Overview
Direct Answer
Federated learning is a machine learning paradigm where model training occurs across distributed devices or organisations without centralising raw data. Participants compute gradients locally and share only model updates with a central aggregator, enabling collaborative model improvement while maintaining data privacy and sovereignty.
How It Works
Local devices train a shared model architecture on their private datasets, computing weight updates independently. These updates are transmitted to a central server, which aggregates contributions using algorithms such as FederatedAveraging to produce an improved global model. The refined model is distributed back to participants for subsequent training rounds, iterating until convergence.
Why It Matters
Organisations adopt this approach to comply with data protection regulations (GDPR, HIPAA) whilst leveraging distributed datasets for model improvement. It reduces data transmission costs, latency, and breach risk by eliminating centralised data repositories, making it valuable for healthcare, finance, and telecommunications sectors handling sensitive information.
Common Applications
Healthcare systems use it to train diagnostic models across hospital networks without sharing patient records. Mobile device manufacturers optimise on-device models using user interaction data. Financial institutions collaboratively develop fraud detection systems whilst maintaining confidentiality of transaction data.
Key Considerations
Communication overhead between devices and server significantly exceeds traditional centralised training, and statistical heterogeneity across decentralised datasets can degrade model convergence. Debugging and monitoring distributed systems presents additional operational complexity.
Cross-References(1)
Cited Across coldai.org3 pages mention Federated Learning
Industry pages, services, technologies, capabilities, case studies and insights on coldai.org that reference Federated Learning — providing applied context for how the concept is used in client engagements.
Referenced By1 term mentions Federated Learning
Other entries in the wiki whose definition references Federated Learning — useful for understanding how this concept connects across Artificial Intelligence and adjacent domains.
More in Artificial Intelligence
AI Chip
Infrastructure & OperationsA semiconductor designed specifically for AI and machine learning computations, optimised for parallel processing and matrix operations.
AI Democratisation
Infrastructure & OperationsThe movement to make AI tools, knowledge, and resources accessible to non-experts and organisations of all sizes.
Perplexity
Evaluation & MetricsA measurement of how well a probability model predicts a sample, commonly used to evaluate language model performance.
Zero-Shot Prompting
Prompting & InteractionQuerying a language model to perform a task it was not explicitly trained on, without providing any examples in the prompt.
AI Red Teaming
Safety & GovernanceThe systematic adversarial testing of AI systems to identify vulnerabilities, failure modes, harmful outputs, and safety risks before deployment.
AI Accelerator
Infrastructure & OperationsSpecialised hardware designed to speed up AI computations, including GPUs, TPUs, and custom AI chips.
Chain-of-Thought Prompting
Prompting & InteractionA prompting technique that encourages language models to break down reasoning into intermediate steps before providing an answer.
AI Tokenomics
Infrastructure & OperationsThe economic model governing the pricing and allocation of computational resources for AI inference, including per-token billing, rate limiting, and credit systems.