John Nabwanda
Cybersecurity UDOMAI-Powered Blockchain Forensics System for Detecting Suspicious Cryptocurrency Transactions and Smart-Contract Threats
Problem Statement
A major challenge is that suspicious transactions may not always have obvious characteristics. Criminal activities can involve complex transaction chains, multiple wallets, rapidly changing addresses, unusual transaction amounts, and interactions with smart contracts.
Proposed Solution
The proposed solution is to develop an AI-powered blockchain forensics system that automatically analyses cryptocurrency transactions and identifies potentially suspicious activities. The system will use open-source pre-trained machine-learning models together with blockchain graph analysis to detect unusual transaction patterns, suspicious wallet relationships, and potential security threats.
Innovation
the project will specifically address class imbalance using techniques such as class weighting, appropriate sampling, and suitable evaluation metrics
Expected Impact
The project will be considered successful if the resulting prototype can reliably identify suspicious transaction patterns while minimizing false alarms