Skip to main navigation Skip to search Skip to main content

Secure computation of hidden markov models

  • University of Notre Dame

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

14 Scopus citations

Abstract

Hidden Markov Model (HMM) is a popular statistical tool with a large number of applications in pattern recognition. In some of such applications, including speaker recognition in particular, the computation involves personal data that can identify individuals and must be protected. For that reason, we develop privacypreserving techniques for HMM and Gaussian mixture model (GMM) computation suitable for use in speaker recognition and other applications. Unlike prior work, our solution uses floating point arithmetic, which allows us to simultaneously achieve high accuracy, provable security guarantees, and reasonable performance. We develop techniques for both two-party HMM and GMM computation based on threshold homomorphic encryption and multi-party computation based on threshold linear secret sharing, which are suitable for secure collaborative computation as well as secure outsourcing.

Original languageEnglish
Title of host publicationICETE 2013 - 10th International Joint Conference on E-Business and Telecommunications; SECRYPT 2013 - 10th International Conference on Security and Cryptography, Proceedings
Pages242-253
Number of pages12
StatePublished - 2013
Event10th International Conference on Security and Cryptography, SECRYPT 2013 - Part of 10th International Joint Conference on E-Business and Telecommunications, ICETE 2013 - Reykjavik, Iceland
Duration: Jul 29 2013Jul 31 2013

Publication series

NameICETE 2013 - 10th International Joint Conference on E-Business and Telecommunications; SECRYPT 2013 - 10th International Conference on Security and Cryptography, Proceedings

Conference

Conference10th International Conference on Security and Cryptography, SECRYPT 2013 - Part of 10th International Joint Conference on E-Business and Telecommunications, ICETE 2013
Country/TerritoryIceland
CityReykjavik
Period07/29/1307/31/13

Keywords

  • Floating point
  • Gaussian mixture models
  • Hidden markov models
  • Secure computation

Fingerprint

Dive into the research topics of 'Secure computation of hidden markov models'. Together they form a unique fingerprint.

Cite this