Researchers develop AI framework to combat DeFi fraud on Ethereum. A new ensemble-based detection system called SLEID has been proposed to identify illicit accounts in decentralized finance transactions. The framework uses machine learning techniques including Isolation Forest models and self-training mechanisms to detect fraudulent activity with minimal labeled data. Testing on nearly seven million Ethereum transactions showed the system achieved significantly higher accuracy than existing methods, with precision improvements of 2.56 percentage points and F1 score gains of 0.90 percentage points. This advancement addresses a critical vulnerability in the DeFi ecosystem as illicit accounts continue evolving their tactics to avoid detection.
Post from MarketNews_en
Log in to interact with content.