Federated AI

What is Federated AI?

Approach where models learn across separate data holders without centralizing the underlying data

In other languages

한국어연합 AI

데이터를 한곳에 모으지 않고 여러 기관이나 장치에서 모델을 함께 학습하는 방식이다

日本語連合型AI

データを各組織に置いたまま、複数拠点で協調してAIモデルを学習・運用する方式

Related Terms

• Hybrid inferenceA computing approach that distributes AI model execution across both local devices and remote cloud servers
• Model-agnosticA system design that can work with various different AI models (like GPT or Claude) without being restricted to a single provider.
• Co-trainingA machine learning technique where models are trained to leverage multiple sources or views of data to improve performance
• Model portabilityDesign principle that allows software systems to switch among AI models with limited rebuilding
• Model orchestrationTechnique for coordinating multiple AI models or agents to handle tasks based on suitability
• Model MergingA technique that combines the weights and capabilities of multiple specialized AI models into a single, unified model.
• Open weight modelsAI models whose trained parameters are publicly available for others to download, run or adapt.