Coinbase upgrades anti-fraud system: Integrating machine learning and rule engines to reduce response time to just hours

Odaily Planet Daily News Coinbase announced that it is optimizing the rule creation process in its anti-fraud system by integrating machine learning models with a rule engine to achieve more efficient risk management. It also proposed a dual-track strategy of "models for long-term defense, rules for rapid response" and built a unified framework to create a feedback loop between the two: rules are used to capture new types of fraudulent behavior and provide feedback to train the model, thereby continuously improving overall defense capabilities.

In terms of specific optimizations, Coinbase has transformed the previously labor-intensive rule creation process into a data-driven and automatically recommended one by restructuring data, automating schema evolution, and introducing Notebook-based analysis tools, significantly improving efficiency. Specifically, rule backtesting performance has improved by more than 10 times, and overall response time has been reduced from days to just hours. Additionally, the new system recommends parameters through machine learning, which helps reduce the false positive rate, effectively combating fraud while minimizing the impact on normal users.

Coinbase stated that the next step will be to promote event-driven automatic rule generation and explore the "one-click conversion" of efficient rules into model features, further moving towards an automated risk management system.