Quantum federated learning (QFL) has emerged as a promising approach for collaboratively training compact quantum neural networks (QNNs) over distributed private data on resource-constrained devices. However, differences in device capabilities make a single shared QNN architecture unsuitable for all clients. While personalized quantum neural architecture search (QNAS) allows each client to select a device-specific QNN, averaging parameters across structurally different QNN architectures mixes semantically inconsistent circuit operations. To address this, prototype-guided personalized QNAS for virtual FL (vFedProtoQNAS) is proposed, where model parameters are never aggregated across clients and federated collaboration is achieved through class-wise prototype sharing. Each client independently searches and trains a client-specific QNN, computes class-wise local prototypes from latent representations, and refines them using global prototypes from the server as federated semantic anchors. Experiments demonstrate that vFedProtoQNAS improves accuracy by 3.70% over FedAvg and enhances class-consistent representation alignment.
Figures & tables
Figure 1: The concept of virtual federated learning.
Figure 2: The overall framework of vFedProtoQNAS.
Figure 3: Prototype margin improvement before and after prototype-guided refinement. A positive margin indicates that a latent representation is closer to the correct class prototype than to any incorrect prototype.
βproto
Accuracy
Precision
Recall
F1-score
0.05
91.25±0.53
91.30±0.52
91.25±0.53
91.23±0.54
0.10
91.25±0.74
91.29±0.71
91.25±0.74
91.24±0.74
0.20
91.19±0.66
91.23±0.66
91.19±0.66
91.18±0.68
0.50
90.94±0.38
90.98±0.35
90.94±0.38
90.93±0.37
Table 1: Sensitivity analysis of vFedProtoQNAS with respect to the prototype alignment weight βproto using MNIST(0-3). Results are reported as mean ± std over five seeds.
α
Accuracy
Precision
Recall
F1-score
0.5
84.32±11.77
82.67±16.72
84.32±11.77
82.64±15.48
1.0
91.15±0.76
91.26±0.67
91.15±0.76
91.15±0.75
10.0
91.16±0.93
91.24±0.89
91.16±0.93
91.17±0.92
Table 2: Sensitivity analysis of vFedProtoQNAS under different non-IID levels using MNIST(0-3). Results are reported as mean ± std over five seeds. A smaller Dirichlet concentration parameter α indicates stronger data heterogeneity.
Method
Accuracy
Precision
Recall
F1-score
FedAvg
87.55±2.38
87.96±2.26
87.55±2.38
87.46±2.39
FixedQNN LocalOnly
90.31±1.01
90.54±0.88
90.31±1.01
90.25±1.04
vFedProtoQNAS w/o QNAS
90.01±0.76
90.21±0.71
90.01±0.76
89.92±0.83
RandomQNAS + FedProto
90.17±0.92
90.32±0.87
90.17±0.92
90.15±0.90
PersonalQNAS
90.45±0.70
90.59±0.66
90.45±0.70
90.44±0.71
vFedProtoQNAS (Ours)
91.25±0.74
91.29±0.71
91.25±0.74
91.24±0.74
Table 3: Comparison with baselines under the default setting βproto=0.10 . Results are reported as mean ± std over five seeds.
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.
Hoang M. Ngo, Quan Nguyen, Wanli Xing +1
Department of Computer & Information Science & Engineering University of Florida · Frost Institute for Data Science and Computing University of Miami
Quantum federated learning enables distributed clients to train quantum neural networks without sharing local data, making it promising for privacy-aware intelligent services. Intelligent services in this context refer to privacy-sensitive distributed decision systems, such as fraud detection and genomic classification, where reliable and fair client-level learning is as important as the accuracy of the aggregate model. However, heterogeneous client data and noisy quantum optimization often cause unstable local updates, client drift, and unfair performance between clients. This paper proposes DUQFL-Prox, a drift-stable quantum federated learning framework based on deep-unfolded local optimization. Instead of using a fixed local optimizer, each client performs adaptive unfolded SPSA updates, while a proximal term keeps the local model close to the global model. A lightweight controller learns step-specific optimization parameters to improve post-aggregation performance. Experiments on financial fraud and genomic classification tasks show that DUQFL-Prox improves stability, generalization, and client fairness compared with standard QFL baselines. The results suggest that deep-unfolded quantum federated learning can support more reliable and fair intelligent services in heterogeneous distributed environments.
Shanika Iroshi Nanayakkara, Shiva Raj Pokhrel
School of IT, Deakin University, VIC 3125, Burwood, Australia
Federated learning (FL) enables collaborative model training across distributed devices without sharing raw data, making it suitable for privacy-sensitive robotic sensing applications. However, multi-agent systems generate heterogeneous and non-independent and identically distributed (non-IID) multimodal sensor streams that degrade conventional FL algorithms, while classical fusion modules introduce substantial parameter overhead and communication cost. This paper proposes QFedAgent, a hybrid quantum-classical personalized FL framework for multi-agent activity recognition. The approach integrates a variational quantum circuit fusion module that models accelerometer--gyroscope interactions through quantum state encoding and entanglement, requiring only 72 quantum rotation parameters versus 33K in classical multi-layer perceptron-based fusion, achieving approximately 10x total parameter reduction. Experiments on the OPPORTUNITY dataset under subject-based non-IID partitions demonstrate 97.7% mean test accuracy, confirming that parameter-efficient quantum fusion remains competitive with conventional federated baselines.
Quoc Bao Phan, Tuy Tan Nguyen
Department of Electrical and Computer Engineering FAMU-FSU College of Engineering, Florida State University Tallahassee, FL 32310, USA