Release of Financial Application Guidance for Multi‑Party Secure Computation and Federated Learning

The Beijing Financial Technology Industry Alliance’s Data Committee organized a research effort on March 29, releasing two full‑text reports: “Multi‑Party Secure Computation Financial Application Status and Implementation Guide” and “Federated Learning Technology Financial Application White Paper,” following an earlier “Privacy Computing Technology Financial Application Research Report.”

Both reports provide detailed analyses of Multi‑Party Secure Computation (MPC) and Federated Learning, outlining relevant policies and standards, summarizing current financial industry use cases, and offering suggestions for platform inter‑connectivity and cross‑institution collaboration.

Ant Group, as a primary contributor, showcased its trusted privacy‑computing framework “YinYu” and the AntChain MORSE product, supplying practical examples and recommendations based on its implementations.

“YinYu” is Ant Group’s self‑developed trusted privacy framework featuring core design principles of security and openness, and includes virtual devices for MPC, TEE, homomorphic encryption, as well as rich federated‑learning algorithms and differential‑privacy mechanisms. AntChain MORSE is a large‑scale commercial privacy‑computing product that has won top honors in iDash competitions for homomorphic encryption and federated learning.

A highlighted case study between Ant Group and Shanghai Pudong Development Bank demonstrates how MPC‑based joint risk‑control modeling enables multiple financial institutions to exchange risk data, collaboratively train models, and make decisions while keeping raw data on‑premises, thereby enhancing model precision and reducing fraud risk.

The reports conclude that MPC‑based joint risk‑control preserves data locality, ensures secure data sharing across institutions, enables “data‑at‑rest, model‑in‑motion” online modeling, improves predictive performance, and mitigates business risk.

In the federated‑learning white paper, Ant Group proposes security recommendations for financial applications, addressing attacks such as protocol deviation, collusion, and differential attacks, and stresses the need for provable security that prevents participants from accessing private data.