A Graph Neural Network Approach for Detecting Smart Contract Anomalies in Collaborative Economy Platforms Based on Blockchain Technology

Abstract

Blockchain technology provides a promising solution for collaborative economy systems by offering a decentralized, transparent, and secure platform. This is mainly accomplished through smart contracts, which are self-executing computer programs that facilitate, verify, and enforce the negotiation or performance of a contract. Digital tokens, on the other hand, are used to represent assets or currencies in these systems. Despite the benefits of Blockchain-based collaborative economy systems, significant security concerns are associated with them. These include the possibility of fraud, risk assessment, bugs in smart contracts, and cyber-attacks. For instance, attackers can exploit vulnerabilities in smart contracts to perform reentrancy and infinite loop attacks, leading to significant financial losses. To address these security challenges, this paper proposes integrating artificial intelligence models to prevent vulnerabilities in smart contracts and detect anomalies. Specifically, Graph Neural Networks models can be utilized to safeguard Blockchain-based collaborative economy platforms from attacks such as reentrancy and infinite loop attacks. According to the findings, this approach can accurately identify both normal and abnormal traffic and classify specific types of attacks. The framework's performance is further evaluated using various metrics to ensure its effectiveness in detecting anomalies, thereby providing an additional layer of security for Blockchain-based collaborative economy systems.