Securing Blockchain Using Propagation Chain Learning

Smart contract vulnerabilities are the most common and severe type of blockchain vulnerability, which may result in very serious economic and property losses. Vulnerability detection and repair are necessary to ensure the security of the blockchain. Currently, the-state-of-art smart contract vulnerability detection methods (e.g. Oyente and Securify) use heuristics based on human-designed algorithms, which have certain shortcomings in different application scenarios. Therefore, this paper proposes a smart contract vulnerability detection method, i.e. CuVuD, which uses Propagation Chain Learning to solve the current vulnerability detection problem. This method first parses the source code, then obtains and trims the propagation chain of smart contracts, and finally detects vulnerabilities in smart contracts. To verify the effectiveness of CuVuD, this paper compares the CuVuD method with seven the-state-of-art smart contract vulnerability detection methods on a large-scale smart contract dataset based on the Solidity language. The experimental results show that CuVuD’s effectiveness in detecting smart contract vulnerabilities is significantly higher than seven the-state-of-art smart contract vulnerability detection methods, significantly improving the ability to detect vulnerabilities.