aiqbench analysis
FedImp addresses the critical non-IID data distribution problem in federated learning by introducing impurity-based weighting to optimize global model aggregation. Experimental results on standard benchmarks demonstrate that this approach significantly accelerates convergence and improves accuracy under highly imbalanced conditions compared to existing methods like FedAvg. This advancement offers a practical pathway for deploying privacy-preserving machine learning systems across heterogeneous device networks with reduced communication overhead.
FedImp: Enhancing Federated Learning Convergence with Impurity-Based Weighting
arXiv:2608.14654v1 Announce Type: new Abstract: Federated Learning (FL) is a collaborative paradigm that enables multiple devices to train a global model while preserving local data privacy. A major challenge in FL is the non-Independent and Identically Distributed (non-IID) nature of data across devices, which hinders training efficiency and slows convergence. To tackle this, we propose Federated Impurity Weighting (FedImp), a novel algorithm that quantifies each device contribution based on the informational content of its local data. These contributions are normalized to compute distinct aggregation weights for the global model update. Extensive experiments on EMNIST and CIFAR-10 datasets show that FedImp significantly improves convergence speed, reducing communication rounds by up to 64.4%, 27.8%, and 66.7% on EMNIST, and 44.2%, 44%, and 25.6% on CIFAR-10 compared to FedAvg, FedProx, and FedAdp, respectively. Under highly imbalanced data distributions, FedImp outperforms all baselines and achieves the highest accuracy. Overall, FedImp offers an effective solution to enhance FL efficiency in non-IID settings.
Abstract:Federated Learning (FL) is a collaborative paradigm that enables multiple devices to train a global model while preserving local data privacy. A major challenge in FL is the non-Independent and Identically Distributed (non-IID) nature of data across devices, which hinders training efficiency and slows convergence. To tackle this, we propose Federated Impurity Weighting (FedImp), a novel algorithm that quantifies each device contribution based on the informational content of its local data. These contributions are normalized to compute distinct aggregation weights for the global model update. Extensive experiments on EMNIST and CIFAR-10 datasets show that FedImp significantly improves convergence speed, reducing communication rounds by up to 64.4%, 27.8%, and 66.7% on EMNIST, and 44.2%, 44%, and 25.6% on CIFAR-10 compared to FedAvg, FedProx, and FedAdp, respectively. Under highly imbalanced data distributions, FedImp outperforms all baselines and achieves the highest accuracy. Overall, FedImp offers an effective solution to enhance FL efficiency in non-IID settings.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2608.14654 [cs.LG]
(or arXiv:2608.14654v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.14654 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
IEEE Transactions on Artificial Intelligence, vol. 7, no. 3, pp. 1652-1665, March 2026
https://doi.org/10.1109/TAI.2025.3605307 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 272 words
Related coverage
Hybrid Quantum Networks Lift Fraud Detection Accuracy to 85%
Quantum Zeitgeist · Aug 4, 2026
Uncertainty-Aware Federated Learning for Joint Communication and Computation in XR Systems
arXiv · Oct 16, 2026
Polar Code Based Federated Learning: Convergence Analysis and Resource Allocation
arXiv · Aug 17, 2026
Efficient Asynchronous Federated Evaluation with Strategy Similarity Awareness for Intent-Based Networking in Industrial Internet of Things
arXiv · Jun 6, 2026