92 Hits in 3.1 sec

Private multiparty sampling and approximation of vector combinations

Yuval Ishai, Tal Malkin, Martin J. Strauss, Rebecca N. Wright
2009 Theoretical Computer Science  
Moreover, these results have some interesting consequences for the general problem of reducing sublinear-communication secure multiparty computation to two-party private information retrieval (PIR).  ...  Sublinear-communication private protocols have primarily been studied only in the two-party case. In contrast, this work focuses on multiparty settings.  ...  In this work, we mostly deal with the special case of secure computation in the semi-honest model, or private computation.  ... 
doi:10.1016/j.tcs.2008.12.062 fatcat:cmt6mwibyjen7ogguojyrdkfh4

A Computational Separation between Private Learning and Online Learning [article]

Mark Bun
2020 arXiv   pre-print
A recent line of work has shown a qualitative equivalence between differentially private PAC learning and online learning: A concept class is privately learnable if and only if it is online learnable with  ...  Studying a special case of this connection, Gonen, Hazan, and Moran (NeurIPS 2019) showed that uniform or highly sample-efficient pure-private learners can be time-efficiently compiled into online learners  ...  In this work, we address the parallel issue of computational complexity. Could there be a computationally efficient black-box conversion from any private learner to an online learner?  ... 
arXiv:2007.05665v1 fatcat:wljwfhh2avcgxfevzglvcv37me

Efficient Private Statistics with Succinct Sketches [article]

Luca Melis and George Danezis and Emiliano De Cristofaro
2016 arXiv   pre-print
of median statistics for Tor hidden services.  ...  In this paper, we present the design and implementation of practical techniques for privately gathering statistics from large data streams.  ...  We would like to thank Chris Newell and Michael Smethurst from the BBC and Aaron Johnson from US Naval Research Labs for motivating our work, respectively, on privacy-preserving recommendation and median  ... 
arXiv:1508.06110v3 fatcat:oqrtjnqvr5hzjjmqxcbpxruoiq

Signal Processing and Machine Learning with Differential Privacy: Algorithms and Challenges for Continuous Data

Anand D. Sarwate, Kamalika Chaudhuri
2013 IEEE Signal Processing Magazine  
[ fIg2] a comparison of computing the mean and the median. (a) outputs of 1,000 runs of the differentially private sample mean algorithm.  ...  (b) outputs of 1,000 runs of the differentially private sample median algorithm.  ... 
doi:10.1109/msp.2013.2259911 pmid:24737929 pmcid:PMC3984544 fatcat:37k4byrgbbgdxcqq76vpv5idzy

Differential Privacy for Statistics: What we Know and What we Want to Learn

Cynthia Dwork, Adam Smith
2010 Journal of Privacy and Confidentiality  
We motivate and review the definition of differential privacy, survey some results on differentially private statistical estimators, and outline a research agenda.  ...  This means that privacy is guaranteed only against adversaries with limited computing time (as is security in modern cryptography).  ...  Mironov and Pandey show a simple computationally differentially private protocol for computing the inner product of two binary vectors (each party holds one of the vectors) with an additive error depending  ... 
doi:10.29012/jpc.v1i2.570 fatcat:ox37yqx7fvb3ncavthwdtwhrii

Differential Privacy in Practice

Hiep H. Nguyen, Jong Kim, Yoonho Kim
2013 Journal of Computing Science and Engineering  
Category: Smart and intelligent computing  ...  Promises of differential privacy help to relieve concerns of privacy loss, which hinder the release of community-valuable data.  ...  For the MapReduce framework, Airavat [54] combined mandatory access control and differential privacy to provide strong security and privacy guarantees for distributed computations on sensitive data.  ... 
doi:10.5626/jcse.2013.7.3.177 fatcat:xrcbyxpzfvh2tfdrr5572rdnku

Pyramid: Enhancing Selectivity in Big Data Protection with Count Featurization [article]

Mathias Lecuyer and Riley Spahn and Roxana Geambasu and Tzu-Kuo Huang and Siddhartha Sen
2017 arXiv   pre-print
We compute differentially private percentiles by adapting the J-List algorithm for the differentially private median described in [31] .  ...  While our IEEE Security & Privacy paper [30] did not address this problem, we have since modified Pyramid to compute feature weights in a differentially private way.  ... 
arXiv:1705.07512v1 fatcat:u2ey66wplnbq7o5oeryjhp75q4

Adversarially Robust Streaming Algorithms via Differential Privacy [article]

Avinatan Hassidim, Haim Kaplan, Yishay Mansour, Yossi Matias, Uri Stemmer
2020 arXiv   pre-print
We establish a connection between adversarial robustness of streaming algorithms and the notion of differential privacy.  ...  other fields, such as machine learning, mechanism design, secure computation, probability theory, secure storage, and more  ...  done using a differentially private algorithm for approximating the median of the responses.  ... 
arXiv:2004.05975v1 fatcat:kgdgss37gjdb5k4zlxo3cmw22m

Pan-private algorithms via statistics on sketches

Darakhshan Mir, S. Muthukrishnan, Aleksandar Nikolov, Rebecca N. Wright
2011 Proceedings of the 30th symposium on Principles of database systems of data - PODS '11  
There are well developed notions of private data analyses with dynamic data, for example, using differential privacy.  ...  Our lower bounds are stronger than those implied by differential privacy or dynamic data streaming alone and hold even if unbounded memory and/or unbounded processing time are allowed.  ...  ., sf(p) = median |X0| p also computed off-line numerically.  ... 
doi:10.1145/1989284.1989290 dblp:conf/pods/MirMNW11 fatcat:gmyhlyger5c3plxjlifzipjw2q

Pyramid: Enhancing Selectivity in Big Data Protection with Count Featurization

Mathias Lecuyer, Riley Spahn, Roxana Geambasu, Tzu-Kuo Huang, Siddhartha Sen
2017 2017 IEEE Symposium on Security and Privacy (SP)  
Post-processing resilience: Any computation on a differentially private data release remains differentially private [29] .  ...  Finally, the weight selection process should be made differentially private so the weights computed on a previous hot window do not reveal anything about that window's data at a later time.  ... 
doi:10.1109/sp.2017.60 dblp:conf/sp/LecuyerSGHS17 fatcat:krycieoionfopnfmpzjsaofsru

Near-optimal private approximation protocols via a black box transformation

David P. Woodruff
2011 Proceedings of the 43rd annual ACM symposium on Theory of computing - STOC '11  
In general it is insufficient to use secure function evaluation or fully homomorphic encryption on a standard, non-private protocol for approximating f .  ...  If the original (non-private) protocol is a simultaneous protocol, e.g., a sketching algorithm, then our only cryptographic assumption is efficient symmetric computationally-private information retrieval  ...  in O * (1) space, 1-pass, and O * (n) time (assuming h can be computed in O * (1) time).  ... 
doi:10.1145/1993636.1993733 dblp:conf/stoc/Woodruff11 fatcat:laimykqhozcxjorxnvzcdrl7ga

Privacy-preserving Targeted Advertising [article]

Theja Tulabandhula, Shailesh Vaya, Aritra Dhar
2018 arXiv   pre-print
Even with numerous optimizations and approximations, collaborative filtering (CF) based approaches require real-time computations involving very large vectors.  ...  Modifying such systems to achieve private recommendations further requires communication of long encrypted vectors, making the whole process inefficient.  ...  Choosing L 1 = log N , L 2 = N ρ and L 3 = 3L 2 allows for sublinear query time.  ... 
arXiv:1710.03275v2 fatcat:3ncwkclgrffx5na7ihrkbqblxm

Data breaches in the catastrophe framework & beyond [article]

Spencer Wheatley, Annette Hofmann, Didier Sornette
2019 arXiv   pre-print
'Fortunately' the total breach in that period turned out to be near the median.  ...  Focusing on data breach severity, measured in private information items ('ids') extracted, we exploit relatively mature open data for U.S. data breaches.  ...  Figure 2 . 2 12-year frequency and median severity by type of breach. Figure 3 . 3 Delay between submission and event time for years 2011 and 2012, at the Open Security Foundation datalossdb.  ... 
arXiv:1901.00699v2 fatcat:afx6fdvaffdhznjk254mj4x7nu

Sanitization of Call Detail Records via Differentially-private Summaries [article]

Mohammad Alaggan, Sébastien Gambs, Stan Matwin, Eriko Souza, and Mohammed Tuhin
2014 arXiv   pre-print
Bloom filter for the purpose of privately counting the number of mobile service users moving from one area (region) to another in a given time frame.  ...  Our sanitization method is both time and space efficient, and ensures differential privacy while solving the shortcomings of a solution recently proposed to this problem.  ...  Differentially-private summaries.  ... 
arXiv:1412.8412v2 fatcat:dirhllpmzfhprguy7wsvcwlgbu

The Algorithmic Foundations of Differential Privacy

Cynthia Dwork, Aaron Roth
2013 Foundations and Trends® in Theoretical Computer Science  
Intuitively, this says that if the adversary is restricted to polynomial time then computationally differentially private mechanisms provide the same degree of privacy as do (ε, ν(κ))-differentially private  ...  We note that if we are only concerned about computationally bounded adversaries, then in principle distributed agents can use secure multiparty computation to simulate private algorithms in the centralized  ...  particular intended computation on the given data set.  ... 
doi:10.1561/0400000042 fatcat:jk3mb6ltcfeqzeyvpyznpohuja
« Previous Showing results 1 — 15 out of 92 results