We show that it is possible to signcantly improve the accuracy of a general class of histogram queries while satisfying dierential privacy. Our approach carefully chooses a set
of queries to evaluate, and then exploits consistency constraints that should hold over the noisy output. In a post-processing phase, we compute the consistent input most likely to have produced the noisy output. The nal output is dierentially-private and consistent, but in addition,
it is often much more accurate. We show, both theoretically and experimentally, that these techniques can be used for estimating the degree sequence of a graph very precisely, and for computing a histogram that can support arbitrary range queries accurately
%0 Journal Article
%1 DBLP:journals/pvldb/HayRMS10
%A Hay, Michael
%A Rastogi, Vibhor
%A Miklau, Gerome
%A Suciu, Dan
%D 2010
%J PVLDB
%K database differential_privacy privacy
%N 1
%P 1021-1032
%T Boosting the Accuracy of Differentially Private Histograms
Through Consistency
%U http://www.comp.nus.edu.sg/~vldb2010/proceedings/files/papers/R91.pdf
%V 3
%X We show that it is possible to signcantly improve the accuracy of a general class of histogram queries while satisfying dierential privacy. Our approach carefully chooses a set
of queries to evaluate, and then exploits consistency constraints that should hold over the noisy output. In a post-processing phase, we compute the consistent input most likely to have produced the noisy output. The nal output is dierentially-private and consistent, but in addition,
it is often much more accurate. We show, both theoretically and experimentally, that these techniques can be used for estimating the degree sequence of a graph very precisely, and for computing a histogram that can support arbitrary range queries accurately
@article{DBLP:journals/pvldb/HayRMS10,
abstract = {
We show that it is possible to signcantly improve the accuracy of a general class of histogram queries while satisfying dierential privacy. Our approach carefully chooses a set
of queries to evaluate, and then exploits consistency constraints that should hold over the noisy output. In a post-processing phase, we compute the consistent input most likely to have produced the noisy output. The nal output is dierentially-private and consistent, but in addition,
it is often much more accurate. We show, both theoretically and experimentally, that these techniques can be used for estimating the degree sequence of a graph very precisely, and for computing a histogram that can support arbitrary range queries accurately},
added-at = {2011-03-25T22:02:40.000+0100},
author = {Hay, Michael and Rastogi, Vibhor and Miklau, Gerome and Suciu, Dan},
bibsource = {DBLP, http://dblp.uni-trier.de},
biburl = {https://www.bibsonomy.org/bibtex/256ab90cf84d4eefafc6e67a3535e2dcb/ytyoun},
interhash = {74afb3bf5086a45c4839947ccce3cbcd},
intrahash = {56ab90cf84d4eefafc6e67a3535e2dcb},
journal = {PVLDB},
keywords = {database differential_privacy privacy},
number = 1,
pages = {1021-1032},
timestamp = {2017-12-08T03:45:05.000+0100},
title = {Boosting the Accuracy of Differentially Private Histograms
Through Consistency},
url = {http://www.comp.nus.edu.sg/~vldb2010/proceedings/files/papers/R91.pdf},
volume = 3,
year = 2010
}