Unlock instant, AI-driven research and patent intelligence for your innovation.

Lightweight convolutional neural network security prediction method

A convolutional neural network and security prediction technology, applied in the field of lightweight convolutional neural network security prediction, can solve the problems of low FV encryption efficiency and difficult application, and achieve the goal of reducing weight parameters, deepening layers, and improving prediction efficiency Effect

Active Publication Date: 2021-01-12
JINAN UNIVERSITY
View PDF4 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

When the neural network level is relatively deep, there will be more parameters, and a large number of ciphertext operations will be generated when the neural network is predicted, making the FV encryption efficiency very low, and it is difficult to apply it to the actual scene

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
  • Lightweight convolutional neural network security prediction method
  • Lightweight convolutional neural network security prediction method
  • Lightweight convolutional neural network security prediction method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0063] A lightweight convolutional neural network security prediction method, such as figure 1 shown, including the following steps:

[0064] Construct a network security prediction model through training samples; specifically: obtain training samples and perform preprocessing; randomly select training samples for convolution and pooling, and output from the fully connected layer; backpropagation to adjust network weights to obtain a network security prediction model . Specific process such as figure 2 shown.

[0065] Perform filter pruning on the network security prediction model to obtain a pruned network security prediction model; specifically: according to each filter, the smaller the sum of the absolute values ​​of the convolution kernels of each channel, the importance of the filter The lower, each layer selects the m least important filters for pruning. The pruning method is to remove some less important filters from a trained model while minimizing the loss of acc...

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 lightweight convolutional neural network security prediction method. The method comprises the following steps of constructing a network security prediction model through a training sample; performing filter pruning on the network security prediction model to obtain a pruned network security prediction model; encoding the model parameters of the pruning network security prediction model to obtain an encoded network security prediction model; inputting the data uploaded by the user into a coding network security prediction model, and encrypting the data uploaded by theuser to obtain a ciphertext; through ciphertext prediction, a ciphertext prediction result being obtained, and then a final data result being obtained. According to the method, prediction and analysisof data are realized, the cloud server is ensured not to acquire any effective information of the user, and the user is ensured not to acquire any information of the model on the cloud server.

Description

technical field [0001] The invention relates to the research field of network security, in particular to a lightweight convolutional neural network security prediction method. Background technique [0002] In recent years, artificial intelligence has developed rapidly in various fields. Due to the blowout of data volume, breakthroughs in computing power, and breakthroughs in algorithms, deep learning has achieved great success in various fields. The main difference between deep learning and traditional machine learning is that its performance increases with the increase of data size. Deep learning technology has amazing applications in many fields such as medical diagnosis, face recognition and credit risk assessment. Large Internet companies collect users' online behavior data, such as users' personal information, web pages they like to browse, and things they often buy, and use these data to train recommendation systems, and then analyze and predict users' interests. Hos...

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
IPC IPC(8): H04L12/24H04L9/00G06N3/04G06N3/08G06F21/60G06F21/62
CPCH04L41/147H04L41/145H04L9/008G06N3/082G06N3/084G06F21/602G06F21/6245G06N3/048G06N3/045Y02D30/50
Inventor 周德华杨诗吟赖俊祚王传胜
Owner JINAN UNIVERSITY