Beyond Weights and Gradients: A Taxonomy of Federated Learning Messages
2026-06-15 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors explain that in federated learning, the information shared between devices is becoming more varied than just model weights or gradients. They create a clear definition and a classification system for these different types of data exchanges, including model structures, statistical summaries, and data-conditioned representations. By examining many recent studies, they show how this variety affects computing needs, communication, and privacy risks. Their work helps clarify the trade-offs involved and offers a framework for improving federated learning systems in the future.
Federated LearningModel WeightsGradientsSynthetic DataFederated AnalyticsPrivacyDecentralized TrainingCommunication CostStatistical SummariesData-Conditioned Representations
Authors
Alvaro Javier Vargas Guerrero, Xinguang Wang, Quang Manh Doan, Guy Nagels
Abstract
Federated Learning is rapidly evolving beyond the exchange of traditional model weights and gradients, yet existing definitions fail to capture the full scope of modern payloads like synthetic data and federated analytics. This paper addresses the gap by proposing a formal mathematical definition of a federated message that accounts for both utility and privacy. We introduce a taxonomy that organizes these exchanges into three categories: model structures, statistical summaries, and data-conditioned representations. By evaluating these groups based on computational demands, communication costs, and privacy risks, we provide a clearer understanding of the trade-offs involved in decentralized training. Our review of 202 recent publications highlights a significant shift since 2021 toward diverse messaging paradigms, signaling a move away from standard deep learning updates toward more specialized information sharing. This framework provides a structured path for future research to optimize federated systems for varying hardware and security requirements.