ExplainabilitySecurityAccountabilitycs.CR
Explainable Ponzi Schemes Detection on Ethereum
Letterio Galletta, Fabio Pinelli
January 12, 2023
ACM Symposium on Applied Computing
17 citations
2 influential
Key Takeaways
-
Ponzi schemes are a significant threat on Ethereum: The research identifies Ponzi schemes as "one of the most common scams" on blockchain, demonstrating that cryptocurrency platforms are actively being exploited for fraudulent financial schemes that deceive users.
-
Machine learning can help protect users: The researchers developed a classifier that outperforms previous detection methods, suggesting that automated tools using artificial intelligence could be deployed to identify fraudulent smart contracts before they harm investors.
-
Explainability is crucial for trust in detection systems: The study emphasizes using "eXplainable AI techniques" to understand why contracts are flagged as Ponzi schemes, rather than treating detection as a "black box"—this transparency is essential for users and regulators to trust automated fraud detection systems.
-
A small set of identifiable warning signs can reveal fraud: The research found that "a small and effective set of features" can reliably detect Ponzi schemes, meaning suspicious patterns in smart contract code are recognizable and don't require complex analysis, making detection more practical and accessible.
-
Public datasets are needed to combat blockchain fraud: By releasing a labelled dataset of 4,422 real-world contracts, the researchers address a critical gap in fraud detection research, enabling the broader community to develop better protective tools—highlighting how transparency and data sharing strengthen security.