Defect detection methods for smart contracts based on multimodality

Abstract

With the popularization of blockchain technology, the application of smart contracts in various scenarios is becoming increasingly widespread. However, due to its inherent complexity and dynamism, the security issues of smart contracts are gradually becoming prominent. In order to effectively detect and repair defects in smart contracts, this paper proposes a multimodal smart contract defect detection method. This method integrates multimodal data information, including visual, semantic, and inheritance relationship structural information. It combines the Transform method to comprehensively analyze and identify potential defects in the operational status of smart contracts. Compared with other mainstream methods, it has been proven that our method can more comprehensively identify and locate defects and improve detection accuracy and coverage.