Glossary
Federated Learning
A machine learning approach where model training occurs across distributed data sources without centralizing the data, preserving privacy.
What is Federated Learning?
Federated learning enables training AI agents on sensitive data that cannot be centralized due to privacy, security, or regulatory constraints. Local devices or organizations train on their own data, then share only model updates (gradients) with a central server that aggregates improvements. Raw data never leaves its source, protecting privacy while still enabling collaborative learning across many data sources.
This approach is valuable for healthcare agents learning from patient data across hospitals, financial agents training on customer data across banks, or mobile agents improving from user interactions across devices. Challenges include communication efficiency, handling non-uniform data distributions, ensuring security of shared gradients (which can leak information), and coordinating training across unreliable participants.
Example
Ten hospitals train a diagnostic agent on their patient data without sharing records. Each hospital's local model trains for several epochs, computes parameter updates, and sends encrypted gradients to a central server. The server aggregates updates to improve the global model, then redistributes it for the next training round.
How Signet addresses this
Signet's Security dimension values privacy-preserving training methods like federated learning. Agents trained using federated approaches on sensitive data demonstrate stronger data protection practices, improving security scores compared to agents requiring data centralization.
Build trust into your agents
Register your agents with Signet to receive a permanent identity and trust score.