Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Privately Evaluating Decision Trees and Random Forests

Privately Evaluating Decision Trees and Random Forests Abstract Decision trees and random forests are common classifiers with widespread use. In this paper, we develop two protocols for privately evaluating decision trees and random forests. We operate in the standard two-party setting where the server holds a model (either a tree or a forest), and the client holds an input (a feature vector). At the conclusion of the protocol, the client learns only the model’s output on its input and a few generic parameters concerning the model; the server learns nothing. The first protocol we develop provides security against semi-honest adversaries. We then give an extension of the semi-honest protocol that is robust against malicious adversaries. We implement both protocols and show that both variants are able to process trees with several hundred decision nodes in just a few seconds and a modest amount of bandwidth. Compared to previous semi-honest protocols for private decision tree evaluation, we demonstrate a tenfold improvement in computation and bandwidth. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Proceedings on Privacy Enhancing Technologies de Gruyter

Privately Evaluating Decision Trees and Random Forests

Download PDF
21 pages

Loading next page...
 
/lp/de-gruyter/privately-evaluating-decision-trees-and-random-forests-lc3vze7MtF
Publisher
de Gruyter
Copyright
Copyright © 2016 by the
ISSN
2299-0984
eISSN
2299-0984
DOI
10.1515/popets-2016-0043
Publisher site
See Article on Publisher Site

Abstract

Abstract Decision trees and random forests are common classifiers with widespread use. In this paper, we develop two protocols for privately evaluating decision trees and random forests. We operate in the standard two-party setting where the server holds a model (either a tree or a forest), and the client holds an input (a feature vector). At the conclusion of the protocol, the client learns only the model’s output on its input and a few generic parameters concerning the model; the server learns nothing. The first protocol we develop provides security against semi-honest adversaries. We then give an extension of the semi-honest protocol that is robust against malicious adversaries. We implement both protocols and show that both variants are able to process trees with several hundred decision nodes in just a few seconds and a modest amount of bandwidth. Compared to previous semi-honest protocols for private decision tree evaluation, we demonstrate a tenfold improvement in computation and bandwidth.

Journal

Proceedings on Privacy Enhancing Technologies – de Gruyter

Published: Oct 1, 2016

References

  • Decision tree classifiers for automated medical diagnosis Neural Applications
    Azar
  • generalization of Paillier s public - key system with applications to electronic voting Sec
    Damgård
  • Specification of curve selection and supported curve parameters in Technical Report Microsoft Research
    Bos
  • How to exchange secrets with oblivious transfer Cryptology ePrint Archive
    Rabin
  • Wallet databases with observers In pages
    Chaum
  • Privacy preserving data mining In pages
    Lindell
  • composition of cryptographic protocols a tutorial part News
    Canetti
  • How to exchange secrets with oblivious transfer Cryptology ePrint Archive
    Rabin
  • Strong conditional oblivious transfer and computing on intervals In
    Blake
  • in In pages
    Halevi
  • Improved garbled circuit Free gates applications In pages
    Kolesnikov
  • proof of security of protocol for two - party computation
    Lindell
  • composition of multiparty cryptographic protocols
    Canetti
  • Enhancing privacy and trust in electronic communities In EC pages
    Huberman
  • Goldreich Foundations of Volume Basic Applications Cambridge University New York
    Cryptography
  • machine learning repository
    Bache
  • Goldreich Foundations of Volume Basic Applications Cambridge University New York
    Cryptography
  • Strong conditional oblivious transfer and computing on intervals In
    Blake
  • Goldwasser learning classification over encrypted data In
    Bost
  • framework for fast privacy - preserving computations In pages
    Bogdanov
  • On achieving the best of both worlds in secure multiparty computation
    Ishai
  • How to generate and exchange secrets extended abstract In pages
    Yao
  • two - step method to construct credit scoring models with data mining techniques of
    Koh
  • Oblivious decision programs from oblivious transfer Efficient reductions
    Mohassel
  • The decision problem In ANTS pages
    Boneh
  • Proofs of partial knowledge and simplified design of witness hiding protocols In pages
    Cramer
  • Lauter Privately evaluating decision trees and random forests IACR Cryptology ePrint Archive
    Wu
  • How to hide circuits in An efficient framework for private function evaluation In pages
    Mohassel
  • Efficient and secure comparison for on - line auctions In pages
    Damgård
  • Location privacy via private proximity testing In
    Narayanan
  • Efficient oblivious transfer protocols In SODA pages
    Naor
  • Improved garbled circuit building blocks and applications to auctions and computing minima Cryptology ePrint Archive Report
    Kolesnikov
  • Founding cryptography on oblivious transfer In pages
    Kilian
  • Privacy - preserving data mining
    Agrawal
  • framework for efficient mixed - protocol secure two - party computation In
    Demmler
  • On achieving the best of both worlds in secure multiparty computation
    Ishai
  • The Elements of Statistical Learning Springer Series in New York
    Hastie
  • Goldwasser Probabilistic encryption of and System
    Micali
  • How to generate and exchange secrets extended abstract In pages
    Yao
  • framework for fast privacy - preserving computations In pages
    Bogdanov
  • Efficient Secure Two Party Protocols Techniques and Constructions
    Hazay
  • Decision tree classifiers for automated medical diagnosis Neural Applications
    Azar
  • Two halves make a whole - reducing data transfer in garbled circuits using half gates In pages
    Zahur
  • Efficient garbling from a fixed - key blockcipher In on Privacy pages
    Bellare
  • Public - key cryptosystems based on composite degree residuosity classes In volume pages
    Paillier
  • Leveled fully homomorphic encryption without bootstrapping In pages
    Brakerski
  • Goldwasser Probabilistic encryption of and System
    Micali
  • secure and optimally efficient multi - authority election scheme In pages
    Cramer
  • Oblivious transfer polynomial evaluation In pages
    Naor
  • proof of security of protocol for two - party computation
    Lindell
  • library for doing number theory http www shoup net ntl
    Shoup
  • https bigml com
    BigML
  • Efficient group signature schemes for large groups extended abstract In pages
    Camenisch
  • Efficient Secure Two Party Protocols Techniques and Constructions
    Hazay
  • Goldwasser The knowledge complexity of interactive proof - systems extended abstract In pages
    Micali
  • Location privacy via private proximity testing In
    Narayanan
  • Enhancing privacy and trust in electronic communities In EC pages
    Huberman
  • Wallet databases with observers In pages
    Chaum
  • Du Building decision tree classifier on private data In
    Zhan
  • Privacy preserving remote diagnostics In pages
    Brickell
  • Goldwasser learning classification over encrypted data In
    Bost
  • Conditional oblivious transfer and timed - release encryption In pages
    Crescenzo
  • Efficient and secure comparison for on - line auctions In pages
    Damgård
  • Model inversion attacks that exploit confidence information and basic countermeasures In pages
    Fredrikson
  • Secure evaluation of private linear branching programs with medical applications In pages
    Barni
  • Random forests
    Breiman
  • Efficient garbling from a fixed - key blockcipher In on Privacy pages
    Bellare
  • generalization of Paillier s public - key system with applications to electronic voting Sec
    Damgård
  • More efficient oblivious transfer and extensions for faster secure computation In pages
    Asharov
  • composition of multiparty cryptographic protocols
    Canetti
  • Random forests
    Breiman
  • Selecting elliptic curves for cryptography : An efficiency and security analysis Cryptology ePrint Archive
    Bos
  • How to hide circuits in An efficient framework for private function evaluation In pages
    Mohassel
  • Leveled fully homomorphic encryption without bootstrapping In pages
    Brakerski
  • Private predictive analysis on encrypted medical data of
    Bos
  • confidential learning on encrypted data In pages
    Graepel
  • Private predictive analysis on encrypted medical data of
    Bos
  • Microsoft Learning https azure microsoft com en us services machine learning
    Azure
  • Specification of curve selection and supported curve parameters in Technical Report Microsoft Research
    Bos
  • Improved garbled circuit Free gates applications In pages
    Kolesnikov
  • Founding cryptography on oblivious transfer In pages
    Kilian
  • Proofs of partial knowledge and simplified design of witness hiding protocols In pages
    Cramer
  • two - step method to construct credit scoring models with data mining techniques of
    Koh
  • fully homomorphic encryption scheme thesis University
    Gentry
  • framework for efficient mixed - protocol secure two - party computation In
    Demmler
  • Privacy preserving data mining In pages
    Lindell
  • Privacy preserving face recognition In PETS pages
    Erkin
  • Privacy - preserving data mining
    Agrawal
  • Conditional oblivious transfer and timed - release encryption In pages
    Crescenzo
  • More efficient oblivious transfer and extensions for faster secure computation In pages
    Asharov
  • confidential learning on encrypted data In pages
    Graepel
  • pseudorandom generator from any one - way function
    Håstad
  • Improved garbled circuit building blocks and applications to auctions and computing minima Cryptology ePrint Archive Report
    Kolesnikov
  • Secure evaluation of private linear branching programs with medical applications In pages
    Barni
  • fully homomorphic encryption scheme thesis University
    Gentry
  • composition of cryptographic protocols a tutorial part News
    Canetti
  • Lauter Privately evaluating decision trees and random forests IACR Cryptology ePrint Archive
    Wu
  • library for doing number theory http www shoup net ntl
    Shoup
  • comparison of non - symmetric entropy - based classification trees and support vector machine for cardiovascular risk stratification Annual International Conference of the IEEE Engineering in Medicine and Biology pages
    Singh
  • Microsoft Learning https azure microsoft com en us services machine learning
    Azure
  • Fiat How to prove yourself : Practical solutions to identification and signature problems In pages
    Shamir
  • in In pages
    Halevi
  • Oblivious transfer polynomial evaluation In pages
    Naor
  • pseudorandom generator from any one - way function
    Håstad
  • Constant - round private function evaluation with linear complexity In pages
    Katz
  • Two halves make a whole - reducing data transfer in garbled circuits using half gates In pages
    Zahur
  • Oblivious decision programs from oblivious transfer Efficient reductions
    Mohassel
  • Privacy preserving face recognition In PETS pages
    Erkin
  • comparison of non - symmetric entropy - based classification trees and support vector machine for cardiovascular risk stratification Annual International Conference of the IEEE Engineering in Medicine and Biology pages
    Singh
  • The Elements of Statistical Learning Springer Series in New York
    Hastie
  • Goldwasser The knowledge complexity of interactive proof - systems extended abstract In pages
    Micali
  • Du Building decision tree classifier on private data In
    Zhan
  • Efficient group signature schemes for large groups extended abstract In pages
    Camenisch
  • machine learning repository
    Bache
  • The decision problem In ANTS pages
    Boneh
  • https bigml com
    BigML
  • Model inversion attacks that exploit confidence information and basic countermeasures In pages
    Fredrikson
  • Constant - round private function evaluation with linear complexity In pages
    Katz
  • Efficient oblivious transfer protocols In SODA pages
    Naor
  • Selecting elliptic curves for cryptography : An efficiency and security analysis Cryptology ePrint Archive
    Bos
  • secure and optimally efficient multi - authority election scheme In pages
    Cramer
  • Fiat How to prove yourself : Practical solutions to identification and signature problems In pages
    Shamir
  • Public - key cryptosystems based on composite degree residuosity classes In volume pages
    Paillier
  • Privacy preserving remote diagnostics In pages
    Brickell