Background:
Digital financial systems are reshaping global transaction landscapes, especially with the rise of blockchain technology and cryptocurrencies. Despite the potential for innovation, these systems remain vulnerable to fraud and anomalous activities. As blockchain platforms gain traction, insufficient oversight has given rise to malicious transactions, making the need for robust detection methods urgent. This project, a collaboration between the University of Zurich (UZH) and Kyoto University, aims to tackle this challenge by focusing on anomaly detection within blockchain transaction networks. The goal is to develop AI-powered models capable of autonomously detecting fraudulent activities, thereby enhancing the security and integrity of blockchain-based financial systems. By combining technical innovations in machine learning with legal expertise, this collaborative effort will contribute to the creation of more secure and efficient digital financial ecosystems.
Research Question and Methodology:
The project is centered on developing novel methods for anomaly detection in blockchain transactions, addressing the increasing threat of fraud and irregular activities. The central research question is: How can machine learning models, such as Boltzmann machines, be leveraged to detect anomalies in blockchain transactions, and how can these models be integrated with smart contracts to automate fraud detection in real-world markets? The hypothesis posits that integrating anomaly detection models with smart contracts can create an automated system that not only detects but also mitigates fraudulent or abnormal transactions on blockchain platforms. The research follows a three-phase approach: Phase 1 focuses on feature extraction from blockchain transaction networks and price data using graph theory, topology, and statistical analysis; Phase 2 will develop the anomaly detection model using machine learning, particularly Boltzmann machines; and Phase 3 will test the model through smart contract deployment in live markets. Additionally, legal experts will assess the regulatory implications of the anomaly detection pipeline, particularly concerning compliance with the Regulation on the Traceability of Transfers of Funds (TFR).
Research Team:
This interdisciplinary project brings together experts from various fields:
- University of Zurich (UZH): The Department of Informatics and UZH Blockchain Center will lead the development of blockchain data analytics and anomaly detection models, utilizing network science and machine learning to identify fraudulent transactions on blockchain platforms.
- Kyoto University (KyotoU): The Graduate School of Advanced Integrated Studies in Human Survivability (GSAIS) will contribute advanced mathematical modeling techniques, focusing on the creation of scalable models for anomaly detection in blockchain transaction networks.
- University of Florence: Legal experts will provide insights into the regulatory and legal aspects of blockchain-based financial transactions, including compliance with the Regulation on the Traceability of Transfers of Funds (TFR).
This collaboration combines blockchain analytics, mathematical modeling, and legal analysis, forming a comprehensive solution aimed at securing crypto asset trading.