Bybit与香港大学学生合作，通过一项实际应用中的数据脱敏挑战，共同开展反洗钱研究

라온

DUBAI, UAE, May 15, 2026 /PRNewswire/ -- Bybit, the world's second-largest cryptocurrency exchange by trading volume, recently concluded an anti-money laundering research collaboration with student teams from The University of Hong Kong (HKU), giving participants practical experience in cryptocurrency investigations, machine learning and anti-money laundering analysis through a real-world demixing challenge.

The collaboration centered on the February 2025 Bybit security breach. Using the incident as a case study, students explored how blockchain analytics and machine learning can be applied to identify laundering pathways associated with cryptocurrency mixer activity and transactions linked to the Lazarus Group.

The project was proposed and supervised by Prof. Doyeon Kim, assistant professor in Accounting and Law at The University of Hong Kong, with Bybit providing the real-world investigative context and industry guidance throughout the research process. This capstone was scoped and sponsored by David Zong, Head of group risk control and security at Bybit, to give students the opportunity to tackle a real-world industry challenge rather than a hypothetical exercise. Zong has closely followed the team's findings and progress throughout the project.

As part of the collaboration, students were tasked with tracing Lazarus-linked funds on the Bitcoin blockchain, understanding the role of wallet addresses used during laundering operations, and developing machine learning approaches capable of identifying mixer-related transaction patterns and potential output addresses connected to illicit activity. Given the privacy-preserving design of cryptocurrency mixers, tracing exact transaction flows is mathematically infeasible, meaning there is no absolute or deterministic method for attribution. The project therefore focused on probabilistic analysis, behavioral clustering and machine learning techniques designed to improve the identification of suspicious transaction patterns and laundering pathways.

Rather than following a fixed methodology, students were given an open project brief and encouraged to define their own analytical approach. Some teams expanded on existing blockchain tracing methods, while others explored alternative machine learning and graph-analysis models.

Over several weeks of research and iteration, student teams refined their methodologies through independent investigation, technical discussions and project reviews. Bybit also conducted a mid-term review session with participating teams, creating an opportunity for students to present progress updates, reflect on research findings and receive feedback to support the next stage of development.

The collaboration was designed to expose students to the uncertainty and problem-solving demands commonly associated with real-world anti-money laundering and blockchain compliance work. For students, the experience provided an opportunity to refine analytical thinking, develop technical frameworks and work through complex investigative challenges without predetermined solutions. For Bybit, the initiative supported engagement with emerging talent while allowing the company to contribute practical industry perspectives to academic research and early-stage technical exploration.

The final project review was held at the Bybit office, where students presented their findings and received additional feedback from the Bybit team.

As part of the research process, students analyzed approximately 49,800 Bitcoin blocks and more than 146 million transactions using clustering models, graph neural networks and graph-based transaction analysis techniques.

According to the project findings, the research identified 10,289 Wasabi-like transactions and generated a blockchain transaction subgraph containing more than 1.6 million address nodes and nearly 6 million transaction edges using Peel Chain methodology. Based on their sampling and testing, a machine learning cluster achieved a 70.5% recall rate against confirmed DPRK linked addresses.

The project additionally examined how cryptocurrency mixers such as Wasabi, CoinJoin transactions and peel chain structures are used to obscure transaction histories and complicate anti-money laundering investigations. Students studied how illicit funds move across wallets, decentralized exchanges, swap protocols and cross-chain systems designed to reduce transaction traceability.

Students involved in the collaboration said the project helped them better understand how blockchain investigations, anti-money laundering systems and cryptocurrency security operations function in practice.

One participant said the experience reinforced the importance of regulation and anti-money laundering technologies in increasing operational barriers for criminal actors while also demonstrating the complexity of solving practical industry problems.

Another student described the project as more informative and engaging than expected and said the learning outcomes exceeded initial expectations.

Several participants said the collaboration provided valuable insight into cryptocurrency investigations, blockchain security incidents and the increasing demand for talent in blockchain security and compliance-related fields.

One student also noted that the project contributed to a more systematic understanding of Bitcoin blockchain structures, transaction mechanisms and the logic behind on-chain money laundering, while discussions surrounding the future of blockchain-based financial systems provided useful perspectives related to future career development in finance and actuarial science.

The collaboration reflects growing interest among universities and digital asset companies in applying machine learning and blockchain analytics to real-world cybersecurity, compliance and financial crime challenges.

Bybit extends its appreciation to the HKU Business School and Professor Kim for their collaboration and support throughout the research initiative.

#Bybit / #NewFinancialPlatform

About Bybit

Bybit is the world's second-largest cryptocurrency exchange by trading volume, serving a global community of over 80 million users. Founded in 2018, Bybit is redefining openness in the decentralized world by creating a simpler, open and equal ecosystem for everyone. With a strong focus on Web3, Bybit partners strategically with leading blockchain protocols to provide robust infrastructure and drive on-chain innovation. Renowned for its secure custody, diverse marketplaces, intuitive user experience, and advanced blockchain tools, Bybit bridges the gap between TradFi and DeFi, empowering builders, creators, and enthusiasts to unlock the full potential of Web3. Discover the future of decentralized finance at Bybit.com.

For more details about Bybit, please visit Bybit Press

For media inquiries, please contact: media@bybit.com

For updates, please follow: Bybit's Communities and Social Media

Coinsori

Bybit与香港大学学生合作，通过一项实际应用中的数据脱敏挑战，共同开展反洗钱研究