Filters








61 Hits in 4.0 sec

Federated Heavy Hitters Discovery with Differential Privacy [article]

Wennan Zhu, Peter Kairouz, Brendan McMahan, Haicheng Sun, Wei Li
2020 arXiv   pre-print
The discovery of heavy hitters (most frequent items) in user-generated data streams drives improvements in the app and web ecosystems, but can incur substantial privacy risks if not done with care.  ...  Finally, we carefully compare our approach to Apple's local differential privacy method for discovering heavy hitters.  ...  Our work differs in that it focuses on federated algorithms for the discovery of heavy hitters.  ... 
arXiv:1902.08534v4 fatcat:ymcshac2tvca5ieyxoshqv6eky

Sample and Threshold Differential Privacy: Histograms and applications [article]

Akash Bharadwaj, Graham Cormode
2022 arXiv   pre-print
Using such histograms, related problems such as heavy hitters and quantiles can be answered with provable error and privacy guarantees.  ...  In this paper, we show how a strong (ϵ, δ)-Differential Privacy (DP) guarantee can be achieved for the fundamental problem of histogram generation in a federated setting, via a highly practical sampling-based  ...  set to O( √ n) for heavy hitter discovery by Zhu et al. (2020) .  ... 
arXiv:2112.05693v2 fatcat:2jcuwpigv5gpdmsl6myqpzoeyy

Secure Bayesian Federated Analytics for Privacy-Preserving Trend Detection [article]

Amit Chaulwar, Michael Huth
2021 arXiv   pre-print
We propose a protocol, named SAFE, for Bayesian federated analytics that offers sufficient privacy for production grade use cases and reduces the computational burden of users and an aggregator.  ...  We illustrate this approach with a trend detection experiment and discuss how this approach could be extended further to make it production-ready.  ...  Federated heavy-hitters discovery (Zhu et al., 2020) is the latest privacy-preserving method that can be used to detect heavy-hitters. However, it is not suitable for finding new trends.  ... 
arXiv:2107.13640v1 fatcat:mvgxsf6u5feihmd7jtgr7lgmbu

A Comprehensive Survey on Local Differential Privacy

Xingxing Xiong, Shubo Liu, Dan Li, Zhaohui Cai, Xiaoguang Niu
2020 Security and Communication Networks  
With the advent of the era of big data, privacy issues have been becoming a hot topic in public.  ...  Local differential privacy (LDP) is a state-of-the-art privacy preservation technique that allows to perform big data analysis (e.g., statistical estimation, statistical learning, and data mining) while  ...  private heavy hitter identification problem.  ... 
doi:10.1155/2020/8829523 fatcat:xjk3vgyambb5xioc2q5hyr2hua

Towards Sparse Federated Analytics: Location Heatmaps under Distributed Differential Privacy with Secure Aggregation [article]

Eugene Bagdasaryan, Peter Kairouz, Stefan Mellem, Adrià Gascón, Kallista Bonawitz, Deborah Estrin, Marco Gruteser
2022 arXiv   pre-print
To achieve this, we revisit distributed differential privacy based on recent results in secure multiparty computation, and we design a scalable and adaptive distributed differential privacy approach for  ...  It aims to ensure differential privacy before data becomes visible to a service provider while maintaining high data accuracy and minimizing resource consumption on users' devices.  ...  Indeed, our work is most closely aligned with the TrieHH algorithm in [67] , an adaptive algorithm for learning heavy hitters with differential privacy.  ... 
arXiv:2111.02356v2 fatcat:5sgu3qkmfbejrknm2xtlpcb7wq

FED-χ^2: Privacy Preserving Federated Correlation Test [article]

Lun Wang, Qi Pang, Shuai Wang, Dawn Song
2021 arXiv   pre-print
In this paper, we propose the first secure federated χ^2-test protocol Fed-χ^2.  ...  To minimize both the privacy leakage and the communication cost, we recast χ^2-test to the second moment estimation problem and thus can take advantage of stable projection to encode the local information  ...  [10] proposed new schemes for federated private mean estimation and heavy hitter estimation. Zhu et al. [46] proposed distributed and privacy-preserving algorithms for heavy hitter estimation.  ... 
arXiv:2105.14618v1 fatcat:mnjxfd3lhvcwtl5sh72pwh3p4e

A Comprehensive Survey on Local Differential Privacy Toward Data Statistics and Analysis in Crowdsensing [article]

Teng Wang, Xuefeng Zhang, Jingyu Feng, Xinyu Yang
2020 arXiv   pre-print
Local differential privacy (LDP) has been proposed as an excellent and prevalent privacy model with distributed architecture, which can provide strong privacy guarantees for each user while collecting  ...  This survey presents a comprehensive and systematic overview of LDP with respect to privacy models, research tasks, enabling mechanisms, and various applications.  ...  heavy hitters discovery [31] , k-way marginal release [32] , empirical risk minimization (ERM) [33] , federated learning [34] , and deep learning [35] .  ... 
arXiv:2010.05253v2 fatcat:uuts5enifreixjt4lf6yjrepl4

A Comprehensive Survey on Local Differential Privacy toward Data Statistics and Analysis

Teng Wang, Xuefeng Zhang, Jingyu Feng, Xinyu Yang
2020 Sensors  
Local differential privacy (LDP) was proposed as an excellent and prevalent privacy model with distributed architecture, which can provide strong privacy guarantees for each user while collecting and analyzing  ...  This survey presents a comprehensive and systematic overview of LDP with respect to privacy models, research tasks, enabling mechanisms, and various applications.  ...  Frequent Items Mining Frequent items mining (also known as heavy hitters identification, top-ω hitters mining, or frequent terms discovery) has played important roles in data statistics.  ... 
doi:10.3390/s20247030 pmid:33302517 pmcid:PMC7763193 fatcat:25iufaivynabdftrzq4rzxsz2e

Locally Private Graph Neural Networks [article]

Sina Sajadmanesh, Daniel Gatica-Perez
2021 arXiv   pre-print
To address this problem, we develop a privacy-preserving, architecture-agnostic GNN learning algorithm with formal privacy guarantees based on Local Differential Privacy (LDP).  ...  Extensive experiments conducted over real-world datasets demonstrate that our method can maintain a satisfying level of accuracy with low privacy loss.  ...  There are also some works focusing on learning problems, such as probability distribution estimation [3, 12, 22] , heavy hitter discovery [6, 8, 54] , frequent new term discovery [49] , marginal release  ... 
arXiv:2006.05535v9 fatcat:xqzrddss6zhzhgzcllxvq53qeu

Locally Private Graph Neural Networks

Sina Sajadmanesh, Daniel Gatica-Perez
2021 Zenodo  
To address this problem, we develop a privacy-preserving, architecture-agnostic GNN learning algorithm with formal privacy guarantees based on Local Differential Privacy (LDP).  ...  Extensive experiments conducted over real-world datasets demonstrate that our method can maintain a satisfying level of accuracy with low privacy loss.  ...  There are also some works focusing on learning problems, such as probability distribution estimation [3, 12, 22] , heavy hitter discovery [6, 8, 54] , frequent new term discovery [49] , marginal release  ... 
doi:10.5281/zenodo.5081878 fatcat:5s4qnep7trbi3pjvmpz5fbg3kq

Secure Federated Submodel Learning [article]

Chaoyue Niu, Fan Wu, Shaojie Tang, Lifeng Hua, Rongfei Jia, Chengfei Lv, Zhihua Wu, Guihai Chen
2019 arXiv   pre-print
Our secure scheme features the properties of randomized response, secure aggregation, and Bloom filter, and endows each client with a customized plausible deniability, in terms of local differential privacy  ...  To integrate efficiency and privacy, we have designed a secure federated submodel learning scheme coupled with a private set union protocol as a cornerstone.  ...  in differential privacy.  ... 
arXiv:1911.02254v2 fatcat:aankusilkjfbtcwboh6odhx4ju

Privacy-preserving Wi-Fi Analytics

Mohammad Alaggan, Mathieu Cunche, Sébastien Gambs
2018 Proceedings on Privacy Enhancing Technologies  
To solve this issue, we propose a privacy-preserving solution for collecting aggregate mobility patterns while satisfying the strong guarantee of ε-differential privacy.  ...  Thus, while there is a strong interest for physical analytics platforms to leverage this information for many purposes, this tracking also threatens the privacy of individuals.  ...  Namely they presented algorithms to estimate the count of distinct events, but also other statistics such as t-Cropped Mean, k-Heavy Hitters, and t-incidence [25] .  ... 
doi:10.1515/popets-2018-0010 dblp:journals/popets/AlagganCG18 fatcat:q3pqvgsepzf7fhpg4l5yajwplm

An Exhaustive Survey on P4 Programmable Data Plane Switches: Taxonomy, Applications, Challenges, and Future Trends [article]

Elie F. Kfoury, Jorge Crichigno, Elias Bou-Harb
2021 arXiv   pre-print
Traditionally, the data plane has been designed with fixed functions to forward packets using a small set of protocols.  ...  The paper then presents a unique, comprehensive taxonomy of applications developed with P4 language; surveying, classifying, and analyzing more than 150 articles; discussing challenges and considerations  ...  TABLE 28 : 28 Heavy hitter schemes comparison.  ... 
arXiv:2102.00643v2 fatcat:izxi645kozdc5ibfsqp2y2foau

An Exhaustive Survey on P4 Programmable Data Plane Switches: Taxonomy, Applications, Challenges, and Future Trends

Elie F. Kfoury, Jorge Crichigno, Elias Bou-Harb
2021 IEEE Access  
Traditionally, the data plane has been designed with fixed functions to forward packets using a small set of protocols.  ...  The paper then presents a unique, comprehensive taxonomy of applications developed with P4 language; surveying, classifying, and analyzing more than 200 articles; discussing challenges and considerations  ...  HEAVY HITTER 1) Background Heavy hitters are a small number of flows that constitute most of the network traffic over a certain amount of time.  ... 
doi:10.1109/access.2021.3086704 fatcat:2jgbxj2cbfbp7fawkxwrztbbia

FedPAQ: A Communication-Efficient Federated Learning Method with Periodic Averaging and Quantization [article]

Amirhossein Reisizadeh, Aryan Mokhtari, Hamed Hassani, Ali Jadbabaie, Ramtin Pedarsani
2020 arXiv   pre-print
In this paper, we present FedPAQ, a communication-efficient Federated Learning method with Periodic Averaging and Quantization.  ...  Due to these systems challenges as well as issues related to statistical heterogeneity of data and privacy concerns, designing a provably efficient federated learning method is of significant importance  ...  Federated heavy hitters discovery with differential privacy was proposed in [Zhu et al., 2019] .  ... 
arXiv:1909.13014v4 fatcat:4l66y4ksirdfzcxkc5oss33lpq
« Previous Showing results 1 — 15 out of 61 results