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

Vulnerability detection method based on function and branch coverage in combination with neural network

A technology of vulnerability detection and neural network, which is applied in the direction of neural learning method, biological neural network model, neural architecture, etc., can solve problems such as dependence on the solving ability and efficiency of constraint solvers, and program state explosion, so as to alleviate the problem of path explosion and improve Execution efficiency and the effect of improving the problem of blind mutation

Pending Publication Date: 2021-10-29
四川省信息安全测评中心
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Symbolic execution can avoid the generation of redundant test inputs, but it depends on the solving ability and efficiency of the constraint solver, and is limited by the program state explosion problem

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
  • Vulnerability detection method based on function and branch coverage in combination with neural network
  • Vulnerability detection method based on function and branch coverage in combination with neural network
  • Vulnerability detection method based on function and branch coverage in combination with neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0038] Such as figure 1 As shown, the vulnerability detection method based on function and branch coverage combined with neural network includes:

[0039] S1. Obtain a program to be tested for vulnerability detection.

[0040] S2. Fuzzing the program to be tested, and monitoring the coverage information of functions and branches during the fuzzing process. Generally, the coverage information of the functions and branches is recorded while monitoring the coverage information of the functions and branches.

[0041] In some embodiments, the fuzzing is machine learning based fuzzing. Specifically, the coverage-oriented LSTM RNN neural network training guides the mutation strategy of the fuzzing test; that is, through the neural network of machine learning, the random number generation of each fuzzer that needs to be tested is reduced, and the accuracy is improved. Rate. Compared with the traditional fuzz test, the fuzz test based on machine learning reduces the blindness of mu...

Embodiment 2

[0049] Such as image 3 As shown, the vulnerability detection method based on function and branch coverage combined with neural network includes:

[0050] S´1. Obtain the program under test to perform vulnerability detection.

[0051] S´2. Fuzz the program to be tested, and monitor the coverage information of functions and branches during the fuzzing process. Generally, the coverage information of the functions and branches is recorded while monitoring the coverage information of the functions and branches.

[0052] S´3. Determine whether there are uncovered functions and branches, if there are uncovered functions and branches, execute S´4, otherwise execute S´8.

[0053] S´4. Obtain the instrumentation source code through disassembly and instrumentation technology.

[0054] S´5. Use instrumented source code to guide mixed symbolic execution to explore uncovered functions and branches, and use instrumented source code to guide mixed symbolic execution to solve uncovered fun...

Embodiment 3

[0060] Such as Figure 4 As shown, the vulnerability detection method based on function and branch coverage combined with neural network includes:

[0061] S´´1. Obtain the program under test to perform vulnerability detection.

[0062] S´´2. Compile the source code of the program to be tested to generate a binary program.

[0063] S´´3. Fuzz the program under test, and monitor the coverage information of functions and branches during the fuzzing process. Generally, the coverage information of the functions and branches is recorded while monitoring the coverage information of the functions and branches.

[0064] S´´4. Determine whether there are uncovered functions and branches, if there are uncovered functions and branches, execute S´´5, otherwise execute S´´9.

[0065] S´´5. Update the source code of uncovered functions and branches by instrumentation.

[0066] S´´6. Use the updated source code to guide mixed symbolic execution to explore uncovered functions and branches...

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 vulnerability detection method based on function and branch coverage in combination with a neural network. The method comprises the following steps: S1, obtaining a to-be-detected program to be subjected to vulnerability detection; S2, performing fuzz testing on the to-be-tested program, and monitoring coverage information of functions and branches in the fuzz testing process; S3, judging whether uncovered functions and branches exist or not, if the uncovered functions and branches exist, executing S4, and if not, executing S7; S4, exploring uncovered functions and branches by using mixed symbols, and solving the uncovered functions and branches; S5, generating a new test case by utilizing hybrid coincidence execution; S6, updating the test case of the fuzzy test by using the new test case, and executing the step S2; and S7, ending vulnerability detection. By combining the advantages of fuzzy testing and symbolic execution, the code coverage rate is increased, and the vulnerability mining capability is further enhanced.

Description

technical field [0001] The invention relates to the field of software system security detection, in particular to a vulnerability detection method combined with neural network based on function and branch coverage. Background technique [0002] With the development of information technology, software has penetrated into all aspects of modern society, and the scale and complexity of software are increasing day by day. However, due to the software loopholes introduced by improper development, its security has been paid more and more attention. In today's society, software security involves various industries of the national economy and the people's livelihood, and hidden dangers of software security vulnerabilities can cause irreparable and serious consequences. Software vulnerability mining is an effective technical means to ensure software security, and a high-coverage, high-efficiency vulnerability mining method can effectively reduce the hidden dangers of vulnerabilities. ...

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): G06F21/57G06F11/36G06N3/04G06N3/08
CPCG06F21/577G06F11/3676G06F11/3684G06F11/3688G06N3/08G06N3/044G06N3/045
Inventor 王丹琛徐鹏李志辉郭伟吴贯锋彭浩
Owner 四川省信息安全测评中心