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

TL;DR

The paper addresses privacy and data locality challenges in AI by surveying federated learning with a strong emphasis on aggregation. It introduces a multilevel taxonomy of aggregation techniques organized around personalization, optimization, and robustness, and couples bibliometric mapping with systematic review to map recent advances. Through experiments comparing representative aggregation methods under IID and non-IID data, it provides empirical insights into how heterogeneity, scalability, and hyperparameters shape performance. The work also outlines concrete future directions, including FL-enabled large language models, 6G integration, digital twins, and federated meta-learning, offering structured guidelines for evaluating new proposals in real-world settings.

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.

Figures (25)

• Figure 1: High-Level Classification of Recent Research Advances in Federated Learning.
• Figure 2: FL Research Area Clustering
• Figure 3: Flowchart For Our Systematic Survey Methodology.
• Figure 4: Quantitative Analysis of Surveyed Literature.
• Figure 5: Our Survey Structure Organization.