Svmuu reported that Coinbase stated 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 where "models are responsible for long-term defense, and rules are responsible for rapid response," and built a unified framework that creates a feedback loop between the two: rules are used to capture new types of fraudulent behavior and then inversely train the model, thereby continuously enhancing overall defense capabilities.
In terms of specific optimizations, Coinbase has transformed the previously manual rule creation process into a data-driven and automatically recommended one by restructuring data architecture, automating Schema evolution, and introducing Notebook-based analysis tools, significantly improving efficiency. Among these improvements, the performance of rule backtesting has increased by more than 10 times, and the overall response time has been shortened from days to hours. Additionally, the new system recommends parameters through machine learning, which helps reduce false positive rates, minimizing the impact on legitimate users while combating fraud.
Coinbase stated that its next steps will be to advance event-driven automatic rule generation and explore the ability to "convert" efficient rules into model features with a single click, further moving towards an automated risk management system.