Federated Learning with Differential Privacy for Cross-Silo Financial Fraud Detection
Views: 0
Downloads: 0
Ahmad Firmansyah
1
,
Laila Saputra
2
1, 2Department of Public Policy, Universitas Padjadjaran
Downloads
Banking regulations prohibit sharing raw transactional logs across competing financial organizations. We design a federated learning framework incorporating differential privacy noise, enabling collaborative model convergence on rare fraud topologies while strictly guaranteeing data governance boundaries.
