Quantum federated learning (QFL) enables collaborative training of variational quantum circuits across decentralized clients, but client drift under non-IID data distributions degrades standard Federated Averaging (FedAvg). Existing aggregation methods either introduce tunable hyperparameters that depend on unknown data heterogeneity or incur structural costs such as order-dependent error propagation. We propose Kernel-Weighted Aggregation (KWA), a hyperparameter-free aggregation strategy that constructs a non-parametric kernel density estimate over the received client parameter vectors at each communication round. Clients in high-density regions receive greater voting power, and isolated clients are automatically attenuated. The kernel bandwidth is the median of all pairwise squared distances. We evaluate KWA against six state-of-the-art QFL methods and five classical robust baselines on five binary MNIST digit-pair tasks under IID, Dir(0.5), and Dir(0.1) distributions with a unified four-qubit benchmark. Under IID data, KWA recovers the performance of FedAvg. Under Dir(0.5), KWA attains the highest average accuracy (0.7120), followed by FedAvg (0.7093). The margin at N=4 clients is not statistically significant. It grows to +0.023 at N=16 in the client scaling experiment. Under this distribution, the five classical robust baselines also rank below FedAvg. Under Dir(0.1), KWA exceeds FedAvg by 0.028 on average, with a pooled paired t-test p=0.068. Larger-circuit and three-class experiments show that the advantage does not yet transfer consistently beyond the four-qubit binary setting. KWA thus provides a hyperparameter-free aggregation strategy that recovers FedAvg under IID conditions, adapts to client drift, and requires no manually tuned hyperparameters.
Quantum federated learning (QFL) enables collaborative training of variational quantum circuits across decentralized clients, but client drift under non-IID data distributions degrades standard Federated Averaging (FedAvg). Existing aggregation methods either introduce tunable hyperparameters that depend on unknown data heterogeneity or incur structural costs such as order-dependent error propagation. We propose Kernel-Weighted Aggregation (KWA), a hyperparameter-free aggregation strategy that constructs a non-parametric kernel density estimate over the received client parameter vectors at each communication round. Clients in high-density regions receive greater voting power, and isolated clients are automatically attenuated. The kernel bandwidth is the median of all pairwise squared distances. We evaluate KWA against six state-of-the-art QFL methods and five classical robust baselines on five binary MNIST digit-pair tasks under IID, Dir(0.5), and Dir(0.1) distributions with a unified four-qubit benchmark. Under IID data, KWA recovers the performance of FedAvg. Under Dir(0.5), KWA attains the highest average accuracy (0.7120), followed by FedAvg (0.7093). The margin at N=4 clients is not statistically significant. It grows to +0.023 at N=16 in the client scaling experiment. Under this distribution, the five classical robust baselines also rank below FedAvg. Under Dir(0.1), KWA exceeds FedAvg by 0.028 on average, with a pooled paired t-test p=0.068. Larger-circuit and three-class experiments show that the advantage does not yet transfer consistently beyond the four-qubit binary setting. KWA thus provides a hyperparameter-free aggregation strategy that recovers FedAvg under IID conditions, adapts to client drift, and requires no manually tuned hyperparameters.
Abstract This paper addresses Quantum Federated Learning (QFL), an adaptation of classical federated learning for quantum computing that enables collaboration among multiple quantum devices. A significant challenge in QFL arises from non-independent and identically distributed (non-IID) data, primarily due to inherent variations in quantum devices and environments, which substantially degrades performance. To mitigate this, we propose novel QFL algorithms that employ weighted averaging and clustering-optimized averaging strategies. Specifically, Quantum Weighted Federated Averaging (QW-FedAvg) is introduced, which utilizes a model aggregation strategy combining quantum weighted averaging with a global data-sharing mechanism to reduce the impact of non-IID data. Building on this, we further propose the Quantum Federated Clustering-Optimized Averaging (QF-COA) algorithm. QF-COA enhances performance by selecting client-side training parameters that demonstrate superior efficacy through a clustering method, thereby further minimizing the adverse effects of non-IID data on model performance. Our models were trained and evaluated using the MNIST and Fashion-MNIST datasets. Simulation results demonstrate that the proposed algorithms improve accuracy by 2% to 5% in non-IID environments compared to traditional quantum federated averaging algorithms. Additionally, these advancements lead to enhanced communication efficiency and increased robustness. These improvements contribute to the greater generalization and adaptability of QFL frameworks, offering both theoretical insights and practical guidance for future research in the field. The source code is available at https://github.com/wang-lab-repository/qavg-gzq.git .
Quantum Federated Learning (QFL) offers a promising framework to train quantum models across distributed clients while keeping data strictly local. Due to its simplicity and low communication overhead, Federated Averaging (FedAvg) is the standard aggregation choice in QFL literature. However, deploying QFL on practical hardware exposes a severe double-drift phenomenon: the global model is simultaneously derailed by client drift from non-IID data and hardware bias from noisy quantum gradient estimates. In this work, we first analyze the convergence of FedAvg under these realistic conditions, mathematically demonstrating that quantum hardware bias creates a persistent error floor that standard averaging cannot correct. To overcome this limitation, we propose Q-ANCHOR, a quantum-aware federated aggregation architecture that anchors server updates with zero-noise extrapolation while applying stateful client correction to suppress both client drift and hardware-induced bias. Our convergence theory proves that Q-ANCHOR successfully mitigates classical client drift while actively reducing the hardware-bias floor. Experimental results demonstrate that Q-ANCHOR achieves significantly more stable training than conventional FL baselines.
Federated learning enables collaborative model training across decentralized clients under privacy constraints. Quantum computing offers potential for alleviating computational and communication burdens in federated learning, yet hybrid classical-quantum federated learning remains susceptible to performance degradation under non-IID data. To address this, we propose FEDCOMPASS, a layered aggregation framework for hybrid classical-quantum federated learning. FEDCOMPASS employs spectral clustering to group clients by class distribution similarity and performs cluster-wise aggregation for classical feature extractors. For quantum parameters, it uses circular mean aggregation combined with adaptive optimization to ensure stable global updates. Experiments on three benchmark datasets show that FEDCOMPASS improves test accuracy by up to 10.22% and enhances convergence stability under non-IID settings, outperforming six strong federated learning baselines.
Edge analytics requires continuous model adaptation under tight latency, bandwidth, and confidentiality constraints, yet conventional federated learning can leak sensitive information through shared updates and often struggles under non-IID client data. This study introduces Federated Quantum Machine Learning that transmits privacy-preserving measurement statistics derived from hybrid parameterized quantum circuits rather than raw samples or high-dimensional gradients. Across 50 edge clients over 20 rounds, the proposed method achieves 0.844 accuracy and 0.836 macro-F1 with ECE 0.048, outperforming a classical FedAvg baseline (0.801 accuracy, 0.788 macro-F1, ECE 0.062) while maintaining predictable edge overhead (845 ms average round latency) and compact messaging (3.1 KB average payload per round). Privacy–measurement sweeps demonstrate accuracy saturation with increasing privacy budget, improving from 0.816 at ε = 0.5 to 0.856 at ε = 8.0 under 512 shots, while calibration improves more sharply, with ECE decreasing from 0.068 to 0.046 over the same range. Increasing shot budget from 128 to 2048 consistently increases accuracy under all privacy settings, confirming that measurement variance is a controllable bottleneck for protected learning. Communication analysis shows that statistics-only messaging remains an order of magnitude smaller than gradient-style baselines (33.9 KB average payload per round), and secure aggregation adds bounded overhead (4.8 KB average payload) with a modest utility shift (0.836 accuracy, ECE 0.051). Robustness experiments under severe non-IID drift (severity 0.9) preserve macro-F1 at 0.801 with expectation-based statistics versus 0.772 for histogram-based statistics, and perturbed-client tests show improved stability with robust aggregation, sustaining 0.836 accuracy at 10% perturbed participation compared to 0.824 under naïve aggregation. These results indicate that low-dimensional, privatized measurement summaries enable practical, reliable, and privacy-aligned federated quantum learning on constrained edge infrastructures.
Federated Learning (FL) has become increasingly popular across different sectors, offering a way for clients to work together to train a global model without sharing sensitive data. It involves multiple rounds of communication between the global model and participating clients, which introduces several challenges like high communication costs, heterogeneous client data, prolonged processing times, and increased vulnerability to privacy threats. In recent years, the convergence of federated learning and parameterized quantum circuits has sparked significant research interest, with promising implications for fields such as healthcare and finance. By enabling decentralized training of quantum models, it allows clients or institutions to collaboratively enhance model performance and outcomes while preserving data privacy. Recognizing that Fisher information can quantify the amount of information that a quantum state carries under parameter changes, thereby providing insight into its geometric and statistical properties. We intend to leverage this property to address the aforementioned challenges. In this work, we propose a Quantum Federated Learning (QFL) algorithm that makes use of the Fisher information computed on local client models, with data distributed across heterogeneous partitions. This approach identifies the critical parameters that significantly influence the quantum model's performance, ensuring they are preserved during the aggregation process. Our research assessed the effectiveness and feasibility of QFL by comparing its performance against other variants, and exploring the benefits of incorporating Fisher information in QFL settings. Experimental results on ADNI and MNIST datasets demonstrate the effectiveness of our approach in achieving better performance and robustness against the quantum federated averaging method.
Abstract—Quantum federated learning (QFL) enables clients to train quantum neural network (QNN) models without sharing private data. We find that aggregation in QFL is unstable under heterogeneous data, unreliable communication, variable fidelity, latency, and quantum hardware noise. Moreover, QFL is non-trivial to address because several QNN parameters are periodic angles, and Euclidean averaging often fails to capture the inherent dynamics. We develop a novel self-consistent midpoint aggregation method for the stable design and implementation of QFLs. We combine QoS-aware client weighting, circular parameter aggregation, and bounded midpoint-based update control. We perform several angular tests and experiments on IBM's real Quantum machines for validation, confirming our approach. Extensive evaluations and experiments on medical and financial datasets show improved stability, lower volatility, and competitive accuracy. Index Terms—Quantum federated learning, adaptive aggregation, geometry-aware aggregation, self-consistent midpoint, quality of service, angular aggregation, quantum neural networks.