Medical purpose neural network robustness verification method and system based on shell protection

A robustness verification and neural network technology, applied in the field of neural network robustness verification for medical purposes, can solve the lack of methods and systems for neural network robustness verification, to improve robustness and reduce time complexity , the effect of improving operating efficiency

Active Publication Date: 2021-04-30
SHANDONG UNIV
View PDF6 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0016] (2) The verification problem of the deep neural network: the verification problem of the neural network is to verify whether the input-output relationship of a neural network is established
[0020] (4) There is a lack of a method and system for robustness verification of neural networks for medical purposes in the prior art

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
  • Medical purpose neural network robustness verification method and system based on shell protection
  • Medical purpose neural network robustness verification method and system based on shell protection
  • Medical purpose neural network robustness verification method and system based on shell protection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0048] This embodiment provides a method for verifying the robustness of the neural network for medical use based on shell preservation;

[0049] Such as figure 1 As shown, the robustness verification method of neural network for medical use based on shell preservation includes:

[0050] S101: Obtain physical examination data of healthy users; generate an input set based on the physical examination data of healthy users;

[0051] S102: Take the input set as the root node of the binary tree, each layer node is obtained by bisecting the super rectangle represented by its parent node, until the super rectangle represented by a certain layer node reaches the specified size limit, then no Continue to bisect, the current layer node is the leaf node, and obtain the binary tree for searching;

[0052] S103: Traverse the binary tree, search each node of the binary tree, input the hyperrectangle represented by each node of the binary tree into the neural network to be verified, and th...

Embodiment 2

[0174] This embodiment provides a robustness verification system for medical use neural networks based on shell preservation;

[0175] The robustness verification system of neural network for medical use based on shell preservation, including:

[0176] The input set generation module is configured to: acquire the physical examination data of the healthy user; generate the input set based on the physical examination data of the healthy user;

[0177] Binary tree construction module, which is configured to: use the input set as the root node of the binary tree, each layer of nodes is obtained by bisecting the super rectangle represented by its parent node, until the super rectangle represented by a certain layer of nodes reaches The specified size limit, then do not continue to bisect, the current layer node is the leaf node, and the binary tree for searching is obtained;

[0178] The security verification module is configured to: traverse the binary tree, search each node of t...

Embodiment 3

[0183]This embodiment also provides an electronic device, including: one or more processors, one or more memories, and one or more computer programs; wherein, the processor is connected to the memory, and the one or more computer programs are programmed Stored in the memory, when the electronic device is running, the processor executes one or more computer programs stored in the memory, so that the electronic device executes the method described in Embodiment 1 above.

[0184] It should be understood that in this embodiment, the processor can be a central processing unit CPU, and the processor can also be other general-purpose processors, digital signal processors DSP, application specific integrated circuits ASIC, off-the-shelf programmable gate array FPGA or other programmable logic devices , discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or...

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 medical purpose neural network robustness verification method and system based on shell protection. Physical examination data of a healthy user is collected; generating an input set based on the physical examination data of the healthy user; the input set serves as a root node of the binary tree, each layer of node is obtained by bisecting a super rectangle represented by a father node of the node until the super rectangle represented by a certain layer of node reaches the specified size limit, bisecting is not continued any more, the current layer of node is a leaf node, and the binary tree for searching is obtained; traversing the binary tree, searching each node of the binary tree, inputting the hyper-rectangle represented by each node of the binary tree into a neural network to be verified, and outputting a reachable set by the neural network to be verified; if it is found that the output reachable set estimation of the hyper-rectangle represented by a certain leaf node is not included in the output limitation, obtaining a conclusion that the current neural network to be verified is unsafe; otherwise, obtaining a conclusion that the current to-be-verified neural network is safe.

Description

technical field [0001] This application relates to the technical field of neural network security verification or trusted artificial intelligence, in particular to a method and system for robust verification of neural networks for medical purposes based on shell preservation. Background technique [0002] The statements in this section merely mention the background art related to this application, and do not necessarily constitute the prior art. [0003] At present, neural networks have been widely used in natural language processing, speech recognition, image recognition, autonomous driving, malware detection, medical fields, etc. Especially in the medical field, with the aging of the population, the existing medical system is becoming more and more difficult to meet the medical needs of the society as a whole. Therefore, emerging technologies such as artificial intelligence based on neural networks are more and more widely used. The neural network has achieved high accura...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/044G06N3/045G06F18/24323
Inventor 郭山清唐朋张云若
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap