Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Fully homomorphic encryption deep learning reasoning method and system based on FPGA

A fully homomorphic encryption and deep learning technology, applied in the field of homomorphic encryption algorithms, it can solve the problems of complex computing, complexity, the combination of homomorphic encryption and deep learning, etc., and achieve the effect of speeding up the inference speed and improving the inference speed.

Pending Publication Date: 2021-04-23
SHANDONG UNIV
View PDF4 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the inventor believes that the homomorphic encryption technology has certain limitations. The data encrypted by the homomorphic encryption algorithm is usually thousands of times larger than before; in addition, the operations supported by homomorphic encryption only support addition and multiplication, which is not yet mature. The scheme allows it to support operations such as comparison; and only multiplication is much more complicated than the multiplication of ordinary numbers, because in order to limit the growth of ciphertexts, it is necessary to perform reproducibility operations (RelinearizationOperation) after multiplication; therefore complex calculations and a large number of The combination of homomorphic encryption and deep learning is limited due to the occupation of storage resources

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Fully homomorphic encryption deep learning reasoning method and system based on FPGA
  • Fully homomorphic encryption deep learning reasoning method and system based on FPGA
  • Fully homomorphic encryption deep learning reasoning method and system based on FPGA

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] Such as figure 1 As shown, this embodiment provides an FPGA-based fully homomorphic encryption deep learning reasoning method, including:

[0052] S1: Obtain the ciphertext encrypted by the homomorphic encryption algorithm and the corresponding encoded plaintext;

[0053] S2: Obtain the multiplication depth, data processing scale and network layer of the initial deep learning network;

[0054] S3: Determine the value range of the coefficient modulus of the polynomial according to the number of items of the plaintext polynomial, determine the value number according to the multiplication depth, and determine the coefficient modulus according to the coefficient modulus selected by the error parameter;

[0055] S4: Determine the weight and deviation of the network layer according to the number of items and coefficient modulus of the polynomial and the scale of data processing, so as to obtain the packaging strategy of the network layer;

[0056] S5: According to the packa...

Embodiment 2

[0100] This embodiment provides an FPGA-based fully homomorphic encryption deep learning reasoning system, including:

[0101] The first obtaining module is used to obtain the ciphertext encrypted by the homomorphic encryption algorithm and the corresponding encoded plaintext;

[0102] The second acquisition module is used to acquire the multiplication depth, data processing scale and network layer of the initial deep learning network;

[0103] The homomorphic encryption parameter determination module is used to determine the value range of the coefficient modulus of the polynomial according to the number of items of the plaintext polynomial, determine the value number according to the multiplication depth, and determine the coefficient modulus according to the coefficient modulus selected by the error parameter;

[0104] A packaging strategy determination module is used to determine the weight and deviation of the network layer according to the number of items and coefficient...

Embodiment 3

[0108] According to the network structure and homomorphic encryption parameters, the optimal design of the network model software level is obtained. However, the complex memory access mode hinders the direct use of multi-threading and other parallel technologies for optimization. Therefore, directly deploying it on the CPU cannot fully utilize the system. Performance; in order to achieve optimal performance indicators such as delay, throughput, and power consumption of the entire system, this embodiment provides a fully homomorphic encryption deep learning reasoning platform based on FPGA, such as Figure 5 shown;

[0109] The platform adopts the code system and HLS tools to construct a reasoning model based on the FPGA-based fully homomorphic encryption deep learning reasoning method described in Example 1, and realizes the homomorphic encryption data reasoning method. The deep learning network that supports CKKS homomorphic encryption will use high The hierarchical synthesis...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a fully homomorphic encryption deep learning reasoning method and system based on an FPGA. The method comprises the steps of obtaining a ciphertext encrypted by a homomorphic encryption algorithm and a coded plaintext; obtaining the multiplication depth, the data processing scale and the network layer of the initial deep learning network; determining a value range of a coefficient module factor of the polynomial according to the term number of the plaintext polynomial, determining the number of values according to the multiplication depth, and determining a coefficient module according to the coefficient module factor selected by the error parameter; determining the weight and deviation of the network layer according to the item number and coefficient module of the polynomial and the data processing scale so as to obtain a packaging strategy of the network layer; according to the packaging strategy and the plaintext, selection of the item number and the coefficient module of the polynomial is judged, a network layer is optimized, an inference model is constructed, and therefore a ciphertext inference result is output to the ciphertext. An accelerator is integrally designed for the combination of homomorphic encryption and a deep learning network by using an FPGA, so that the reasoning speed of homomorphic encryption data on the deep learning network is increased.

Description

technical field [0001] The invention relates to the technical field of homomorphic encryption algorithms, in particular to an FPGA-based fully homomorphic encryption deep learning reasoning method and system. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] Deep learning (Deep Learning) is a method of machine learning (Machine Learning), and machine learning is a branch of artificial intelligence. Deep learning uses multiple processing layers (neural networks) that contain complex structures or consist of multiple nonlinear transformations. The calculation is applied in image recognition, speech recognition and other technologies. Taking the medical field as an example, deep learning will be used to process patients' diagnostic information to predict their condition and other conditions; under the requirements of morality and law, it is ne...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06F21/60G06N5/04H04L9/00
CPCG06F21/602H04L9/008G06N5/04
Inventor 鞠雷诸怡兰韩明钦周梓梦郭山清
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products