11], [12].”
Distributed Mechanism Design for Multicast Transmission
Heydaribeni, N., & Anastasopoulos, A. (2018) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2018.8618979 · Cited by 10
Local Differential Privacy for Multi-Agent Distributed Optimal Power Flow
Dobbe, R., Pu, Y., Zhu, J., Ramchandran, K., & Tomlin, C. J. (2020) · IEEE PES Innovative Smart Grid Technologies Europe · doi:10.1109/isgt-europe47291.2020.9248851 · Cited by 10
“Similarly, other works also limit protection to either individual objective functions [10] or individual constraint [7].”
BackgroundMethodInfluentialWang, Y. (2023) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc49753.2023.10383541 · Cited by 13
“Note that the DP approach in [38] addresses distributed optimization problems, which can be viewed as a dynamic average consensus problem if we regard the gradient of an agent’s objective function as the input to the agent.”
BackgroundMethodInfluentialSolving bin-packing problems under privacy preservation: Possibilities and trade-offs
Hoogervorst, R., Zhang, Y., Tillem, G., Erkin, Z., & Verwer, S. (2019) · Information Sciences · doi:10.1016/j.ins.2019.05.011 · Cited by 14
“In [27] , the authors assume the private user information is encoded in its cost functions and study the problem of differentially private distributed convex optimization.”
BackgroundRezazadeh, N., & Kia, S. S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.05286 · Cited by 10
“Additive noises have also been used as a privacy preservation mechanism in other distributed algorithms such as distributed optimization [13] and distributed estimation [14], [15].”
MethodLiu, B., Xie, F., & Chai, L. (2025) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2025.3546087 · Cited by 9
“…considerable interest, and many privacy-preserving algorithms have been proposed,broadlydividedintothreecategories:differentialprivacy-basedmethods[29],[30],[31],[32],[33],[34],[35],[36], [37],structure-basedmethods[38],[39],[40],[41],[42],and…”
Bartan, B., & Pilancı, M. (2023) · IEEE Transactions on Information Theory · doi:10.1109/tit.2023.3247559 · Cited by 9
“We note that the approximation error of the algorithm in [38] depends on parameters that Algorithm 1 does not have such as the initial step size, the step size decay rate, and noise decay rate.”
BackgroundMethodInfluentialDifferentially Private Consensus-Based Distributed OptimizationPreprint
Showkatbakhsh, M., Karakus, C., & Diggavi, S. N. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1903.07792 · Cited by 9
“It gives us the condition under which the distributed algorithm is ( ε , δ ) -DP. Working with ( ε , δ ) -DP as opposed to ε -DP in [5], enables us to derive a convergence result with a diminishing regret bound when the privacy requirement weakens.”
BackgroundInfluentialPrivacy-Preserving Distributed Optimization and Learning
Chen, Z., & Wang, Y. (2024) · Elsevier eBooks · doi:10.1016/b978-0-443-14081-5.00125-2 · Cited by 10
Differentially private and communication-efficient distributed nonconvex optimization algorithms
Xie, A., Yi, X., Wang, X., Cao, M., & Ren, X. (2025) · Automatica · doi:10.1016/j.automatica.2025.112338 · Cited by 12
Data-weighted ensemble learning for privacy-preserving distributed learning
Xie, L., Plis, S., & Sarwate, A. D. (2016) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2016.7472089 · Cited by 9
“The extensive existing work on distributed learning either ignores privacy constraints [1] [2] [3] [4] [5] [6] [7] [8] or assumes an asymptotic scenario where each site has a large number of samples [9] [10] [11].”
BackgroundLocally Differentially Private Distributed Online Learning With Guaranteed Optimality
Chen, Z., & Wang, Y. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3482977 · Cited by 8
“Unlike dynamic regret, which emphasizes the accumulated error [29], [43], instantaneous regret captures the real-time performance of the online algorithm, thereby having a better fit to spatiotemporally fluctuating environments [49].”
InfluentialChen, Y., Ma, Q., Jin, P., & Xu, S. (2025) · IEEE Transactions on Fuzzy Systems · doi:10.1109/tfuzz.2025.3542497 · Cited by 8
“For a given distributed NE problem H with an initial state τ (0) , the observation mapping function is defined as [36] (cid:2) -differential privacy): For a given > 0 , if for any two adjacent distributed NE seeking problems H and H (cid:9) , any set of observation sequences O s ⊆ O , and any…”
A gradient‐free distributed optimization method for convex sum of nonconvex cost functions
Pang, Y., & Hu, G. (2022) · International Journal of Robust and Nonlinear Control · doi:10.1002/rnc.6266 · Cited by 8
“Recent researches on the privacy issue in distributed optimization have been reported in these works [13], [19]–[24], where almost all of them employ randomized perturbation techniques in the message sharing steps [20]–[23] or in the cost functions [13].”
BackgroundPersA-FL : personalized asynchronous federated learning
Toghani, M. T., Lee, S., & Uribe, C. A. (2024) · Optimization methods & software · doi:10.1080/10556788.2023.2280056 · Cited by 10
“Nevertheless, FL poses challenges such as data heterogeneity (statistical diversity among clients) [11, 17, 34, 42], fairness [14, 43], privacy [24, 30, 55, 69], unreliable communication [60, 64], and staleness [3, 44, 53, 70].”
BackgroundMechanism Design for Demand Management in Energy Communities
Wei, X., & Anastasopoulos, A. (2021) · Games · doi:10.3390/g12030061 · Cited by 7
Computation of privacy-preserving prices in smart grids
Koufogiannis, F., Han, S., & Pappas, G. J. (2014) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2014.7039715 · Cited by 7
“From a theoretical point of view, providing privacy in distributed optimization problems can severely deteriorate the performance of the system when incorporated in a naive way [5].”
BackgroundDifferential Privacy of Online Distributed optimization under Adversarial Nodes
Hou, M., Li, D., Wu, X., & Shen, X. (2019) · Chinese Control Conference (CCC) · doi:10.23919/chicc.2019.8865820 · Cited by 7
Graph-Homomorphic Perturbations for Private Decentralized Learning
Vlaski, S., & Sayed, A. H. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9413811 · Cited by 7
Private and Accurate Decentralized Optimization via Encrypted and Structured Functional Perturbation
Zhou, Y., & Pu, S. (2022) · IEEE Control Systems Letters · doi:10.1109/lcsys.2022.3232210 · Cited by 7
“[12] observed that, by fixing the other parameters, the accuracy level of the obtained solution is in the order of O( 1 2 ), where is the privacy budget that is inversely proportional to privacy.”
InfluentialPrivacy-Preserving Dual Stochastic Push-Sum Algorithm for Distributed Constrained Optimization
Gu, C., Jiang, L., Li, J., & Wu, Z. (2023) · Journal of Optimization Theory and Applications · doi:10.1007/s10957-023-02173-9 · Cited by 7
“Our algorithm can tackle the DCO problems with globally coupled constraints by exploiting Lagrangian dual methods, which is different from the differentially private distributed methods [14, 17, 21, 38] that are suitable to solve the DCO problems without coupled constraints.”
MethodInfluentialPrivate Stochastic Dual Averaging for Decentralized Empirical Risk Minimization
Liu, C., Johansson, K. H., & Shi, Y. (2022) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2022.07.233 · Cited by 6
Distributed online bandit linear regressions with differential privacy
Dai, M., Ho, D. W. C., Zhang, B., Yuan, D., & Xu, S. (2023) · Journal of the Franklin Institute · doi:10.1016/j.jfranklin.2023.09.001 · Cited by 6
Differentially private dual gradient tracking for distributed resource allocation
Huo, W., Chen, X., Huang, L., Johansson, K. H., & Shi, L. (2025) · Automatica · doi:10.1016/j.automatica.2025.112521 · Cited by 6
Privacy-preserving federated Bayesian optimization with learnable noise
Liu, Q., Yan, Y., & Jin, Y. (2023) · Information Sciences · doi:10.1016/j.ins.2023.119739 · Cited by 6
Average Consensus for Switching Topology Networks with Privacy Protection
Wang, W., Li, D., Wu, X., & Xue, S. (2019) · Chinese Automation Congress · doi:10.1109/cac48633.2019.8996648 · Cited by 6
“One solution is differential privacy [13], [14], which randomizes query results by adding noise that satisfies a certain distribution.”
BackgroundMuffliato: Peer-to-Peer Privacy Amplification for Decentralized Optimization and AveragingPreprint
Cyffers, E., Even, M., Bellet, A., & Massoulié, L. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2206.05091 · Cited by 6
“[32] Zhenqi Huang, Sayan Mitra, and Nitin Vaidya.”
BackgroundMethodChen, Z., & Wang, Y. (2024) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2024.3519938 · Cited by 12
“Lemma 1: At time t ∈ N , if agent i injects into each of its transmitted messages a noise vector ϑ it consisting of n independent Laplace noises Lap ( ν it ) with parameter ν it such that (cid:5) Tt =1 Proof: The lemma can be obtained following the same line of reasoning of [24, Lemma 2].”
InfluentialPrivacy-Preserving Distributed Optimisation using Stochastic PDMM
Jordan, S. O., Li, Q., & Heusdens, R. R. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447844 · Cited by 5
“First, one can use differential privacy [8] to ensure user privacy, see e.g. [9, 10, 11, 12].”
BackgroundDifferentially Private Reward Functions for Markov Decision Processes
Benvenuti, A., Hawkins, C., Fallin, B., Chen, B., Bialy, B., Dennis, M., & Hale, M. (2024) · IEEE Conference on Control Technology and Applications (CCTA) · doi:10.1109/ccta60707.2024.10666610 · Cited by 5
Xing, M., Ma, D., Zhao, J., & Wong, P. K. (2024) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2024.3496488 · Cited by 5
Guaranteed Privacy of Distributed Nonconvex Optimization via Mixed-Monotone Functional Perturbations
Khajenejad, M., & Martínez, S. (2022) · IEEE Control Systems Letters · doi:10.1109/lcsys.2022.3231223 · Cited by 5
Differential Private Discrete Noise Adding Mechanism: Conditions and Properties
Qin, S., He, J., Fang, C., & Lam, J. (2022) · Proceedings of the ... American Control Conference/Proceedings of the American Control Conference · doi:10.23919/acc53348.2022.9867334 · Cited by 5
“To quantify the discrete inputs of the mechanism A , we define the m -adjacency by referring to the studies in [11].”
MethodPrivacy in Multi-Agent Systems
Wang, Y. (2024) · Elsevier eBooks · doi:10.1016/b978-0-443-14081-5.00089-1 · Cited by 5
Hu, R., Xu, W., Sun, L. N., & Cao, J. (2025) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2025.3579593 · Cited by 4
“1 / 2 (cid:4) , which aligns with the findings in [29].”
InfluentialCrowd-ML: A library for privacy-preserving machine learning on smart devices
Hamm, J., Luken, J., & Xie, Y. (2017) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2017.7953387 · Cited by 4
“However, noise can be chosen to decrease geometrically [8], decrease proportionally to the learning rate (see Sec.”
BackgroundDifferentially private formation control: Privacy and network co-design
Hawkins, C., & Hale, M. (2026) · Systems & Control Letters · doi:10.1016/j.sysconle.2026.106434 · Cited by 4
Distributed submodular maximization: trading performance for privacy
Rezazadeh, N., & Kia, S. S. (2022) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc51059.2022.9992452 · Cited by 4
“Our privacy preservation mechanism differs from existing methods for distributed algorithms such as differential privacy [18], [19], which rely on adding additive noises to the inter-agent communication messages.”
MethodMulti-owner multi-user privacy
Koufogiannis, F., & Pappas, G. J. (2016) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2016.7798524 · Cited by 8
“The literature of differential privacy was initiated in a database context [3], [5], such as recommender systems [4], and was later extended to Kalman filtering [6], optimization problems [7], consensus algorithms [8] etc.”
Xiao, Y., Hou, X., Cai, J., & Hu, J. (2018) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc.2018.8619017 · Cited by 4
“, filtering problem [16], electric vehicle charging schedule problem [17], distributed optimization problem [18], [19], consensus problem [20], [21], and distributed control [22].”
BackgroundPrivacy-Preserving Quantum Annealing for Quadratic Unconstrained Binary Optimization (QUBO) Problems
Xie, M., Zhang, Y., Zhong, S., & Li, Q. (2024) · IEEE International Conference on Quantum Computing and Engineering (QCE) · doi:10.1109/qce60285.2024.00160 · Cited by 4
“Notable works include [12]–[21].”
Dynamics Based Privacy Protection for Average Consensus on Directed GraphsPreprint
Gao, H. B., & Wang, Y. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1812.02255 · Cited by 5
Differentially Private Distributed Computation via Public-Private Communication NetworksPreprint
Wang, L., Liu, Y., Manchester, I. R., & Shi, G. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2101.01376 · Cited by 4
“We artificially and evenly allocate all training samples to all nodes of G in Figure 4a, so that all three protocols, including the PPSC-Gossip-DCO algorithm with the private network Gp in Figure 4c, Huang2015 [9], Nozari2018 [12], can be executed over G with uniform privacy budgets = 0.001, 0.01, 0.1 for L = 3000 ste…”
BackgroundMethodInfluentialEnsuring Truthfulness in Distributed Aggregative Optimization
Chen, Z., Egerstedt, M., & Wang, Y. (2025) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2025.3606265 · Cited by 3
Adversarial Robustness in Distributed Quantum Machine Learning
Kananian, P., & Jacobsen, H. (2026) · Quantum science and technology · doi:10.1007/978-3-032-11153-1_11 · Cited by 3
A Communication-Efficient Local Differentially Private Algorithm in Federated Optimization
Alam, S. E., Shukla, D., & Rao, S. (2023) · IEEE Access · doi:10.1109/access.2023.3283503 · Cited by 3
“In another work on differentially private distributed algorithms, Huang et al. [75] also use convex functions.”
Distributed Online Convex Optimization With Statistical Privacy
Dai, M., Ho, D. W. C., Zhang, B., Yuan, D., & Xu, S. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3492144 · Cited by 3
“Another popular method to preserve privacy is differential privacy [29], [30], [31], [32], which uses independent randomized noises to cover the local information of participating agents.”
Bipartite Consensus for Multi-Agent Systems with Differential Privacy Constraint
Zuo, Z., Tian, R., & Wang, Y. (2021) · Chinese Control Conference (CCC) · doi:10.23919/ccc52363.2021.9550694 · Cited by 3
“By the method of adding random noises to the agents’ states at each iteration, the differential privacy mechanism can be achieved for MASs [27, 28], Nozari et al. [29] studied average consensus problem of the multi-agent based on the condition of differential privacy, which has a close relationship…”
MethodPrivacy-Preserving Average Consensus via Homomorphic Encryption
Qi, M., Wang, F., Liu, Z., & Chen, Z. (2023) · Chinese Control Conference (CCC) · doi:10.23919/ccc58697.2023.10241210 · Cited by 3
Near-optimal Differentially Private Client Selection in Federated Settings
Alam, S. E., Shukla, D., & Rao, S. (2023) · Allerton Conference on Communication, Control, and Computing · doi:10.1109/allerton58177.2023.10313432 · Cited by 3
“Huang et al. in [20] developed differentially private distributed algorithms using convex functions.”
BackgroundLiu, Z., Wang, W., & Guo, F. (2023) · Annual Conference of the IEEE Industrial Electronics Society (IECON) · doi:10.1109/iecon51785.2023.10312089 · Cited by 3
Compressed Differentially Private Distributed Optimization with Linear Convergence
Xie, A., Yi, X., Wang, X., Cao, M., & Ren, X. (2023) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2023.10.1029 · Cited by 3
“To preserve the privacy of each agent, Huang et al. (2015); Zhu et al. (2018); Ding et al. (2021); Chen et al. (2021) proposed several differentially private distributed optimization algorithms by introducing the notion of dif- ferential privacy (Dwork, 2008), which has mathematically provable…”
BackgroundMethodDifferentially Private Nash Equilibrium Seeking in Quadratic Network Games
Wang, L., Ding, K., Leng, Y., Ren, X., & Shi, G. (2024) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2024.3398729 · Cited by 3
“Of more relevance to our article are the works in [18], [19], [20], and [21] focusing on distributed differentially private optimization.”
Local Differential Privacy in Decentralized OptimizationPreprint
Xiao, H., Ye, Y., & Devadas, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1902.06101 · Cited by 5
“The second line corresponds to the differentially private optimization schemes (Zhang & Zhu, 2017), (Zhang et al., 2018a), (Guo & Gong, 2018), (Huang et al., 2015), (Han et al., 2017), (Lou et al., 2018).”
BackgroundMethodInfluentialDifferentially Private Distributed Online LearningPreprint
Li, C., & Zhou, P. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1505.06556 · Cited by 3
“Besides, Huang et al. [14] is closely related to our work.”
BackgroundA Private and Finite-Time Algorithm for Solving a Distributed System of Linear EquationsPreprint
Gade, S., Liu, J., & Vaidya, N. H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2004.04680 · Cited by 3
“Differential privacy is employed in [14], [17] for distributed convex optimization, however, it suffers from a fundamental privacy - accuracy trade-off [16].”
BackgroundSelf-citationFully Privacy-Preserving Distributed Optimization Based on Secret SharingPreprint
Tian, N., Guo, Q., Sun, H., & Zhou, X. (2021) · doi:10.36227/techrxiv.15087774.v1 · Cited by 3
Differentially-Private Distributed Optimization with Guaranteed Optimality
Wang, Y., & Nedić, A. (2023) · IEEE Conference on Decision and Control (CDC) · doi:10.1109/cdc49753.2023.10383285 · Cited by 3
“Allowing stepsize sequences to be non-summable is key to avoiding optimization errors in [34] and achieve almost sure convergence.”
BackgroundMethodInfluentialPPDOG: Privacy-Preserving Distributed Optimization With Performance Guarantees in Networked Systems
Yu, D., Shi, X., Lin, Y., Wu, C., & Chen, J. (2025) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3582347 · Cited by 2
“Huang et al. established foundational work by demonstrating that achieving (cid:2) -DP necessitates algorithm convergence accuracy of O ( 1 2 ) [50].”
Zhang, K., Wang, Y., Song, Z., & Li, Z. (2025) · IEEE Transactions on Automatic Control · doi:10.1109/tac.2025.3627270 · Cited by 2
“In addition, the saddle-point theorem holds, i.e., given an optimal primal– dual pair ( { ˜ u ∗ i } , λ ∗ ) , the following relationship holds for any λ ∈ R Np + and ˜ u i ∈ ˜ U i ( x i ( t )) : A standard approach for solving (9) is the distributed dual-gradient method [6], [22].”
InfluentialDifferentially private projected gradient tracking for distributed constrained optimization
Yang, Z., He, W., & Chen, G. (2026) · Automatica · doi:10.1016/j.automatica.2026.112851 · Cited by 2
Differentially Private Accelerated Distributed Algorithm for Aggregative Optimization
Liu, B., Li, D., & Chai, L. (2026) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2026.3674758 · Cited by 2
“…privacy-preserving techniques, di ff erential privacy (DP) has gained significant attention due to its rigorous theoretical foundation, strong privacy guarantees, and ease of implementation [28], which distinguish it from encryption-based approaches [29] and state-decomposition-based methods [30].”
Differentially Private Multi-Agent Constraint Optimization
Damle, S., Triastcyn, A., Faltings, B. V., & Gujar, S. (2021) · IEEE/WIC/ACM International Conference on Web Intelligence · doi:10.1145/3486622.3493929 · Cited by 2
State Estimation with Differential Privacy under Scale Enlargement of the Sensor Network
Yan, X., & Yang, C. (2022) · Asian Control Conference (ASCC) · doi:10.23919/ascc56756.2022.9828316 · Cited by 2
Distributed Load Shedding via Differentially Private Average Consensus Algorithm
Qiao, Z., Guo, F., Pan, X., Sun, Y., & Wang, L. (2022) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc55256.2022.10034229 · Cited by 2
“In the last decades, such notion has received a significant attention and become a golden metric in the data privacy due to its rigorous formulation and robustness to post-processing and side information, with extensive applications in the fields of learning [16] and control [17].”
BackgroundZeroth-order gradient tracking for decentralized learning with privacy guarantees
Zhao, Z., Xia, L., Jiang, L., Gao, W., & Ge, Q. (2024) · ISA Transactions · doi:10.1016/j.isatra.2024.06.033 · Cited by 2
Guaranteed Feasibility in Differentially Private Linearly Constrained Convex Optimization
Benvenuti, A., Bialy, B., Dennis, M., & Hale, M. (2024) · IEEE Control Systems Letters · doi:10.1109/lcsys.2024.3513232 · Cited by 2
Privacy-Preserving Distributed Maximum Consensus Without Accuracy Loss
Yu, W., Heusdens, R. R., Pang, J., & Li, Q. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888860 · Cited by 2
“Common privacy preservation techniques in consensus problems include differential privacy (DP) [4]–[8], secure multi-party computation (SMPC) [9]–[16], subspace perturbation [17]–[19] and variants of it [20]–[22].”
Metadata
- Authors
- Huang, Z., Mitra, S., & Vaidya, N.
- Affiliations
- University of Illinois Urbana-Champaign (US)
- Venue
- Proceedings of the 16th International Conference on Distributed Computing and Networking
- Publisher
- ACM
- Published
- 2015-01-04
- Pages
- 1-10
- OpenAlex
- W2092845654
- Cited by
- 266
- References
- 17
- Citations by year
- 2026: 15 · 2025: 24 · 2024: 37 · 2023: 33 · 2022: 27 · 2021: 24 · 2020: 23 · 2019: 26 · 2018: 23 · 2017: 14 · 2016: 12 · 2015: 5
- FWCI
- 10.14
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- International Conference on Distributed Computing and Networking (ICDCN) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Science Foundation (CAREER CNS 10-54247)
- NSA (SoS W911NSF-13-0086)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Engineering › Control and Systems Engineering › Distributed Control Multi-Agent Systems
- Physical Sciences › Computer Science › Computer Networks and Communications › Distributed Sensor Networks and Detection Algorithms
- Keywords
- differential privacy, distributed optimization, multi-agent systems, privacy-accuracy trade-off, convergence analysis, iterative methods, privacy-preserving optimization
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.