Code auditing method and device, electronic equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence, and provides a code auditing method and device. The code auditing method comprises the following steps: acquiring a to-be-detected first code fragment; processing the first code snippet to obtain a first code attribute graph corresponding to the first code snippet; inputting the first code attribute graph into a code auditingmodel, wherein the code auditing model is a machine learning model obtained by training based on N second code attribute graphs corresponding to N second code snippets and N third code attribute graphs corresponding to N third code snippets; wherein the second code snippets are code snippets with bugs, and the third code snippets are code snippets obtained after bugs in the second code snippets are repaired; and obtaining an output of the code auditing model to obtain a detection result of auditing the first code snippet. The invention further provides a training method and device of the codeauditing module, electronic equipment and a medium.

Description

technical field [0001] The present disclosure relates to the technical field of artificial intelligence, and more specifically, to a code audit method and device, a code audit model training method and training device, an electronic device, and a medium. Background technique [0002] In the past time, a large number of enterprises around the world have been attacked by cyber attacks, with a total loss of tens of billions of dollars. Nowadays, the network permeates all aspects of work and life, and network security is particularly important for enterprises. Code auditing is an analysis method aimed at discovering security vulnerabilities, program errors, and source code violations of program coding standards. Source code security audit is a crucial step in routine penetration testing services and software architecture security checks. [0003] Existing code audits mostly use a combination of source code scanning and manual analysis and confirmation. Code audit tools commonl...

AI Technical Summary

Problems solved by technology

[0004] In the process of conceiving the technical solution of the present disclosure, the inventor found the following defects in the prior art: although dynamic program analysis can achieve high precision, it is difficult to detect all abnormal code fragments; when performing static program analysis, due to Many code fragments with vulnerabilities are very different from normal code, and cannot be distinguished when they are not running. There will be many false negatives and false positives, resulting in low accuracy

Images