Beyond Weights and Gradients: A Taxonomy of Federated Learning Messages

2026-06-15

AI Summary

Existing definitions of federated learning struggle to encompass the diverse forms of communication beyond model weights and gradients, such as synthetic data and statistical summaries. This work proposes a formal mathematical definition of “federated messages” and establishes a tripartite classification framework encompassing model architectures, statistical summaries, and data-conditioned representations. Through a systematic review of 202 publications and a multidimensional evaluation incorporating computational overhead, communication costs, and privacy risks, the study presents the first unified taxonomy that captures the heterogeneous communication paradigms in modern federated learning. It further reveals an emerging trend since 2021 toward diversified information sharing, offering a structured theoretical foundation and design trade-off guidance for optimizing federated systems under heterogeneous hardware constraints and varying security requirements.

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.

Problem

• Federated Learning
• federated messages
• privacy
• communication payloads
• taxonomy

Innovation

• federated learning
• message taxonomy
• privacy-utility trade-off
• data-conditioned representations
• federated analytics