Online security detection method and device for low-label-dependence block chain smart contract

By adopting a multivariate timing encoding mechanism based on behavior perception and a smart contract APT attack noise addition strategy in the online security detection of smart contracts, combined with the adaptive learning and offline semi-supervised training of the smart contract-guided attack noise denoising model, the problem of dependence on attack labeling data and difficulty in capturing APT strategies in the existing technology is solved, and efficient online security detection of smart contracts is achieved.

CN119722066APending Publication Date: 2025-03-28西交网络空间安全研究院 +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202411840412.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-13
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

Existing smart contract online security detection solutions rely heavily on attack labeled data and are difficult to capture the abnormal behavior characteristics of advanced sustainable attack (APT) strategies in dynamic blockchain environments.

Method used

By designing a multivariate timing encoding mechanism based on behavior perception, the behavior status of smart contracts is dynamically monitored, and an attack noise sequence is generated using the smart contract APT attack noise plus strategy. At the same time, a smart contract-guided attack noise denoising model is used to adaptively learn the noise expression characteristics, and the model is trained through offline semi-supervised training strategies to achieve online security detection.

Benefits of technology

Reduced dependence on scarce attack labeled data, improved the robustness and adaptability of online security detection of smart contracts, and was able to efficiently detect APT attacks in a dynamic blockchain environment.

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

Abstract

The invention discloses a low-label-dependence block chain smart contract online security detection method and device, and the method comprises the steps: starting from a real-time transaction sequence of a block chain smart contract, and for the transaction behavior characteristics of the smart contract, dynamically monitoring the behavior state of the smart contract through employing a designed multi-element time sequence coding mechanism based on behavior perception; for the normal behavior state of the smart contract, generating a smart contract APT attack noise sequence by using a designed smart contract APT attack noise adding strategy; in different block chain service scenes, adaptively learning noise expression features of an APT attack noise sequence of the smart contract by using the designed smart contract guided attack noise denoising model, and training the smart contract guided attack noise denoising model by using an offline semi-supervised training strategy; and performing online security detection on the block chain smart contract by using the trained smart contract guide attack noise denoising model. The objective of the invention is to improve robustness and adaptivity of intelligent contract online security detection, reduce dependence on scarce attack annotation data, effectively mine evolution characteristics of attack transaction behaviors, and realize efficient detection of intelligent contract APT attacks in a dynamic and low-annotation block chain environment.
Need to check novelty before this filing date? Find Prior Art