What is Federated Learning?
TL;DR
A distributed learning approach where models are trained across multiple devices without centralizing data. A privacy-preserving AI technology.
Federated Learning: Definition & Explanation
Federated learning is a distributed training approach where learning occurs locally on each device or site, sharing only model updates (gradients or parameters) rather than centralizing raw data on a central server. Proposed by Google in 2016 and first applied to improve smartphone keyboard predictions, it enables AI model building with data that is difficult to share due to privacy or data sovereignty concerns — such as medical and financial data. It aligns well with data protection regulations like GDPR and is used for collaborative medical AI training across hospitals and joint AI development among multiple companies. Combined with differential privacy and secure aggregation, it provides even stronger privacy guarantees.