Towards Heterogeneous Quantum Federated Learning: Challenges and Solutions

TL;DR

This work addresses the critical challenge of heterogeneity in quantum federated learning by classifying it into data and system categories and examining its impact on convergence and aggregation. It surveys existing mitigation strategies, identifies their limitations, and introduces SPQFL—a sporadic personalized QFL framework—as a case study to demonstrate robust performance under quantum noise and non-IID data. The proposed approach combines sporadic participation with personalization to improve stability and accuracy across diverse quantum devices, validated via simulations on standard datasets and realistic noise models. The paper also outlines open research directions, including scalable, noise-aware, and network-dynamics-aware QFL architectures for practical quantum networks.

Abstract

Quantum federated learning (QFL) combines quantum computing and federated learning to enable decentralized model training while maintaining data privacy. QFL can improve computational efficiency and scalability by taking advantage of quantum properties such as superposition and entanglement. However, existing QFL frameworks largely focus on homogeneity among quantum \textcolor{black}{clients, and they do not account} for real-world variances in quantum data distributions, encoding techniques, hardware noise levels, and computational capacity. These differences can create instability during training, slow convergence, and reduce overall model performance. In this paper, we conduct an in-depth examination of heterogeneity in QFL, classifying it into two categories: data or system heterogeneity. Then we investigate the influence of heterogeneity on training convergence and model aggregation. We critically evaluate existing mitigation solutions, highlight their limitations, and give a case study that demonstrates the viability of tackling quantum heterogeneity. Finally, we discuss potential future research areas for constructing robust and scalable heterogeneous QFL frameworks.

Figures (2)

• Figure 1: Proposed SPQFL architecture in which a set of distributed quantum devices collaborate with a quantum server to train a shared QML model.
• Figure 2: Comparison of SPQFL with other state-of-the-art approaches across datasets. Top row: classification accuracy; bottom row: cross-entropy loss.