Financial institutions face increasingly sophisticated cyber threats, but collaboration on stronger detection models is constrained by privacy, regulation, and the risks of centralized data sharing. This keynote explores how federated learning enables cross-institutional threat detection by allowing organizations to train shared machine learning models without exchanging raw security, customer, or transaction data.

The session presents federated learning as more than a privacy technique. It is a resilience enabler that can improve detection of fraud, account compromise, insider threats, anomalous behavior, and emerging attack campaigns while preserving data sovereignty. It also connects this approach to DevOps and platform engineering practices such as secure model delivery, continuous validation, observability, drift monitoring, incident response integration, and policy-driven governance.

A practical framework will be outlined for deploying federated cyber-defense in regulated financial environments, addressing secure aggregation, explainability, operational telemetry, resilience testing, failure isolation, and governance. The central message is that the future of financial cybersecurity lies in collaborative, privacy-preserving, and operationally reliable AI.

https://icccnet.co.uk/InvitedSpeakers/2026