Leveraging AI and Machine Learning to Develop an Efficient Smart Contract Vulnerability Detection Method

In recent years, blockchain technology has become a source of new hope with its broad spectrum of applications, e.g., finance, health care, supply-chain management or intrusion detection, to name a few. The main attributes of blockchain technology are transparency, decentralization, reliable database, collective maintenance, trackability, security and credibility, digital cryptocurrency, and programmable (smart) contracts.

A smart contract is a self-executing program that runs on a blockchain with a unique address. It allows transactions to be carried out between different entities without a central authority and enables the code to execute autonomously upon meeting specified conditions. A smart contract can store information as internal state variables and define custom functions to manipulate or update its state. The operations in a smart contract are published as transactions.

In recent years, many vulnerabilities of smart contracts have been found. Hackers used these vulnerabilities to attack the corresponding contracts, and it has caused lots of economic losses. Thus, it is need of an hour to identify the potential problems of smart contracts. Existing methods of detecting vulnerabilities in smart contracts, such as Oyente and Securify, are based on symbolic execution. However, this technique is time-consuming, as it requires the exploration of all possible execution paths in a contract.

This thesis aims to develop an AI-powered and machine learning-enabled method to efficiently identify vulnerabilities in smart contracts. The work would involve building a dataset of smart contracts with source codes written in commonly used smart contract programming languages (e.g. Solidity, Javascript, Rust, etc.), identifying critical keywords, functions, or patterns within the code that are commonly associated with security issues, developing a machine learning model specifically tailored for identifying vulnerabilities in smart contracts, implementing robust evaluation metrics for measuring the effectiveness of the proposed vulnerability detection method and assessing the effectiveness of the proposed method by applying it to a diverse range of real-world smart contracts, including those from popular blockchain platforms.

How to apply

Formal applications can be made via the University of Bradford web site; applicants should register an account and select 'Full-time PhD in Engineering' as the course.