Federated Learning Survey: A Multi-Level Taxonomy of Aggregation Techniques, Experimental Insights, and Future Frontiers

TL;DR

This survey provides a comprehensive overview of federated learning, categorizing aggregation techniques, analyzing experimental results under various data distributions, and outlining future research directions to address heterogeneity, efficiency, security, and privacy challenges.

Contribution

It offers a structured taxonomy of aggregation strategies in federated learning, combining bibliometric analysis with systematic review, and presents experimental insights and future research directions.

Findings

Aggregation strategies vary across architectures and synchronization methods.

Performance differs significantly between IID and non-IID data distributions.

Identifies key challenges and promising directions for future FL research.

Abstract

The integration of IoT and AI has unlocked innovation across industries, but growing privacy concerns and data isolation hinder progress. Traditional centralized ML struggles to overcome these challenges, which has led to the rise of Federated Learning (FL), a decentralized paradigm that enables collaborative model training without sharing local raw data. FL ensures data privacy, reduces communication overhead, and supports scalability, yet its heterogeneity adds complexity compared to centralized approaches. This survey focuses on three main FL research directions: personalization, optimization, and robustness, offering a structured classification through a hybrid methodology that combines bibliometric analysis with systematic review to identify the most influential works. We examine challenges and techniques related to heterogeneity, efficiency, security, and privacy, and provide a comprehensive overview of aggregation strategies, including architectures, synchronization methods, and diverse federation objectives. To complement this, we discuss practical evaluation approaches and present experiments comparing aggregation methods under IID and non-IID data distributions. Finally, we outline promising research directions to advance FL, aiming to guide future innovation in this rapidly evolving field.