A Comprehensive Empirical Study of Heterogeneity in Federated Learning

Federated learning (FL) is becoming a popular paradigm for collaborative learning over distributed, private datasets owned by non-trusting entities. FL has seen successful deployment in production environments, and it has been adopted in services such as virtual keyboards, auto-completion, item recommendation, and several IoT applications. However, FL comes with the challenge of performing training over largely heterogeneous datasets, devices, and networks that are out of the control of the centralized FL server. Motivated by this inherent challenge, we aim to empirically characterize the impact of device and behavioral heterogeneity on the trained model. We conduct an extensive empirical study spanning nearly 1.5K unique configurations on five popular FL benchmarks. Our analysis shows that these sources of heterogeneity have a major impact on both model quality and fairness, causing up to 4.6× and 2.2× degradation in the quality and fairness, respectively, thus shedding light on the importance of considering heterogeneity in FL system design.

Abdelmoniem, A. M., Ho, C.-Y., Papageorgiou, P., & Canini, M. (2023). A Comprehensive Empirical Study of Heterogeneity in Federated Learning. IEEE Internet of Things Journal, 1–1. https://doi.org/10.1109/jiot.2023.3250275

The work was conducted in part while Ahmed was with KAUST, KSA and while Pantelis was on an internship at KAUST, KSA We thank Muhammad Bilal for his help during the execution of the work. This publication is based upon work supported by King Abdullah University of Science and Technology (KAUST) under Award No. ORA-CRG10-2021-4699.

Institute of Electrical and Electronics Engineers (IEEE)

IEEE Internet of Things Journal


Additional Links

Permanent link to this record