Homomorphic
Encryption for Arithmetic of
Approximate Numbers

HEaaN provides the ability to analyze data while encrypted.
Therefore, even if a hacker stole the information while analyst computes on data,
the hacker will never get any information from it.

Existing homomorphic encryption scheme has difficulty in commercialization
since the size of ciphertexts exponentially increases when iterating operations with it.
However HEaaN discards small values of number which does not significantly
affect the calculation result, in order to increase the calculation speed.
HEaaN is the sole homomorphic encryption scheme that dramatically improves
the calculation speed by creating an approximate calculation method.

Given that most machine learning and medical data analytics require
real-number calculation, HEaaN is the only solution that provides real-number
calculation among homomorphic encryptions that can perform calculations
while protecting sensitive information.

HEaaN

Homomorphic Encryption provides the operation between ciphertexts, and it has the advantage that information is never exposed during data collection, storage, and processing. Accordingly, active research and development are being conducted in applications where personal information needs to be protected among cloud computing and Big Data applications. Compared to the computation with unencrypted data, computation with homomorphic encryption requires longer computation time and bigger size of data which makes inefficiency. Therefore, one of the most important performance measures of the homomorphic encryption algorithm is to minimize those inefficiencies.

IDASH
  • Host

    UCSD Medical School

    UCSD Medical School

  • Sponsor

    미국 국립보건원

    National Institutes of Health

  • Goal

    Analysis of genetic information
    by cloud computing without information leakage

HEaaN has been adopted as a candidate algorithm for international standardization organizations such as International Standard Organization (ISO), International Telecommunication Union (ITU), and Internet Engineering Task Force (IETF) by showing the best performance in this performance measure. As the algorithm of homomorphic encryption which enables efficient real number computations, HEaaN’s excellence has been proven through academics and world-class competitions, and its original technology has been secured through 14 patents.

Based on the patented technology, HEaaN is designed to be equipped with a module that enables basic encryption/decryption functions and efficient practical computation between ciphertexts. The typical operations using ciphertexts provided by HEaaN include addition/multiplication/rotation, bootstrapping, basic statistics, logistic regression, and machine learning. HEaaN includes a library that optimizes these core functions and modules with a software architecture that provides optimal adaptability to various applications. In addition HEaaN offers APIs (Application Programming Interfaces) which support application to important application system such as finance and health care, and various white papers that help to optimize those systems.

동형암호의 원리동형암호의 원리
2017 TRACK 3 : BEST-PERFORMING TEAMS

Evaluated on (three datasets of 1422 records for training/157 records for testing + 18 features)

2017 TRACK 3 : BEST-PERFORMING TEAMS
Teams AUC
0.7136
Encryption Secure learing Decryption Overall time
(mins)
Size(MB) Time
(mins)
Time
(mins)
Memory
(MB)
Size
(MB)
Time
(mins)
SNU 0.6934 537.667 0.060 10.250 2775.333 64.875 0.050 10.360
CEA LIST 0.6930 53.000 1.303 2206.057 238.255 0.350 0.003 2207.363
KU Leuven 0.6722 4904.000 4.304 155.695 7266.727 10.790 0.913 160.912
EPFL 0.6584 1011.750 1.633 15.089 1498.513 7.125 0.017 16.739
MAR 0.6574 1945.600 11.335 385.021 26299.344 76.000 0.033 396.390
Waseda* 0.7154 20.390 1.178 2.077 7635.600 20.390 2.077 5.332
Saarland N/A 65536.000 16.633 48.356 29752.527 65536 7.355 57.344

* Interactive mechanism, no complete guarantee on 80-bit security at “analyst” side

** Program ends with errors

2018 TRACK 2 : Secure Parallel Genome Wied Associaion Studies using Homomorphic Encryption
Team Submission Schemes End to End Performance Evaluation result (F1-Score)
at different cutoffs
Running time
(mins)
Peak Memory
(M)
0.01 0.001 0.0001 0.00001
Gold Semi Gold Semi Gold Semi Gold Semi
A*FHE A*FHE -1 + HEAAN 922.48 3,777 0.977 0.999 0.986 0.999 0.985 0.999 0.966 0.998
A*FHE -2 1,632.97 4,093 0.882 0.905 0.863 0.877 0.827 0.843 0.792 0.826
Chimera Version 1 + TFHE & HEAAN
(Chimera)
201.73 10,375 0.979 0.993 0.987 0.991 0.988 0.989 0.982 0.974
Version 2 215.95 15,166 0.339 0.35 0.305 0.309 0.271 0.276 0.239 0.253
Drlft Blue Delft Blue HEAAN 1,844.82 10,814 0.965 0.969 0.956 0.944 0.951 0.935 0.884 0.849
UC San Diego Logistic Regr + HEAAN 1.66 14,901 0.983 0.993 0.993 0.987 0.991 0.989 0.995 0.967
Linear Regr 0.42 3,387 0.982 0.989 0.980 0.971 0.982 0.968 0.925 0.89
Duality
Technologies
Logistic Regr + CKKS (Aka HEAAN),
pkg:palisade
3.8 10,230 0.982 0.993 0.991 0.993 0.993 0.991 0.990 0.973
Chi2 test 0.09 1,512 0.968 0.983 0.981 0.985 0.980 0.985 0.939 0.962
Seoul National University SNU-1 HEAAN 52.49 15,204 0.975 0.984 0.976 0.973 0.975 0.969 0.932 0.905
SNU-2 52.37 15,177 0.976 0.988 0.979 0.975 0.974 0.969 0.939 0.909
IBM IBM-Complex CKKS (Aka HEAAN),
Pkg:HEIIB
23.35 8,651 0.913 0.911 0.169 0.188 0.067 0.077 0.053 0.06
IBM-Real 52.65 15,613 0.542 0.526 0.279 0.28 0.241 0.255 0.218 0.229

2020 TRACK 1 : Secure multi-label Tumor classification using Homomorphic Encryption-1st Place

1.png

 

2.png

3.png

 


 

Thesis

Main Papers(Google Scholar Profile, dplp)

Video

Video

  • Lecture from Seoul National Unviersity, Department of Mathematical Science.

    Date
    2019.09.19
    Explanation
    Homomorphic Encryption, which provides the ability to compute on data while it is encrypted, was first proposed in 1978 and has been put to practical use in recent years of research. This lecture provides the concept of homomorphic encryption, recent research results and its application to machine learning.
    Link
  • SAMSUNG SDS REAL 2019 - KEYNOTE 2 (Key Technololgies in Digital Transformation)

    Date
    2019.05.08
    Explanation
    This video provides the development direction and characteristics of homomorphic encryption, and contents of cooperation with Samsung SDS
    Link
  • Introducing an evaluation model using encrypted credit information.

    Date
    2019.01.23
    Explanation
    At the seminar on Personal Identification Prevention Encryption Technology, Lee Wook-Jae, the director from the Korea Credit Bureau (KCB), presents the description of the homomorphic credit evaluation model using homomorphic encryption.
    Link
  • Seminar on Personal Identification Prevention Technology (Homomorphic Encryption Technology Trend and Demonstration)

    Date
    2019.01.22
    Explanation
    International standardization trends & Domestic and foreign use cases - Professor Jung Hee Cheon (Department of Mathematical Science, Seoul National University)
    Link
검색 닫기