Intelligent Agents for Auction-based Federated Learning: A Survey

TL;DR

The paper addresses the challenge of incentive design in auction-based federated learning by surveying intelligent-agent approaches across data consumers, data owners, and the auctioneer. It proposes a unique multi-tier taxonomy organized by stakeholders, auction mechanisms, and agent goals, and analyzes limitations, evaluation metrics, and future directions. The work highlights methodological gaps, such as dynamic decision making, multi-agent interactions, privacy/security, online auctions, and explainability, that are essential for practical stakeholder-oriented decision support in IA-AFL ecosystems. Overall, the survey provides a structured roadmap to guide researchers entering IA-AFL and informs the design of robust, fair, and efficient AFL incentive mechanisms.

Abstract

Auction-based federated learning (AFL) is an important emerging category of FL incentive mechanism design, due to its ability to fairly and efficiently motivate high-quality data owners to join data consumers' (i.e., servers') FL training tasks. To enhance the efficiency in AFL decision support for stakeholders (i.e., data consumers, data owners, and the auctioneer), intelligent agent-based techniques have emerged. However, due to the highly interdisciplinary nature of this field and the lack of a comprehensive survey providing an accessible perspective, it is a challenge for researchers to enter and contribute to this field. This paper bridges this important gap by providing a first-of-its-kind survey on the Intelligent Agents for AFL (IA-AFL) literature. We propose a unique multi-tiered taxonomy that organises existing IA-AFL works according to 1) the stakeholders served, 2) the auction mechanism adopted, and 3) the goals of the agents, to provide readers with a multi-perspective view into this field. In addition, we analyse the limitations of existing approaches, summarise the commonly adopted performance evaluation metrics, and discuss promising future directions leading towards effective and efficient stakeholder-oriented decision support in IA-AFL ecosystems.

Figures (2)

• Figure 1: An overview of the AFL ecosystem.
• Figure 2: The IA-AFL taxonomy. DC, DO, SW and SC denote data consumer, data owner, social welfare and social cost, respectively.