Byte code level intelligent contract vulnerability detection method and system based on BERT extension

Through the BERT-based BERT extension, the problem of accurate and low detection methods is solved, and a more efficient smart contract vulnerability detection effect is achieved.

CN119720216APending Publication Date: 2025-03-28HAINAN UNIV
0 Cites 0 Cited by

Patent Information

Application Number
CN202411776205.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-05
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

Traditional smart contract vulnerability detection methods are limited in the effectiveness of bytecode, resulting in the problem of low accuracy of vulnerability detection.

Method used

Bytecode-level smart contract vulnerability detection method based on BERT extension is adopted, by collecting bytecode-level smart contracts, converting them into opcodes, performing semantic extraction, building control flow diagrams, finding vulnerability locations, extracting vulnerability fragments, and inputting them into the BERT model for feature extraction and feature fusion, and generating vulnerability detection reports.

Benefits of technology

It improves the accuracy and effectiveness of smart contract vulnerability detection and improves the detection effect of smart contract vulnerabilities.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to a byte code level intelligent contract vulnerability detection method and system based on BERT extension. The method comprises the following steps: collecting a byte code level smart contract, converting a byte code in the byte code level smart contract into an operation code, and performing semantic extraction to obtain a semantic feature; dividing the operation code to obtain each independent basic block, and constructing a control flow diagram of a byte code level based on each basic block and the jump relation; and extracting vulnerability segments in the control flow graph, inputting the vulnerability segments into a BERT model for feature extraction, performing feature fusion on the extracted features and semantic features to obtain fused full features, and completing a vulnerability detection task. According to the technical scheme, the independent basic blocks are obtained by dividing the operation codes, the control flow graph is constructed based on the basic blocks and the jump relation, and the control flow graph contains rich intelligent contract service logics, so that intelligent contract detection work based on BERT is more effectively realized, and the detection effect on intelligent contract vulnerabilities is improved.
Need to check novelty before this filing date? Find Prior Art