eprint.iacr.org will be offline for approximately an hour for routine maintenance again at 10pm UTC on Wednesday, April 17.

Paper 2014/331

Machine Learning Classification over Encrypted Data

Raphael Bost, Raluca Ada Popa, Stephen Tu, and Shafi Goldwasser

Abstract

Machine learning classification is used in numerous settings nowadays, such as medical or genomics predictions, spam detection, face recognition, and financial predictions. Due to privacy concerns in some of these applications, it is important that the data and the classifier remain confidential. In this work, we construct three major classification protocols that satisfy this privacy constraint: hyperplane decision, Naïve Bayes, and decision trees. These protocols may also be combined with AdaBoost. They rely on a library of building blocks for constructing classifiers securely, and we demonstrate the versatility of this library by constructing a face detection classifier. Our protocols are efficient, taking milliseconds to a few seconds to perform a classification when running on real medical datasets.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
Published elsewhere. Major revision. NDSS 2015
DOI
10.14722/ndss.2015.23241
Keywords
public-key cryptographyimplementationapplicationsmachine learning
Contact author(s)
raphael_bost @ alumni brown edu
History
2015-01-12: last of 3 revisions
2014-05-13: received
See all versions
Short URL
https://ia.cr/2014/331
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2014/331,
      author = {Raphael Bost and Raluca Ada Popa and Stephen Tu and Shafi Goldwasser},
      title = {Machine Learning Classification over Encrypted Data},
      howpublished = {Cryptology ePrint Archive, Paper 2014/331},
      year = {2014},
      doi = {10.14722/ndss.2015.23241},
      note = {\url{https://eprint.iacr.org/2014/331}},
      url = {https://eprint.iacr.org/2014/331}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.