7,697 Hits in 4.8 sec

Privacy for Free: Communication-Efficient Learning with Differential Privacy Using Sketches [article]

Tian Li, Zaoxing Liu, Vyas Sekar, Virginia Smith
2019 arXiv   pre-print
Using these derived privacy guarantees, we propose a novel sketch-based framework (DiffSketch) for distributed learning, where we compress the transmitted messages via sketches to simultaneously achieve  ...  communication efficiency and provable privacy benefits.  ...  A Proof for Theorem 2 Theorem: (ε-differential privacy of Count Sketch) For a sketching algorithm M using Count Sketch S t×k with t arrays of k bins, for any input vector D with length n satisfying Assumption  ... 
arXiv:1911.00972v2 fatcat:hmsabj5wyjem5ikmgtb5yyeso4

How to Make Private Distributed Cardinality Estimation Practical, and Get Differential Privacy for Free [article]

Changhui Hu, Jin Li, Zheli Liu, Xiaojie Guo, Yu Wei, Xuan Guang, Grigorios Loukides, Changyu Dong
2020 IACR Cryptology ePrint Archive  
On the other hand, the use of sketches has gained popularity in data mining, because sketches often give rise to highly efficient and scalable sub-linear algorithms.  ...  We investigated the question and the findings are interesting: we can get security, we can get scalability, and somewhat unexpectedly, we can also get differential privacy -for free.  ...  Acknowledgement We thank shepherd Mathias Lécuyer as well as the anonymous reviewers for their insightful comments.  ... 
dblp:journals/iacr/HuLLGWGLD20 fatcat:35wk2lx6hbgttph67bhwutkuxq

Privacy for 5G-Supported Vehicular Networks

Meng Li, Liehuang Zhu, Zijian Zhang, Chhagan Lal, Mauro Conti, Fabio Martinelli
2021 IEEE Open Journal of the Communications Society  
Meanwhile, the development of such communication capabilities, along with the development of sensory devices and the enhancement of local computing powers, have lead to an inevitable reality of massive  ...  Unfortunately, 5GVN are still confronted with a variety of privacy threats. Such threats are targeted at users' data, identity, location, and trajectory.  ...  (Priv-2SP-SP) [7] Differentially private scheduling for ridesharing (JDP-Ride) [8] Privacy-preserving and accountable ride-hailing (ORide) [9] Privacy-enhanced ride-hailing (PrivateRide) [109] Efficient  ... 
doi:10.1109/ojcoms.2021.3103445 fatcat:cznswnoczrb4dgfg7hhiywifpe

The Price of Differential Privacy for Low-Rank Factorization [article]

Jalaj Upadhyay
2018 arXiv   pre-print
Our algorithms when private matrix is updated in an arbitrary manner promise differential privacy with respect to two stronger privacy guarantees than previously studied, use space and time comparable  ...  When data matrices are distributed over multiple servers, we give a non-interactive differentially private algorithm with communication cost independent of dimension.  ...  The author wishes to thank Adam Smith for many useful discussions. Even though he declined to be coauthor in this paper, this paper would have not existed without his insightful inputs.  ... 
arXiv:1604.01429v5 fatcat:eikycsbu3ncklpoini4mzmeh4m

SecureML: A System for Scalable Privacy-Preserving Machine Learning

Payman Mohassel, Yupeng Zhang
2017 2017 IEEE Symposium on Security and Privacy (SP)  
In this paper, we present new and efficient protocols for privacy preserving machine learning for linear regression, logistic regression and neural network training using the stochastic gradient descent  ...  Machine learning is widely used in practice to produce predictive models for applications such as image processing, speech and text recognition.  ...  Acknowledgements We thank Jing Huang from Visa Research for helpful discussions on machine learning, and Xiao Wang from University of Maryland for his help on the EMP toolkit.  ... 
doi:10.1109/sp.2017.12 dblp:conf/sp/MohasselZ17 fatcat:y6lwerwfwnfghjkia76yqenmiy

Generalised Differential Privacy for Text Document Processing [chapter]

Natasha Fernandes, Mark Dras, Annabelle McIver
2019 Research Series on the Chinese Dream and China's Development Path  
In this paper we combine ideas from "generalised differential privacy" and machine learning techniques for text processing to model privacy for text documents.  ...  We show that our mechanism satisfies privacy with respect to a metric for semantic similarity, thereby providing a balance between utility, defined by the semantic content of texts, with the obfuscation  ...  Machine Learning and Differential Privacy.  ... 
doi:10.1007/978-3-030-17138-4_6 dblp:conf/post/FernandesDM19 fatcat:unwda7ocnrcz5o4dihwlct6veq

Generalised Differential Privacy for Text Document Processing [article]

Natasha Fernandes, Mark Dras, Annabelle McIver
2019 arXiv   pre-print
In this paper we combine ideas from "generalised differential privacy" and machine learning techniques for text processing to model privacy for text documents.  ...  We show that our mechanism satisfies privacy with respect to a metric for semantic similarity, thereby providing a balance between utility, defined by the semantic content of texts, with the obfuscation  ...  Machine Learning and Differential Privacy Outside of author attribution, there is quite a body of work on introducing differential privacy to machine learning: [13] gives an overview of a classical machine  ... 
arXiv:1811.10256v2 fatcat:pdhkjngqvvgsvlsonv2vexgjg4

FedCG: Leverage Conditional GAN for Protecting Privacy and Maintaining Competitive Performance in Federated Learning [article]

Yuezhou Wu, Yan Kang, Jiahuan Luo, Yuanqin He, Qiang Yang
2022 arXiv   pre-print
terms of prediction accuracy (e.g., with differential privacy).  ...  Federated learning (FL) aims to protect data privacy by enabling clients to build machine learning models collaboratively without sharing their private data.  ...  The discriminators communicate with each other in a peer-to-peer fashion for improving computational efficiency.  ... 
arXiv:2111.08211v2 fatcat:oebydezpfnfijmgvggebd6jdb4

Security and privacy for 6G: A survey on prospective technologies and challenges [article]

Van-Linh Nguyen, Po-Ching Lin, Bo-Chao Cheng, Ren-Hung Hwang, Ying-Dar Lin
2021 arXiv   pre-print
However, for now, security and privacy issues for 6G remain largely in concept.  ...  This survey provides a systematic overview of security and privacy issues based on prospective technologies for 6G in the physical, connection, and service layers, as well as through lessons learned from  ...  Blockchain, distributed learning, federated learning, homomorphic encryption, and differential privacy are potential technologies that support protecting personal data, complying with GDPR standards, and  ... 
arXiv:2108.11861v1 fatcat:o7fgtm4tobfk5knltojut6xpyi

Contextual Linear Types for Differential Privacy [article]

Matías Toro, David Darais, Chike Abuah, Joe Near, Damián Árquez, Federico Olmedo, Éric Tanter
2021 arXiv   pre-print
recent differential privacy literature.  ...  We present Jazz, a language and type system which uses linear types and latent contextual effects to support both advanced variants of differential privacy and higher-order programming.  ...  Differentially Private Deep Learning with Adaptive Clipping The current state-of-the-art in differentially private machine learning is noisy gradient descent [3] : at each iteration of training, compute  ... 
arXiv:2010.11342v2 fatcat:gwgcsvx2mzewlpui56256g4noa

Decentralized Systems for Privacy Preservation (Dagstuhl Seminar 13062)

Sonja Buchegger, Jon Crowcroft, Balachander Krishnamurthy, Thorsten Strufe, Marc Herbstritt
2013 Dagstuhl Reports  
This report documents the program and the outcomes of Dagstuhl Seminar 13062 "Decentralized Systems for Privacy Preservation".  ...  The results were mixed: some clear agenda setting outputs emerged with some less clear ones.  ...  The customers get a service for free in return for risking their privacy.  ... 
doi:10.4230/dagrep.3.2.22 dblp:journals/dagstuhl-reports/BucheggerCKS13 fatcat:6skor4xlgbgq7ivvnz26qep6om

Achieving Better Privacy for the 3GPP AKA Protocol

Pierre-Alain Fouque, Cristina Onete, Benjamin Richard
2016 Proceedings on Privacy Enhancing Technologies  
Proposed by the 3rd Generation Partnership Project (3GPP) as a standard for 3G and 4G mobile-network communications, the AKA protocol is meant to provide a mutually-authenticated key-exchange between clients  ...  In this paper we use the approach of provable security to show that these variants still fail to guarantee the privacy of mobile clients.  ...  communication is that of client privacy.  ... 
doi:10.1515/popets-2016-0039 dblp:journals/popets/FouqueOR16 fatcat:urbqt6xemnglbfqieedt7qyf4e

Do privacy laws affect the location decisions of internet firms? Evidence for privacy havens

Fabrice Rochelandet, Silvio H. T. Tai
2013 European Journal of Law and Economics  
One way for a firm to avoid such legal restrictions is to locate or to expand its business in less privacy protective countries. Our empirical results support this 'noprivacy haven' hypothesis.  ...  So the more a jurisdiction makes collecting and using these data easy, the more attractive the country is, if all other things remain constant.  ...  These aspects are potentially significant for internet firms looking for efficient communication networks.  ... 
doi:10.1007/s10657-013-9428-6 fatcat:xmythmqvbzhm7pqprkiq5wv7da

Koi: A Location-Privacy Platform for Smartphone Apps

Saikat Guha, Mudit Jain, Venkata N. Padmanabhan
2012 Symposium on Networked Systems Design and Implementation  
is easier for applications to build upon.  ...  We verify the non-tracking properties of Koi using a theorem prover, illustrate how privacy guarantees can easily be added to a wide range of location-based applications, and show that our public deployment  ...  Koi's privacy model differ from PIR and SMC, thus allowing for a much more efficient protocol that has constant communication and computational overhead.  ... 
dblp:conf/nsdi/GuhaJP12 fatcat:7iguou7l55b6be2l3i5ioen5vu

No free lunch in data privacy

Daniel Kifer, Ashwin Machanavajjhala
2011 Proceedings of the 2011 international conference on Management of data - SIGMOD '11  
Differential privacy is a powerful tool for providing privacypreserving noisy query answers over statistical databases.  ...  In both cases the notion of participation varies, the use of differential privacy can lead to privacy breaches, and differential privacy does not always adequately limit inference about participation.  ...  Let us start with two more forms of differential privacyattribute and bit differential privacy. Definition 2.4. (Attribute Differential Privacy).  ... 
doi:10.1145/1989323.1989345 dblp:conf/sigmod/KiferM11 fatcat:fmaigs6zqbam5erf3dq2osntqe
« Previous Showing results 1 — 15 out of 7,697 results