Filters








6,397 Hits in 3.5 sec

Privacy-Preserving Distributed Linear Regression on High-Dimensional Data

Adrià Gascón, Phillipp Schoppmann, Borja Balle, Mariana Raykova, Jack Doerner, Samee Zahur, David Evans
2017 Proceedings on Privacy Enhancing Technologies  
We propose privacy-preserving protocols for computing linear regression models, in the setting where the training dataset is vertically distributed among several parties.  ...  Our technique improves on Nikolaenko et al.'s method for privacy-preserving ridge regression (S&P 2013), and can be used as a building block in other analyses.  ...  Distributed Linear Regression on High-Dimensional Data Unauthenticated Download Date | 7/26/18 9:13 AM Privacy-Preserving Distributed Linear Regression on High-Dimensional Data  ... 
doi:10.1515/popets-2017-0053 dblp:journals/popets/GasconSB0DZE17 fatcat:hpn4a3ulf5dstojfrvjesrjf6y

Difficulty in estimating visual information from randomly sampled images [article]

Masaki Kitayama, Hitoshi Kiya
2020 arXiv   pre-print
In particular, the random sampling method that was proposed for privacy-preserving machine learning, is compared with typical dimensionality reduction methods.  ...  In this paper, we evaluate dimensionality reduction methods in terms of difficulty in estimating visual information on original images from dimensionally reduced ones.  ...  In machine learning, P is used for reducing the number of random variables for avoiding negative effects of high-dimensional data.  ... 
arXiv:2012.08751v1 fatcat:4egnhp2uezautk35wpemasiclm

Efficient differentially private learning improves drug sensitivity prediction

Antti Honkela, Mrinal Das, Arttu Nieminen, Onur Dikmen, Samuel Kaski
2018 Biology Direct  
However, differentially private learning with current methods does not improve predictions with feasible data sizes and dimensionalities.  ...  private regression method in the accuracy of private drug sensitivity prediction.  ...  Availability of data and materials Datasets used in these studies have been derived from public resources and have been referred to in the paper.  ... 
doi:10.1186/s13062-017-0203-4 pmid:29409513 pmcid:PMC5801888 fatcat:b6nvoxpjl5chdci55ug3sc3dt4

Privacy-preserving cox regression for survival analysis

Shipeng Yu, Glenn Fung, Romer Rosales, Sriram Krishnan, R. Bharat Rao, Cary Dehing-Oberije, Philippe Lambin
2008 Proceeding of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining - KDD 08  
The proposed model is based on linearly projecting the data to a lower dimensional space through an optimal mapping obtained by solving a linear programming problem.  ...  Privacy-preserving data mining (PPDM) is an emergent research area that addresses the incorporation of privacy preserving concerns to data mining techniques.  ...  DISCUSSION AND CONCLUSIONS We have designed an approach for privacy-preserving data mining based on a linear (lossy) projection of the original data onto a lower-dimensional space.  ... 
doi:10.1145/1401890.1402013 dblp:conf/kdd/YuFRKRDL08 fatcat:vpie2vp6szcxrgbuzlheczhpiy

LoPub: High-Dimensional Crowdsourced Data Publication with Local Differential Privacy [article]

Xuebin Ren, Chia-Mu Yu, Weiren Yu, Shusen Yang, Xinyu Yang, Julie A. McCann, Philip S. Yu
2017 arXiv   pre-print
Then, we develop a Locally privacy-preserving high-dimensional data Publication algorithm, LoPub, by taking advantage of our distribution estimation techniques.  ...  Here, based on EM and Lasso regression, we propose efficient multi-dimensional joint distribution estimation algorithms with local privacy.  ...  To this end, we present LoPub, a locally privacy-preserving data publication scheme for high-dimensional crowdsourced data.  ... 
arXiv:1612.04350v2 fatcat:rymvsunigzbr3mhsjh26qhl6ve

High-Dimensional Crowdsourced Data Distribution Estimation with Local Privacy

Xuebin Ren, Chia-Mu Yu, Weiren Yu, Shusen Yang, Xinyu Yang, Julie McCann
2016 2016 IEEE International Conference on Computer and Information Technology (CIT)  
Then, we develop a Locally privacy-preserving high-dimensional data Publication algorithm, LoPub, by taking advantage of our distribution estimation techniques.  ...  To this end, based on Expectation Maximization (EM) algorithm and Lasso regression, we first propose efficient multi-dimensional joint distribution estimation algorithms that maintain local privacy.  ...  The server gathers all the data and estimates high-dimensional crowdsourced data distribution with local privacy, aiming to release a privacy-preserving dataset to third-parties for conducting data analysis  ... 
doi:10.1109/cit.2016.57 dblp:conf/IEEEcit/RenYYYYM16 fatcat:zb3cijvymja2pa7ioqynyja7w4

A Knowledge Transfer Framework for Differentially Private Sparse Learning

Lingxiao Wang, Quanquan Gu
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We study the problem of estimating high dimensional models with underlying sparse structures while preserving the privacy of each training example.  ...  We develop a differentially private high-dimensional sparse learning framework using the idea of knowledge transfer.  ...  Sparse linear regression We consider the following linear regression problem in the high-dimensional regime (Tibshirani 1996) : y = Xθ * + ξ, where y ∈ R n is the response vector, X ∈ R n×d denotes the  ... 
doi:10.1609/aaai.v34i04.6090 fatcat:tyinzutwgjeolhy4eg6eacqm3e

Privacy preserving data visualizations

Demetris Avraam, Rebecca Wilson, Oliver Butters, Thomas Burton, Christos Nicolaides, Elinor Jones, Andy Boyd, Paul Burton
2021 EPJ Data Science  
We apply the proposed methods to generate privacy-preserving data visualizations for exploratory data analysis and inferential regression plot diagnostics, and we discuss their strengths and limitations  ...  we propose the use of anonymization techniques to generate privacy-preserving visualizations that retain the statistical properties of the underlying data while still adhering to strict data disclosure  ...  data in the Population Health Sciences  ... 
doi:10.1140/epjds/s13688-020-00257-4 pmid:33442528 pmcid:PMC7790778 fatcat:lmbanwnvyvcgxe3me2t2ays4ty

Compressed and Privacy-Sensitive Sparse Regression

Shuheng Zhou, John Lafferty, Larry Wasserman
2009 IEEE Transactions on Information Theory  
A primary motivation for this compression procedure is to anonymize the data and preserve privacy by revealing little information about the original data.  ...  Recent research has studied the role of sparsity in high-dimensional regression and signal reconstruction, establishing theoretical limits for recovering sparse models.  ...  ACKNOWLEDGMENT The authors wish to thank Avrim Blum, Steve Fienberg, and Pradeep Ravikumar for helpful comments on this work.  ... 
doi:10.1109/tit.2008.2009605 fatcat:qqb4j5rtpfaltbxuhel3puy7au

Collect at Once, Use Effectively: Making Non-interactive Locally Private Learning Possible [article]

Kai Zheng, Wenlong Mou, Liwei Wang
2017 arXiv   pre-print
For the high-dimensional world, we discover that under ℓ_2-norm assumption on data points, high-dimensional sparse linear regression and mean estimation can be achieved with logarithmic dependence on dimension  ...  Non-interactive Local Differential Privacy (LDP) requires data analysts to collect data from users through noisy channel at once.  ...  In the following we summarize our contributions. (1) High Dimensional Estimation: One of exciting findings in this paper is about local privacy for high-dimensional data.  ... 
arXiv:1706.03316v1 fatcat:d5gi3b6wczcutcu5xnticoqlly

Cloud-enabled privacy-preserving collaborative learning for mobile sensing

Bin Liu, Yurong Jiang, Fei Sha, Ramesh Govindan
2012 Proceedings of the 10th ACM Conference on Embedded Network Sensor Systems - SenSys '12  
In this paper, we consider the design of a system in which Internet-connected mobile users contribute sensor data as training samples, and collaborate on building a model for classification tasks such  ...  Constructing the model can naturally be performed by a service running in the cloud, but users may be more inclined to contribute training samples if the privacy of these data could be ensured.  ...  It leverages a non-linear mapping to map x x x into a very high-dimensional feature space.  ... 
doi:10.1145/2426656.2426663 dblp:conf/sensys/0004JSG12 fatcat:gcp6iizt3ng7jbxohnkibxa4lq

Privacy-Preserving Data Sharing in High Dimensional Regression and Classification Settings

Stephen E. Fienberg, Jiashun Jin
2012 Journal of Privacy and Confidentiality  
We characterize the notions of "cautious", "regular", and "generous" data sharing in terms of their privacy-preserving implications for the parties and their share of data, with focus on the "feature privacy  ...  We focus on the problem of multi-party data sharing in high dimensional data settings where the number of measured features (or the dimension) p is frequently much larger than the number of subjects (or  ...  section, we relate the approach of privacy-preserving noise addition to high dimensional regression analysis.  ... 
doi:10.29012/jpc.v4i1.618 fatcat:pfvkqnhuhjbudmfkilxwr6kf4q

A Knowledge Transfer Framework for Differentially Private Sparse Learning [article]

Lingxiao Wang, Quanquan Gu
2019 arXiv   pre-print
We study the problem of estimating high dimensional models with underlying sparse structures while preserving the privacy of each training example.  ...  We develop a differentially private high-dimensional sparse learning framework using the idea of knowledge transfer.  ...  Sparse linear regression We consider the following linear regression problem in the high-dimensional regime (Tibshirani, 1996) : y = Xθ * + ξ, where y ∈ R n is the response vector, X ∈ R n×d denotes the  ... 
arXiv:1909.06322v1 fatcat:gbrytwg3qzh3bawuwygltmmbnm

AutoGAN-based Dimension Reduction for Privacy Preservation [article]

Hung Nguyen, Di Zhuang, Pei-Yuan Wu, Morris Chang
2019 arXiv   pre-print
Furthermore, such techniques usually result in low performance with a high number of queries. To address these problems, we propose a dimension reduction-based method for privacy preservation.  ...  In the experiments, we test our method on popular face image datasets and show that our method can retain data utility and resist data reconstruction, thus protecting privacy.  ...  of logistic regression or linear regression models  ... 
arXiv:1902.10799v1 fatcat:546hittb5veofipn5cotmkaeee

Differentially Private Confidence Intervals for Empirical Risk Minimization [article]

Yue Wang, Daniel Kifer, Jaewoo Lee
2018 arXiv   pre-print
The process of data mining with differential privacy produces results that are affected by two types of noise: sampling noise due to data collection and privacy noise that is designed to prevent the reconstruction  ...  The algorithms can provide confidence intervals that satisfy differential privacy (as well as the more recently proposed concentrated differential privacy) and can be used with existing differentially  ...  High dimensional regression problems were also studied in [5, 24] . Bassily et al.  ... 
arXiv:1804.03794v1 fatcat:2tgwyj6lpzbl5f54lft5o7rr4e
« Previous Showing results 1 — 15 out of 6,397 results