Privacy-preserving cross-bank AML detection via reputation-weighted federated learning

Multiple banks train one money-laundering detector together without sharing raw customer data. Each bank’s model update is privacy-protected, and more trustworthy participants get more influence on the shared model.