HEaaN.STAT

HEaaN.STAT


Since the early stages of development, HEaaN has been designed in a form appropriate for reality by collaborating with various companies, which allows HEaaN to support services that can be applied to the customer’s business without losing originality of functions and performance. One of the most basic and frequently used functions on encrypted data by many applications is basic statistics.

Among the HEaaN’s various products, HEaaN.STAT offers basic statistics, sorting between ciphertexts, and logistic regression which has already proven its practicality and efficiency by cooperating with financial sectors.

 

HEaaN.ML

Privacy Preserving Machine Learning (PPML) is recognized for its necessity and importance from various industries that want to utilize AI (Artificial Intelligence). Among the various technical approaches in PPML, HEaaN.ML provides a solution through Machine Learning (ML) functions utilizing homomorphic encryption. HEaaN.ML provides DNN (Deep Neural Network), CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), and random forest functions for each program. In addition, in order to improve its performance, we are proceeding with supplying high-performance semiconductors through the use of GPUs (Graphics Processing Unit) and FPGAs (Field Programmable Gate Arrays) through joint research with academic world and hardware companies.

 

Learning Phase

데이터, 머신러닝, 함수


Prediction Phase

개인데이터x, 함수, 함수값

Demo Site – HEaaN Platform

HEaaN.STAT provides statistical computation and machine learning such as logistic regression based on homomorphic encryption functions.
For customers who want to develop their products and service using HEaaN.STAT, we offer a demo site where they can experience the features on the web.

데이터 소유자, 데이터 서비스 제공자, 키 관리자

 

The above figure is a diagram representing the institution of HEaaN.STAT by usage type, and the description of each institution is as follows.

  • 1

    Data Owner (DO) : The institution that owns the original data. They encrypt the original data and deliver it to the Data Service Provider (DSP).

  • 2

    Data Service Provider (DSP) : The institution that performs data analysis services using encrypted data received from Data Owner.

  • 3

    Key Manager (KM) : The institution that generates keys and manages secret keys. It distributes keys to each institution and provides data decryption services.

  •  

    The fundamental functions provided by HEaaN.STAT are commands for key generation, data encryption, data decryption, data merging and data management. Moreover, the functions to support the operation among ciphertexts consist of basic operation, basic statistics, bin statistics, and logistic regression.

    • 1

      Basic arithmetic operations support addition, subtraction, multiplication, constant addition, constant subtraction, and constant multiplication among columns.

    • 2

      Basic statistics provide the ability to compute the mean, variance, Pearson's correlation coefficient, percentile, and quartile. For Further information on bin statistics, logistic regression, and the detailed description of each is in the manual.

    검색 닫기