Quantization in Federated Learning: Methods, Challenges and Future Directions DOI: XXXXXXX.XXXXXXX Journal: JACM Volume: 37 4 111 8 CCS: Computing methodologies Distributed computing methodologies CCS: Computing methodologies CCS: Computing methodologies Distributed algorithms CCS: Federated Learning Quantization

Farwa Ikram, Dipanwita Thakur, Antonella Guzzo, Giancarlo Fortino · 2026

Read Quantization in Federated Learning: Methods, Challenges and Future Directions DOI: XXXXXXX.XXXXXXX Journal: JACM Volume: 37 4 111 8 CCS: Computing methodologies Distributed computing methodologies CCS: Computing methodologies CCS: Computing methodologies Distributed algorithms CCS: Federated Learning Quantization by Farwa Ikram, Dipanwita Thakur, Antonella Guzzo, Giancarlo Fortino (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-26822.txt.

Introduction

Farwa Ikram, Dipanwita Thakur, Antonella Guzzo, and Giancarlo Fortino review quantization for federated learning. Federated learning trains models across many devices without gathering their data. Its growth is throttled by communication costs, mixed device power, and data that differ from device to device. Quantization, shrinking numbers to fewer bits, cuts both traffic and on-device computing.

The review is the first built specifically around federated learning. Its taxonomy uses dimensions unique to the setting: device differences, aggregation consistency, scheduling, robustness to skewed data, privacy integration, and energy. The authors analyse how quantization interacts with client drift, partial participation, and secure aggregation. Cross-method insights and open problems close the survey. The aim is a map for builders who must compress without breaking the federation's guarantees.

Catalog record

CREATORFarwa Ikram, Dipanwita Thakur, Antonella Guzzo, Giancarlo Fortino
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE144 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.26822

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Quantization in Federated Learning: Methods, Challenges and Future Directions DOI: XXXXXXX.XXXXXXX Journal: JACM Volume: 37 4 111 8 CCS: Computing methodologies Distributed computing methodologies CCS: Computing methodologies CCS: Computing methodologies Distributed algorithms CCS: Federated Learning Quantization Farwa Ikram, Dipanwita Thakur, Antonella Guzzo, Giancarlo Fortino Abstract . Federated Learning (FL) has become a foundational paradigm for privacy-preserving distributed intelligence, yet its scalability remains fundamentally constrained by communication bottlenecks, device heterogeneity, and the challenges of training under statistically non-IID data. Quantization is one of the most effective mechanisms for mitigating these limitations, reducing both uplink/downlink payloads and on-device computation. This paper provides the first FL-centric systematic review of quantization, introducing a novel taxonomy organized around FL-specific dimensions, including client heterogeneity, aggregation consistency, communication-scheduling adaptation, non-IID robustness, privacy/security integration, and hardware/energy co-optimization. Beyond cataloging existing methods, we analyze how quantization interacts with core FL behaviors such as client drift, partial participation, convergence stability, secure aggregation, and differential privacy. We further identify cross-method insights, open research gaps, and design guidelines for practitioners deploying quantized FL on mobile, IoT, and edge platforms. This survey thus establishes quantization not merely as a compression technique, but as a fundamental systems component shaping the performance, robustness, and practicality of modern FL. 1. Introduction The rapid growth of Internet of Things (IoT) devices, mobile systems, and embedded sensors has enabled large-scale distributed intelligence at the network edge. While these devices generate vast amounts of data, transmitting raw data to centralized servers for model training raises serious concerns related to privacy, bandwidth, latency, and energy consumption. Federated Learning (FL) ( McMahan et al. 2017 ) addresses these challenges by enabling multiple clients to collaboratively train a shared global model without sharing local data. By keeping data localized, FL reduces communication overhead and preserves privacy, making it well suited for edge and mobile environments. Despite its advantages, deploying FL in real-world settings remains challenging due to device heterogeneity, limited computational resources, and high communication costs. In large-scale FL systems, the exchange of model updates between clients and the server often dominates training time and energy consumption, leading to latency, unstable convergence, and reduced participation of resource-constrained devices ( Zhao et al. 2022 ) . Consequently, improving communication efficiency is a central challenge in scalable FL. To address this issue, numerous communication-efficient techniques have been proposed, including client scheduling, sparsification, pruning, knowledge distillation, low-rank approximation, and quantization ( Konečný et al. 2017 ; Reisizadeh et al. 2020 ; Thakur et al. 2025b ) . However, many of these methods introduce additional overhead or complexity. For example, pruning requires transmitting masks, knowledge distillation demands additional training pipelines ( Wu et al. 2022 ) , and sparsification may degrade convergence under non-IID data. In contrast, quantization consistently reduces both communication and computation cost by lowering numerical precision, making it particularly attractive for resource-constrained and energy-limited FL deployments. Model quantization compresses model parameters, gradients, or activations into low-bit representations, enabling efficient communication and computation on edge devices. While quantization improves efficiency, it also introduces challenges such as accuracy degradation, convergence instability, and sensitivity to data heterogeneity ( Khan et al. 2021 ) . Understanding these trade-offs is therefore critical for designing robust and scalable FL systems. 1.1. Comparison with Related Survey and Motivation A large number of surveys have studied FL from diverse perspectives, including FL fundamentals ( Aledhari et al. 2020 ; Li et al. 2020b ; Liu et al. 2020 ; Liu et al. 2024 ; Lo et al. 2021 ; Ren et al. 2025 ; Wahab et al. 2021 ; Yang et al. 2019 ; Zhang et al. 2021 ) , mobile and edge FL ( Duan et al. 2023 ; Jia et al. 2025 ; Lim et al. 2020 ; Wu et al. 2024 ) , communication efficiency ( Almanifi et al. 2023 ; Liu et al. 2020 ; Niknam et al. 2020 ; Shahid et al. 2021 ; Zhao et al. 2023 ) , heterogeneity ( Pfeiffer et al. 2023 ; Ye et al. 2023 ) , privacy and security ( Chen et al. 2023c ; Khraisat et al. 2024 ; Kumar et al. 2024 ; Shao et al. 2024 ) , and green FL ( Thakur et al. 2025b ) . As summarized in Table 1 , it is evident that existing surveys on FL primarily concentrate on isolated challenges such as privacy and security, statistical and system heterogeneity, communication efficiency, computation efficiency, and energy efficiency. These surveys typically treat quantization as a subcomponent of communication efficiency or model compression. However, no existing survey provides a dedicated and systematic review of quantization techniques in FL. Prior works do not analyze quantization as a standalone design dimension, nor do they provide a comprehensive taxonomy of quantization strategies and their trade-offs in terms of accuracy, convergence, communication cost, and system heterogeneity. In contrast, this paper is, to the best of our knowledge, the first survey exclusively focused on quantization in FL, offering a unified and comparative synthesis across algorithmic, system, and security dimensions. Table 1 . Comparison of Existing Surveys Related to Federated Learning References Focus Area Key Contributions Limitations ( Aledhari et al. 2020 ; Li et al. 2020b ; Liu et al. 2022 ; Liu et al. 2024 ; Lo et al. 2021 ; Ren et al. 2025 ; Wahab et al. 2021 ; Yang et al. 2019 ; Zhang et al. 2021 ) FL Fundamentals Introduce FL concepts, architectures, and applications. Do not analyze communication efficiency or quantization. ( Duan et al. 2023 ; Jia et al. 2025 ; Lim et al. 2020 ; Liu et al. 2020 ; Niknam et al. 2020 ; Wu et al. 2024 ) FL in Wireless / Edge Networks Study FL integration with edge, 5G/6G, and wireless systems. Limited discussion of quantized model compression. ( Khan et al. 2021 ; Nguyen et al. 2021 ) FL for IoT Survey FL-enabled IoT services and applications. Lack in-depth analysis of communication and quantization. ( Qi et al. 2024 ; Sah and Singh 2022 ) Model Aggregation Review aggregation strategies and convergence issues. Communication compression and energy efficiency not addressed. ( Almanifi et al. 2023 ; Liu et al. 2020 ; Niknam et al. 2020 ; Shahid et al. 2021 ; Zhao et al. 2023 ) Communication-Efficient FL Review sparsification, quantization, scheduling, and client selection. Quantization analysis remains high-level. ( Pfeiffer et al. 2023 ; Ye et al. 2023 ) Heterogeneity-Aware FL Address statistical, device, and system heterogeneity. Energy efficiency and adaptive quantization underexplored. ( Chen et al. 2023c ; Shao et al. 2024 ; Kumar et al. 2024 ; Khraisat et al. 2024 ) Privacy and Security in FL Analyze privacy risks, attacks, and secure aggregation. Limited treatment of quantization–privacy interactions. ( Thakur et al. 2025b ) Green FL Study energy efficiency and carbon-aware FL. Quantization techniques not considered. Ours Quantization in FL Systematic survey and taxonomy of quantization techniques in FL. Addresses communication, robustness, heterogeneity, and privacy. As evidenced by the growing number of recent studies on quantized FL, there is a clear need for a systematic review that consolidates existing knowledge, identifies limitations, and outlines future research directions. This motivates the present PRISMA-based systematic literature review. Furthermore, the number of publications related

Read the complete text