Financial fraud detection is challenged by decentralized data, severe class imbalance, and privacy constraints. This paper presents QuantumChain, a secure Quantum Federated Learning (QFL) framework that combines hybrid quantum-classical neural networks, encrypted federated aggregation, blockchain-based auditability, and quantum-secure communication. Each client trains a local hybrid model in which a variational quantum circuit is embedded between classical neural layers, while model updates are protected through homomorphic encryption, threshold secret sharing, and QKD-based keying. A permissioned blockchain records aggregation events and supports reputation-weighted trust among participants. We evaluate QuantumChain on financial transaction data using a compact, size-matched classical baseline to isolate the effect of the quantum layer. Results show that the HQNN achieves comparable accuracy while improving fraud-class recall in most settings, reaching 94.6% recall compared with 93.2% for the classical model. The Deep QLayer improves performance in full-data settings, suggesting that added circuit depth helps recover representational capacity when the shallow circuit becomes limited. Mixed-state simulations further show that the recall trend persists under non-ideal quantum evolution. In federated deployment with 10 heterogeneous clients, global accuracy increases from 97.7% to 98.8% over five rounds before stabilizing. These results show that QuantumChain can integrate depth-aware hybrid quantum models into a secure federated fraud-detection pipeline while maintaining stable global convergence.
Comments
Accepted at the IEEE International Conference on Quantum Computing and Engineering (QCE), 2026
Financial fraud detection is challenged by decentralized data, severe class imbalance, and privacy constraints. This paper presents QuantumChain, a secure Quantum Federated Learning (QFL) framework that combines hybrid quantum-classical neural networks, encrypted federated aggregation, blockchain-based auditability, and quantum-secure communication. Each client trains a local hybrid model in which a variational quantum circuit is embedded between classical neural layers, while model updates are protected through homomorphic encryption, threshold secret sharing, and QKD-based keying. A permissioned blockchain records aggregation events and supports reputation-weighted trust among participants. We evaluate QuantumChain on financial transaction data using a compact, size-matched classical baseline to isolate the effect of the quantum layer. Results show that the HQNN achieves comparable accuracy while improving fraud-class recall in most settings, reaching 94.6% recall compared with 93.2% for the classical model. The Deep QLayer improves performance in full-data settings, suggesting that added circuit depth helps recover representational capacity when the shallow circuit becomes limited. Mixed-state simulations further show that the recall trend persists under non-ideal quantum evolution. In federated deployment with 10 heterogeneous clients, global accuracy increases from 97.7% to 98.8% over five rounds before stabilizing. These results show that QuantumChain can integrate depth-aware hybrid quantum models into a secure federated fraud-detection pipeline while maintaining stable global convergence.
Financial services organizations are caught in a paradoxical situation whereby the identification of complex and cross-border fraud cases requires the global visibility of data, and at the same time, increasingly rigid data-sovereignty laws, like the GDPR and CCPA, require sensitive data to be isolated locally. The existing Federated Learning (FL) technology provides a partial solution, but is heavily limited by the need to face the high dimensionality and non-identically-distributed (non-IID) nature of financial transaction data sets. This paper, in turn, proposes a new framework called QFL-Fin, which combines Variational Quantum Circuits (VQCs) and federated learning. QFL-Fin can convert transaction characteristics that are high-dimensional in nature into quantum states through the exploitation of the high-dimensional expressiveness of the Hilbert space to enable a more effective dissection of fraudulent patterns compared to classical deep neural networks.
Detecting fraudulent financial transactions is inherently challenging due to extreme class imbalance, complex high-dimensional feature spaces, and the adaptive strategies employed by fraudsters.To overcome these limitations, this work presents a hybrid quantum-classical fraud detection framework that strategically combines conventional machine learning models with quantum computing techniques to enhance predictive reliability.Rich feature representations are constructed by integrating transactional, behavioral, and graph-based attributes to capture both local and relational fraud characteristics.A Quantum Autoencoder (QAE) is utilized to perform efficient dimensionality reduction, enabling compact feature encoding while retaining essential discriminatory patterns.The reduced feature set is subsequently processed using Cat-Boost, alongside Quantum Neural Networks (QNNs) or Variation Quantum Classifiers (VQCs) to leverage complementary learning capabilities.Model outputs are aggregated through a stacking-based meta-classification strategy, resulting in improved generalization and decision stability.Performance evaluation is carried out using fraud-oriented metrics, including recall, precision, accuracy, and false positive rate.Experimental analysis indicates that the proposed hybrid approach consistently surpasses standalone classical and quantum models, achieving superior fraud recall with notably fewer false alarms.These results demonstrate the robustness, scalability, and practical viability of quantum-classical integration, positioning the proposed framework as a promising foundation for next-generation financial fraud detection systems.
This work introduces a Quantum Federated Neural Network for Financial Fraud Detection (QFNNFFD), an advanced framework merging Quantum Machine Learning (QML) and quantum computing with Federated Learning (FL) for financial fraud detection. Using quantum technologies' computational power and the robust data privacy protections offered by FL, QFNN-FFD emerges as a secure and efficient method for identifying fraudulent transactions within the financial sector. Implementing a dual-phase training model across distributed clients enhances data integrity and enables superior performance metrics, achieving precision rates consistently above 95%. Additionally, QFNN-FFD demonstrates exceptional resilience by maintaining an impressive 80% accuracy, highlighting its robustness and readiness for real-world applications. This combination of high performance, security, and robustness against noise positions QFNN-FFD as a transformative advancement in financial technology solutions and establishes it as a new approach for privacy-focused fraud detection systems. This framework facilitates the broader adoption of secure, quantum-enhanced financial services and inspires future innovations that could use QML to tackle complex challenges in other areas requiring high confidentiality and accuracy.
Recent advancements in Quantum Neural Networks (QNNs) have demonstrated theoretical and experimental performance superior to their classical counterparts in a wide range of applications. However, existing centralized QNNs cannot solve many real-world problems because collecting large amounts of training data to a common public site is time-consuming and, more importantly, violates data privacy. Federated Learning (FL) is an emerging distributed machine learning framework that allows collaborative model training on decentralized data residing on multiple devices without breaching data privacy. Some initial attempts at Quantum Federated Learning (QFL) either only focus on improving the QFL performance or rely on a trusted quantum server that fails to preserve data privacy. In this work, we propose CryptoQFL, a QFL framework that allows distributed QNN training on encrypted data. CryptoQFL is (1) secure, because it allows each edge to train a QNN with local private data, and encrypt its updates using quantum homomorphic encryption before sending them to the central quantum server; (2) communication-efficient, as CryptoQFL quantize local gradient updates to ternary values, and only communicate non-zero values to the server for aggregation; and (3) computation-efficient, as CryptoQFL presents an efficient quantum aggregation circuit with significantly reduced latency compared to state-of-the-art approaches.
Federated learning enables decentralized, privacy-preserving training but remains vulnerable to privacy leakage in the quantum era. Quantum federated learning (QFL) offers a promising path toward enhanced security and efficiency. However, a practical and experimentally validated QFL protocol utilizing near-term quantum techniques to address data privacy has been lacking. Here, we present QuNetQFL, a QFL protocol implemented on quantum networks, in which local model updates are masked with distributed quantum secret keys, offering information-theoretic security during aggregation. We experimentally validate the protocol on a 4-client quantum network and benchmark its performance using the generated keys on quantum and real-world datasets. Adding a single quantum client substantially improves global accuracy for classifying multipartite entangled and nonstabilizer quantum datasets. For language tasks, we apply QuNetQFL to sentiment analysis by federated fine-tuning of a hybrid classical-quantum language model, achieving comparable and robust performance in simulation and on real quantum hardware. Large-scale simulations further demonstrate scalability to 200 clients for handwritten-digit recognition, with rapid convergence and a 75% reduction in communication cost via model compression. Our work establishes a practical and scalable route to quantum-secure federated learning for the emerging quantum internet.
Detecting financial fraud is a critical aspect of modern intelligent financial systems. Despite the advances brought by deep learning in predictive accuracy, challenges persist-particularly in capturing complex, high-dimensional nonlinear features. This study introduces a novel hybrid quantum recurrent neural network for fraud detection (HQRNN-FD). The model utilizes variational quantum circuits (VQCs) incorporating angle encoding, data reuploading, and hierarchical entanglement to project transaction features into quantum state spaces, thereby facilitating quantum-enhanced feature extraction. For sequential analysis, the model integrates a recurrent neural network (RNN) with a self-attention mechanism to effectively capture temporal dependencies and uncover latent fraudulent patterns. To mitigate class imbalance, the synthetic minority over-sampling technique (SMOTE) is employed during preprocessing, enhancing both class representation and model generalizability. Experimental evaluations reveal that HQRNN-FD attains an accuracy of 0.972 on publicly available fraud detection datasets, outperforming conventional models by 2.4%. In addition, the framework exhibits robustness against quantum noise and improved predictive performance with increasing qubit numbers, validating its efficacy and scalability for imbalanced financial classification tasks.