Dr Marcin Bernaś presented the latest developments of the FORCE project at the UBBKryptoday 2026 conference, held at the University of Bielsko-Biała, Poland. During his session, titled “AI vs Financial Anomalies in DeFi: How Hidden Money Laundering Patterns are Detected”, Dr Bernaś showcased the technical progress achieved within Work Package 2: Blockchain Analytics & AI for AML.
A key focus of the presentation was a comprehensive comparative analysis of state-of-the-art machine learning approaches for AML detection in cryptocurrency transactions. The study covered a range of model families, including Graph Neural Networks, transformer-based architectures, classical machine learning models (such as Random Forest and XGBoost), as well as hybrid approaches combining multiple methodologies. The evaluation was conducted on several benchmark and real-world datasets, including the Elliptic Dataset (as the primary benchmark), the extended Elliptic++ dataset, synthetic datasets designed for controlled experimentation, and real or anonymized multi-chain data spanning Bitcoin, and Ethereum.
In addition, Dr Bernaś demonstrated an early-stage MVP of an anomaly radar system capable of real-time detection of suspicious financial patterns.
