Fast Homomorphic Evaluation of Deep Discretized Neural Networks View Full Text


Ontology type: schema:Chapter     


Chapter Info

DATE

2018-07-24

AUTHORS

Florian Bourse , Michele Minelli , Matthias Minihold , Pascal Paillier

ABSTRACT

The rise of machine learning as a service multiplies scenarios where one faces a privacy dilemma: either sensitive user data must be revealed to the entity that evaluates the cognitive model (e.g., in the Cloud), or the model itself must be revealed to the user so that the evaluation can take place locally. Fully Homomorphic Encryption (FHE) offers an elegant way to reconcile these conflicting interests in the Cloud-based scenario and also preserve non-interactivity. However, due to the inefficiency of existing FHE schemes, most applications prefer to use Somewhat Homomorphic Encryption (SHE), where the complexity of the computation to be performed has to be known in advance, and the efficiency of the scheme depends on this global complexity. In this paper, we present a new framework for homomorphic evaluation of neural networks, that we call FHE–DiNN, whose complexity is strictly linear in the depth of the network and whose parameters can be set beforehand. To obtain this scale-invariance property, we rely heavily on the bootstrapping procedure. We refine the recent FHE construction by Chillotti et al. (ASIACRYPT 2016) in order to increase the message space and apply the sign function (that we use to activate the neurons in the network) during the bootstrapping. We derive some empirical results, using TFHE library as a starting point, and classify encrypted images from the MNIST dataset with more than 96% accuracy in less than 1.7 s. Finally, as a side contribution, we analyze and introduce some variations to the bootstrapping technique of Chillotti et al. that offer an improvement in efficiency at the cost of increasing the storage requirements. More... »

PAGES

483-512

References to SciGraph publications

  • 2016. Fully Secure Functional Encryption for Inner Products, from Standard Assumptions in ADVANCES IN CRYPTOLOGY – CRYPTO 2016
  • 2014. Algorithms in HElib in ADVANCES IN CRYPTOLOGY – CRYPTO 2014
  • 2010. Fully Homomorphic Encryption with Relatively Small Key and Ciphertext Sizes in PUBLIC KEY CRYPTOGRAPHY – PKC 2010
  • 2015-04-14. Bootstrapping for HElib in ADVANCES IN CRYPTOLOGY -- EUROCRYPT 2015
  • 2015-04-14. FHEW: Bootstrapping Homomorphic Encryption in Less Than a Second in ADVANCES IN CRYPTOLOGY -- EUROCRYPT 2015
  • 2012. Homomorphic Evaluation of the AES Circuit in ADVANCES IN CRYPTOLOGY – CRYPTO 2012
  • 2006. Differential Privacy in AUTOMATA, LANGUAGES AND PROGRAMMING
  • 2013. Practical Bootstrapping in Quasilinear Time in ADVANCES IN CRYPTOLOGY – CRYPTO 2013
  • 2010. On Ideal Lattices and Learning with Errors over Rings in ADVANCES IN CRYPTOLOGY – EUROCRYPT 2010
  • 2010. Fully Homomorphic Encryption over the Integers in ADVANCES IN CRYPTOLOGY – EUROCRYPT 2010
  • 2013. Homomorphic Encryption from Learning with Errors: Conceptually-Simpler, Asymptotically-Faster, Attribute-Based in ADVANCES IN CRYPTOLOGY – CRYPTO 2013
  • 2015-03-17. Simple Functional Encryption Schemes for Inner Products in PUBLIC-KEY CRYPTOGRAPHY -- PKC 2015
  • 2014. Faster Bootstrapping with Polynomial Error in ADVANCES IN CRYPTOLOGY – CRYPTO 2014
  • 1989-12. Approximation by superpositions of a sigmoidal function in MATHEMATICS OF CONTROL, SIGNALS, AND SYSTEMS
  • 2011. Fully Homomorphic Encryption from Ring-LWE and Security for Key Dependent Messages in ADVANCES IN CRYPTOLOGY – CRYPTO 2011
  • 2008. Concurrently Secure Identification Schemes Based on the Worst-Case Hardness of Lattice Problems in ADVANCES IN CRYPTOLOGY - ASIACRYPT 2008
  • 2010. Faster Fully Homomorphic Encryption in ADVANCES IN CRYPTOLOGY - ASIACRYPT 2010
  • 2016. Faster Fully Homomorphic Encryption: Bootstrapping in Less Than 0.1 Seconds in ADVANCES IN CRYPTOLOGY – ASIACRYPT 2016
  • 2016. Minimizing the Number of Bootstrappings in Fully Homomorphic Encryption in SELECTED AREAS IN CRYPTOGRAPHY – SAC 2015
  • 2017. Faster Packed Homomorphic Operations and Efficient Circuit Bootstrapping for TFHE in ADVANCES IN CRYPTOLOGY – ASIACRYPT 2017
  • 2013. On the Minimal Number of Bootstrappings in Homomorphic Circuits in FINANCIAL CRYPTOGRAPHY AND DATA SECURITY
  • Identifiers

    URI

    http://scigraph.springernature.com/pub.10.1007/978-3-319-96878-0_17

    DOI

    http://dx.doi.org/10.1007/978-3-319-96878-0_17

    DIMENSIONS

    https://app.dimensions.ai/details/publication/pub.1105780951


    Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
    Incoming Citations Browse incoming citations for this publication using opencitations.net

    JSON-LD is the canonical representation for SciGraph data.

    TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

    [
      {
        "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
        "about": [
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Artificial Intelligence and Image Processing", 
            "type": "DefinedTerm"
          }, 
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Information and Computing Sciences", 
            "type": "DefinedTerm"
          }
        ], 
        "author": [
          {
            "affiliation": {
              "alternateName": "Orange (France)", 
              "id": "https://www.grid.ac/institutes/grid.89485.38", 
              "name": [
                "Orange Labs, Applied Crypto Group, Cesson-S\u00e9vign\u00e9, France"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Bourse", 
            "givenName": "Florian", 
            "id": "sg:person.013454106661.29", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013454106661.29"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "French Institute for Research in Computer Science and Automation", 
              "id": "https://www.grid.ac/institutes/grid.5328.c", 
              "name": [
                "DIENS, \u00c9cole normale sup\u00e9rieure, CNRS, PSL Research University, Paris, France", 
                "Inria, Paris, France"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Minelli", 
            "givenName": "Michele", 
            "id": "sg:person.015114730252.70", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.015114730252.70"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Ruhr University Bochum", 
              "id": "https://www.grid.ac/institutes/grid.5570.7", 
              "name": [
                "Horst G\u00f6rtz Institut f\u00fcr IT-Security, Ruhr-Universit\u00e4t Bochum, Bochum, Germany"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Minihold", 
            "givenName": "Matthias", 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "CryptoExperts (France)", 
              "id": "https://www.grid.ac/institutes/grid.470554.7", 
              "name": [
                "CryptoExperts, Paris, France"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Paillier", 
            "givenName": "Pascal", 
            "id": "sg:person.012202553435.44", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012202553435.44"
            ], 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "sg:pub.10.1007/978-3-662-44371-2_31", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1003203986", 
              "https://doi.org/10.1007/978-3-662-44371-2_31"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-46447-2_33", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1006063386", 
              "https://doi.org/10.1007/978-3-662-46447-2_33"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-46447-2_33", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1006063386", 
              "https://doi.org/10.1007/978-3-662-46447-2_33"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-40041-4_5", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1006864000", 
              "https://doi.org/10.1007/978-3-642-40041-4_5"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-32009-5_49", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1011363442", 
              "https://doi.org/10.1007/978-3-642-32009-5_49"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1060590.1060603", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1012332159"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-540-89255-7_23", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1021516251", 
              "https://doi.org/10.1007/978-3-540-89255-7_23"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-540-89255-7_23", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1021516251", 
              "https://doi.org/10.1007/978-3-540-89255-7_23"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-17373-8_22", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1021970215", 
              "https://doi.org/10.1007/978-3-642-17373-8_22"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-17373-8_22", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1021970215", 
              "https://doi.org/10.1007/978-3-642-17373-8_22"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/bf02551274", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1023250347", 
              "https://doi.org/10.1007/bf02551274"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/bf02551274", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1023250347", 
              "https://doi.org/10.1007/bf02551274"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-13190-5_2", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1023605606", 
              "https://doi.org/10.1007/978-3-642-13190-5_2"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-13190-5_2", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1023605606", 
              "https://doi.org/10.1007/978-3-642-13190-5_2"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-13190-5_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1025305533", 
              "https://doi.org/10.1007/978-3-642-13190-5_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-13190-5_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1025305533", 
              "https://doi.org/10.1007/978-3-642-13190-5_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-41320-9_13", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1028621484", 
              "https://doi.org/10.1007/978-3-642-41320-9_13"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/11787006_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1035160724", 
              "https://doi.org/10.1007/11787006_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/11787006_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1035160724", 
              "https://doi.org/10.1007/11787006_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-319-31301-6_2", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1036619867", 
              "https://doi.org/10.1007/978-3-319-31301-6_2"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-46800-5_24", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1037378777", 
              "https://doi.org/10.1007/978-3-662-46800-5_24"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-46800-5_24", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1037378777", 
              "https://doi.org/10.1007/978-3-662-46800-5_24"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-13013-7_25", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1039389499", 
              "https://doi.org/10.1007/978-3-642-13013-7_25"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-46800-5_25", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1042169019", 
              "https://doi.org/10.1007/978-3-662-46800-5_25"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-46800-5_25", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1042169019", 
              "https://doi.org/10.1007/978-3-662-46800-5_25"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2554797.2554799", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1042352254"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-44371-2_17", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1047019051", 
              "https://doi.org/10.1007/978-3-662-44371-2_17"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-22792-9_29", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1049197805", 
              "https://doi.org/10.1007/978-3-642-22792-9_29"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-22792-9_29", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1049197805", 
              "https://doi.org/10.1007/978-3-642-22792-9_29"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-40041-4_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1049650086", 
              "https://doi.org/10.1007/978-3-642-40041-4_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/0893-6080(91)90009-t", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1050371510"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/0893-6080(91)90009-t", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1050371510"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-53015-3_12", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1051267410", 
              "https://doi.org/10.1007/978-3-662-53015-3_12"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/5.726791", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061179979"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/tc.2015.2470255", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061536082"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/335191.335438", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1063168485"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-662-53887-6_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1084896599", 
              "https://doi.org/10.1007/978-3-662-53887-6_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-319-70694-8_14", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093076441", 
              "https://doi.org/10.1007/978-3-319-70694-8_14"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/sp.2017.12", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093587091"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14722/ndss.2015.23241", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1095873103"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1137/1.9781611974782.160", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1098556665"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "2018-07-24", 
        "datePublishedReg": "2018-07-24", 
        "description": "The rise of machine learning as a service multiplies scenarios where one faces a privacy dilemma: either sensitive user data must be revealed to the entity that evaluates the cognitive model (e.g., in the Cloud), or the model itself must be revealed to the user so that the evaluation can take place locally. Fully Homomorphic Encryption (FHE) offers an elegant way to reconcile these conflicting interests in the Cloud-based scenario and also preserve non-interactivity. However, due to the inefficiency of existing FHE schemes, most applications prefer to use Somewhat Homomorphic Encryption (SHE), where the complexity of the computation to be performed has to be known in advance, and the efficiency of the scheme depends on this global complexity. In this paper, we present a new framework for homomorphic evaluation of neural networks, that we call FHE\u2013DiNN, whose complexity is strictly linear in the depth of the network and whose parameters can be set beforehand. To obtain this scale-invariance property, we rely heavily on the bootstrapping procedure. We refine the recent FHE construction by Chillotti et al. (ASIACRYPT 2016) in order to increase the message space and apply the sign function (that we use to activate the neurons in the network) during the bootstrapping. We derive some empirical results, using TFHE library as a starting point, and classify encrypted images from the MNIST dataset with more than 96% accuracy in less than 1.7 s. Finally, as a side contribution, we analyze and introduce some variations to the bootstrapping technique of Chillotti et al. that offer an improvement in efficiency at the cost of increasing the storage requirements.", 
        "editor": [
          {
            "familyName": "Shacham", 
            "givenName": "Hovav", 
            "type": "Person"
          }, 
          {
            "familyName": "Boldyreva", 
            "givenName": "Alexandra", 
            "type": "Person"
          }
        ], 
        "genre": "chapter", 
        "id": "sg:pub.10.1007/978-3-319-96878-0_17", 
        "inLanguage": [
          "en"
        ], 
        "isAccessibleForFree": false, 
        "isFundedItemOf": [
          {
            "id": "sg:grant.3792429", 
            "type": "MonetaryGrant"
          }, 
          {
            "id": "sg:grant.7738466", 
            "type": "MonetaryGrant"
          }
        ], 
        "isPartOf": {
          "isbn": [
            "978-3-319-96877-3", 
            "978-3-319-96878-0"
          ], 
          "name": "Advances in Cryptology \u2013 CRYPTO 2018", 
          "type": "Book"
        }, 
        "name": "Fast Homomorphic Evaluation of Deep Discretized Neural Networks", 
        "pagination": "483-512", 
        "productId": [
          {
            "name": "doi", 
            "type": "PropertyValue", 
            "value": [
              "10.1007/978-3-319-96878-0_17"
            ]
          }, 
          {
            "name": "readcube_id", 
            "type": "PropertyValue", 
            "value": [
              "4a5bdf9807cf386b8b216a6f853300476bdfae53463055841a38a190be72bade"
            ]
          }, 
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "pub.1105780951"
            ]
          }
        ], 
        "publisher": {
          "location": "Cham", 
          "name": "Springer International Publishing", 
          "type": "Organisation"
        }, 
        "sameAs": [
          "https://doi.org/10.1007/978-3-319-96878-0_17", 
          "https://app.dimensions.ai/details/publication/pub.1105780951"
        ], 
        "sdDataset": "chapters", 
        "sdDatePublished": "2019-04-16T05:01", 
        "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
        "sdPublisher": {
          "name": "Springer Nature - SN SciGraph project", 
          "type": "Organization"
        }, 
        "sdSource": "s3://com-uberresearch-data-dimensions-target-20181106-alternative/cleanup/v134/2549eaecd7973599484d7c17b260dba0a4ecb94b/merge/v9/a6c9fde33151104705d4d7ff012ea9563521a3ce/jats-lookup/v90/0000000325_0000000325/records_100809_00000000.jsonl", 
        "type": "Chapter", 
        "url": "https://link.springer.com/10.1007%2F978-3-319-96878-0_17"
      }
    ]
     

    Download the RDF metadata as:  json-ld nt turtle xml License info

    HOW TO GET THIS DATA PROGRAMMATICALLY:

    JSON-LD is a popular format for linked data which is fully compatible with JSON.

    curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-96878-0_17'

    N-Triples is a line-based linked data format ideal for batch operations.

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-96878-0_17'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-96878-0_17'

    RDF/XML is a standard XML format for linked data.

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-96878-0_17'


     

    This table displays all metadata directly associated to this object as RDF triples.

    215 TRIPLES      23 PREDICATES      56 URIs      19 LITERALS      8 BLANK NODES

    Subject Predicate Object
    1 sg:pub.10.1007/978-3-319-96878-0_17 schema:about anzsrc-for:08
    2 anzsrc-for:0801
    3 schema:author Nad7536b5b10448f19ea0bd532606c509
    4 schema:citation sg:pub.10.1007/11787006_1
    5 sg:pub.10.1007/978-3-319-31301-6_2
    6 sg:pub.10.1007/978-3-319-70694-8_14
    7 sg:pub.10.1007/978-3-540-89255-7_23
    8 sg:pub.10.1007/978-3-642-13013-7_25
    9 sg:pub.10.1007/978-3-642-13190-5_1
    10 sg:pub.10.1007/978-3-642-13190-5_2
    11 sg:pub.10.1007/978-3-642-17373-8_22
    12 sg:pub.10.1007/978-3-642-22792-9_29
    13 sg:pub.10.1007/978-3-642-32009-5_49
    14 sg:pub.10.1007/978-3-642-40041-4_1
    15 sg:pub.10.1007/978-3-642-40041-4_5
    16 sg:pub.10.1007/978-3-642-41320-9_13
    17 sg:pub.10.1007/978-3-662-44371-2_17
    18 sg:pub.10.1007/978-3-662-44371-2_31
    19 sg:pub.10.1007/978-3-662-46447-2_33
    20 sg:pub.10.1007/978-3-662-46800-5_24
    21 sg:pub.10.1007/978-3-662-46800-5_25
    22 sg:pub.10.1007/978-3-662-53015-3_12
    23 sg:pub.10.1007/978-3-662-53887-6_1
    24 sg:pub.10.1007/bf02551274
    25 https://doi.org/10.1016/0893-6080(91)90009-t
    26 https://doi.org/10.1109/5.726791
    27 https://doi.org/10.1109/sp.2017.12
    28 https://doi.org/10.1109/tc.2015.2470255
    29 https://doi.org/10.1137/1.9781611974782.160
    30 https://doi.org/10.1145/1060590.1060603
    31 https://doi.org/10.1145/2554797.2554799
    32 https://doi.org/10.1145/335191.335438
    33 https://doi.org/10.14722/ndss.2015.23241
    34 schema:datePublished 2018-07-24
    35 schema:datePublishedReg 2018-07-24
    36 schema:description The rise of machine learning as a service multiplies scenarios where one faces a privacy dilemma: either sensitive user data must be revealed to the entity that evaluates the cognitive model (e.g., in the Cloud), or the model itself must be revealed to the user so that the evaluation can take place locally. Fully Homomorphic Encryption (FHE) offers an elegant way to reconcile these conflicting interests in the Cloud-based scenario and also preserve non-interactivity. However, due to the inefficiency of existing FHE schemes, most applications prefer to use Somewhat Homomorphic Encryption (SHE), where the complexity of the computation to be performed has to be known in advance, and the efficiency of the scheme depends on this global complexity. In this paper, we present a new framework for homomorphic evaluation of neural networks, that we call FHE–DiNN, whose complexity is strictly linear in the depth of the network and whose parameters can be set beforehand. To obtain this scale-invariance property, we rely heavily on the bootstrapping procedure. We refine the recent FHE construction by Chillotti et al. (ASIACRYPT 2016) in order to increase the message space and apply the sign function (that we use to activate the neurons in the network) during the bootstrapping. We derive some empirical results, using TFHE library as a starting point, and classify encrypted images from the MNIST dataset with more than 96% accuracy in less than 1.7 s. Finally, as a side contribution, we analyze and introduce some variations to the bootstrapping technique of Chillotti et al. that offer an improvement in efficiency at the cost of increasing the storage requirements.
    37 schema:editor Nf4302124b37c415e85eb96c029867ccd
    38 schema:genre chapter
    39 schema:inLanguage en
    40 schema:isAccessibleForFree false
    41 schema:isPartOf N96b5951e4d4e42c39ee7bb9f726998fd
    42 schema:name Fast Homomorphic Evaluation of Deep Discretized Neural Networks
    43 schema:pagination 483-512
    44 schema:productId N0c5fee0b85574affbaeeaf7407d6eca3
    45 N960a71ac5898458b810c31daf43f6cf6
    46 Nf06981b8ea134d5693f706a73e83430b
    47 schema:publisher Nc569168bfb134f128d52441130cb4e14
    48 schema:sameAs https://app.dimensions.ai/details/publication/pub.1105780951
    49 https://doi.org/10.1007/978-3-319-96878-0_17
    50 schema:sdDatePublished 2019-04-16T05:01
    51 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    52 schema:sdPublisher Ne96707fea23f431a89e71ea19d85dc0c
    53 schema:url https://link.springer.com/10.1007%2F978-3-319-96878-0_17
    54 sgo:license sg:explorer/license/
    55 sgo:sdDataset chapters
    56 rdf:type schema:Chapter
    57 N0a12a422795f4b1f95b71c2b88fcdad6 rdf:first N559bc24711d340ba8db9a29b7856cf7b
    58 rdf:rest N1520ab924057451b9c394db2c360ba77
    59 N0c5fee0b85574affbaeeaf7407d6eca3 schema:name dimensions_id
    60 schema:value pub.1105780951
    61 rdf:type schema:PropertyValue
    62 N1520ab924057451b9c394db2c360ba77 rdf:first sg:person.012202553435.44
    63 rdf:rest rdf:nil
    64 N31830653b7684102b07e143c685f0dd0 schema:familyName Shacham
    65 schema:givenName Hovav
    66 rdf:type schema:Person
    67 N3da6798946a64bc3bc057d27be47a1a7 rdf:first Nd8ced7d576b4442bb33c6c8e2c8cf445
    68 rdf:rest rdf:nil
    69 N559bc24711d340ba8db9a29b7856cf7b schema:affiliation https://www.grid.ac/institutes/grid.5570.7
    70 schema:familyName Minihold
    71 schema:givenName Matthias
    72 rdf:type schema:Person
    73 N960a71ac5898458b810c31daf43f6cf6 schema:name readcube_id
    74 schema:value 4a5bdf9807cf386b8b216a6f853300476bdfae53463055841a38a190be72bade
    75 rdf:type schema:PropertyValue
    76 N96b5951e4d4e42c39ee7bb9f726998fd schema:isbn 978-3-319-96877-3
    77 978-3-319-96878-0
    78 schema:name Advances in Cryptology – CRYPTO 2018
    79 rdf:type schema:Book
    80 Nad7536b5b10448f19ea0bd532606c509 rdf:first sg:person.013454106661.29
    81 rdf:rest Nee95c37a5beb43b4b25c5b56b6b00011
    82 Nc569168bfb134f128d52441130cb4e14 schema:location Cham
    83 schema:name Springer International Publishing
    84 rdf:type schema:Organisation
    85 Nd8ced7d576b4442bb33c6c8e2c8cf445 schema:familyName Boldyreva
    86 schema:givenName Alexandra
    87 rdf:type schema:Person
    88 Ne96707fea23f431a89e71ea19d85dc0c schema:name Springer Nature - SN SciGraph project
    89 rdf:type schema:Organization
    90 Nee95c37a5beb43b4b25c5b56b6b00011 rdf:first sg:person.015114730252.70
    91 rdf:rest N0a12a422795f4b1f95b71c2b88fcdad6
    92 Nf06981b8ea134d5693f706a73e83430b schema:name doi
    93 schema:value 10.1007/978-3-319-96878-0_17
    94 rdf:type schema:PropertyValue
    95 Nf4302124b37c415e85eb96c029867ccd rdf:first N31830653b7684102b07e143c685f0dd0
    96 rdf:rest N3da6798946a64bc3bc057d27be47a1a7
    97 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
    98 schema:name Information and Computing Sciences
    99 rdf:type schema:DefinedTerm
    100 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
    101 schema:name Artificial Intelligence and Image Processing
    102 rdf:type schema:DefinedTerm
    103 sg:grant.3792429 http://pending.schema.org/fundedItem sg:pub.10.1007/978-3-319-96878-0_17
    104 rdf:type schema:MonetaryGrant
    105 sg:grant.7738466 http://pending.schema.org/fundedItem sg:pub.10.1007/978-3-319-96878-0_17
    106 rdf:type schema:MonetaryGrant
    107 sg:person.012202553435.44 schema:affiliation https://www.grid.ac/institutes/grid.470554.7
    108 schema:familyName Paillier
    109 schema:givenName Pascal
    110 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012202553435.44
    111 rdf:type schema:Person
    112 sg:person.013454106661.29 schema:affiliation https://www.grid.ac/institutes/grid.89485.38
    113 schema:familyName Bourse
    114 schema:givenName Florian
    115 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013454106661.29
    116 rdf:type schema:Person
    117 sg:person.015114730252.70 schema:affiliation https://www.grid.ac/institutes/grid.5328.c
    118 schema:familyName Minelli
    119 schema:givenName Michele
    120 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.015114730252.70
    121 rdf:type schema:Person
    122 sg:pub.10.1007/11787006_1 schema:sameAs https://app.dimensions.ai/details/publication/pub.1035160724
    123 https://doi.org/10.1007/11787006_1
    124 rdf:type schema:CreativeWork
    125 sg:pub.10.1007/978-3-319-31301-6_2 schema:sameAs https://app.dimensions.ai/details/publication/pub.1036619867
    126 https://doi.org/10.1007/978-3-319-31301-6_2
    127 rdf:type schema:CreativeWork
    128 sg:pub.10.1007/978-3-319-70694-8_14 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093076441
    129 https://doi.org/10.1007/978-3-319-70694-8_14
    130 rdf:type schema:CreativeWork
    131 sg:pub.10.1007/978-3-540-89255-7_23 schema:sameAs https://app.dimensions.ai/details/publication/pub.1021516251
    132 https://doi.org/10.1007/978-3-540-89255-7_23
    133 rdf:type schema:CreativeWork
    134 sg:pub.10.1007/978-3-642-13013-7_25 schema:sameAs https://app.dimensions.ai/details/publication/pub.1039389499
    135 https://doi.org/10.1007/978-3-642-13013-7_25
    136 rdf:type schema:CreativeWork
    137 sg:pub.10.1007/978-3-642-13190-5_1 schema:sameAs https://app.dimensions.ai/details/publication/pub.1025305533
    138 https://doi.org/10.1007/978-3-642-13190-5_1
    139 rdf:type schema:CreativeWork
    140 sg:pub.10.1007/978-3-642-13190-5_2 schema:sameAs https://app.dimensions.ai/details/publication/pub.1023605606
    141 https://doi.org/10.1007/978-3-642-13190-5_2
    142 rdf:type schema:CreativeWork
    143 sg:pub.10.1007/978-3-642-17373-8_22 schema:sameAs https://app.dimensions.ai/details/publication/pub.1021970215
    144 https://doi.org/10.1007/978-3-642-17373-8_22
    145 rdf:type schema:CreativeWork
    146 sg:pub.10.1007/978-3-642-22792-9_29 schema:sameAs https://app.dimensions.ai/details/publication/pub.1049197805
    147 https://doi.org/10.1007/978-3-642-22792-9_29
    148 rdf:type schema:CreativeWork
    149 sg:pub.10.1007/978-3-642-32009-5_49 schema:sameAs https://app.dimensions.ai/details/publication/pub.1011363442
    150 https://doi.org/10.1007/978-3-642-32009-5_49
    151 rdf:type schema:CreativeWork
    152 sg:pub.10.1007/978-3-642-40041-4_1 schema:sameAs https://app.dimensions.ai/details/publication/pub.1049650086
    153 https://doi.org/10.1007/978-3-642-40041-4_1
    154 rdf:type schema:CreativeWork
    155 sg:pub.10.1007/978-3-642-40041-4_5 schema:sameAs https://app.dimensions.ai/details/publication/pub.1006864000
    156 https://doi.org/10.1007/978-3-642-40041-4_5
    157 rdf:type schema:CreativeWork
    158 sg:pub.10.1007/978-3-642-41320-9_13 schema:sameAs https://app.dimensions.ai/details/publication/pub.1028621484
    159 https://doi.org/10.1007/978-3-642-41320-9_13
    160 rdf:type schema:CreativeWork
    161 sg:pub.10.1007/978-3-662-44371-2_17 schema:sameAs https://app.dimensions.ai/details/publication/pub.1047019051
    162 https://doi.org/10.1007/978-3-662-44371-2_17
    163 rdf:type schema:CreativeWork
    164 sg:pub.10.1007/978-3-662-44371-2_31 schema:sameAs https://app.dimensions.ai/details/publication/pub.1003203986
    165 https://doi.org/10.1007/978-3-662-44371-2_31
    166 rdf:type schema:CreativeWork
    167 sg:pub.10.1007/978-3-662-46447-2_33 schema:sameAs https://app.dimensions.ai/details/publication/pub.1006063386
    168 https://doi.org/10.1007/978-3-662-46447-2_33
    169 rdf:type schema:CreativeWork
    170 sg:pub.10.1007/978-3-662-46800-5_24 schema:sameAs https://app.dimensions.ai/details/publication/pub.1037378777
    171 https://doi.org/10.1007/978-3-662-46800-5_24
    172 rdf:type schema:CreativeWork
    173 sg:pub.10.1007/978-3-662-46800-5_25 schema:sameAs https://app.dimensions.ai/details/publication/pub.1042169019
    174 https://doi.org/10.1007/978-3-662-46800-5_25
    175 rdf:type schema:CreativeWork
    176 sg:pub.10.1007/978-3-662-53015-3_12 schema:sameAs https://app.dimensions.ai/details/publication/pub.1051267410
    177 https://doi.org/10.1007/978-3-662-53015-3_12
    178 rdf:type schema:CreativeWork
    179 sg:pub.10.1007/978-3-662-53887-6_1 schema:sameAs https://app.dimensions.ai/details/publication/pub.1084896599
    180 https://doi.org/10.1007/978-3-662-53887-6_1
    181 rdf:type schema:CreativeWork
    182 sg:pub.10.1007/bf02551274 schema:sameAs https://app.dimensions.ai/details/publication/pub.1023250347
    183 https://doi.org/10.1007/bf02551274
    184 rdf:type schema:CreativeWork
    185 https://doi.org/10.1016/0893-6080(91)90009-t schema:sameAs https://app.dimensions.ai/details/publication/pub.1050371510
    186 rdf:type schema:CreativeWork
    187 https://doi.org/10.1109/5.726791 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061179979
    188 rdf:type schema:CreativeWork
    189 https://doi.org/10.1109/sp.2017.12 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093587091
    190 rdf:type schema:CreativeWork
    191 https://doi.org/10.1109/tc.2015.2470255 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061536082
    192 rdf:type schema:CreativeWork
    193 https://doi.org/10.1137/1.9781611974782.160 schema:sameAs https://app.dimensions.ai/details/publication/pub.1098556665
    194 rdf:type schema:CreativeWork
    195 https://doi.org/10.1145/1060590.1060603 schema:sameAs https://app.dimensions.ai/details/publication/pub.1012332159
    196 rdf:type schema:CreativeWork
    197 https://doi.org/10.1145/2554797.2554799 schema:sameAs https://app.dimensions.ai/details/publication/pub.1042352254
    198 rdf:type schema:CreativeWork
    199 https://doi.org/10.1145/335191.335438 schema:sameAs https://app.dimensions.ai/details/publication/pub.1063168485
    200 rdf:type schema:CreativeWork
    201 https://doi.org/10.14722/ndss.2015.23241 schema:sameAs https://app.dimensions.ai/details/publication/pub.1095873103
    202 rdf:type schema:CreativeWork
    203 https://www.grid.ac/institutes/grid.470554.7 schema:alternateName CryptoExperts (France)
    204 schema:name CryptoExperts, Paris, France
    205 rdf:type schema:Organization
    206 https://www.grid.ac/institutes/grid.5328.c schema:alternateName French Institute for Research in Computer Science and Automation
    207 schema:name DIENS, École normale supérieure, CNRS, PSL Research University, Paris, France
    208 Inria, Paris, France
    209 rdf:type schema:Organization
    210 https://www.grid.ac/institutes/grid.5570.7 schema:alternateName Ruhr University Bochum
    211 schema:name Horst Görtz Institut für IT-Security, Ruhr-Universität Bochum, Bochum, Germany
    212 rdf:type schema:Organization
    213 https://www.grid.ac/institutes/grid.89485.38 schema:alternateName Orange (France)
    214 schema:name Orange Labs, Applied Crypto Group, Cesson-Sévigné, France
    215 rdf:type schema:Organization
     




    Preview window. Press ESC to close (or click here)


    ...