Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Hybrid Quantum Intelligence for Detecting Sophisticated Financial Fraud in Dynamic Transaction Ecosystems

Mudimela Madhusudhan, Pramoda Patro

Published Jul 14, 2026• 0 citations
Publisher

Abstract

Abstract Fraud develops rapidly when transactions happen quickly through electronic pipelines. The traditional fraud detection methods, which are reliable but slow to adapt, cannot keep pace with fraudsters' rapid development of new schemes. They also have trouble adapting to complex relationships and creating explanations that regulators can understand. In this paper, we are proposing a multi-pronged attack on the fraud problem using hybrid quantum-classical techniques. QAFEN-CNN utilizes the concept of adaptive quantum feature entanglement, with volatility-based depth controls to achieve a separation increase of 8%. VQTM-LSTM uses parameterized quantum embeddings as input to each recurrent gate to achieve a 7.9% increase in long-horizon recall. QGCL-Net utilizes quantum enhanced graph construction to model transactional ecosystems, achieving a 12.3% increase in the quality of sub-graphs constructed. Lastly, QVAE-DD proposes a dual-decoder quantum variational autoencoder that generates both input representations and salient explanation representations. The last proposal, QMRL-XAI, utilizes quantum meta-reinforcement learning in combination with counter-factual reasoning to generate policy adaptations after only five episodes. All of these proposals together demonstrate how the use of quantum computing can provide improved accuracy, improve resistance to class imbalance, and produce explanations that can be understood by regulators — a fascinating outline for future applications of quantum computing in anti-fraud analytics.

DOI
10.21203/rs.3.rs-9881318/v1

Related papers

Quantum Classical Synergy for Fraud Detection

P Karthikeyan, N Babu, K Lakshmi Pathi +3 more

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.

International Journal of Emerging Research in Science Engineering and Management•Mar 30, 2026•0 citations

Design of an Iterative Explainable Quantum Machine Learning Method for Adaptive Fraud Detection in Financial Transaction Systems

Pramoda Patro Mudimela Madhusudhan

Financial transaction systems are becoming increasingly complex and large, in this scenario, fraud detection patterns have become more sophisticated than the conventional detection mechanisms by themselves. While machine learning models have been developed for operational purposes, very few of them are interpretable, sensitive to class imbalance or do not generalize to new fraudulent behaviors. Research on quantum machine learning (QML) for fraud detection is still nascent, with many efforts concentrating on classification without addressing explainability or even attempting to deal with graph-structured and sequential data comprehensively. This study aims to fill this gap by presenting a new explainable quantum machine learning architecture to detect fraud from financial transaction datasets. Three core quantum models are developed and validated: (1) Hybrid Quantum-Loss Adaptive Variational Autoencoder (HQLAVAE) for anomaly detection-has a fidelity aware, class-adaptive loss function that enhances reconstruction of typical transaction patterns and sensitively detects anomalies; (2) The third is Quantum Feature Relevance Graph Attribution (Q-FRGA): an explainable quantum graph neural network particularly capturing transactional relationships with respect to node-level fraud-attributed predictions from Hilbert-space-based relevance scoring; and (3) Quantum Environment-Adaptive Policy Refinement (Q-EAPR), a reinforcement learning model integrating distributional quantum Bellman updates with a novel tracking mechanism for policies to enable online processing of fraud behavior changes. Moreover, two cross-model innovations for enhanced explainability and model integration include Cross-Modal Quantum Counterfactual Explainer (CQCE), generating explainable perturbation-based rationales for single predictions as well as Quantum Residual Explainability-Integrated Ensemble Framework (Q-REIEF) using an ensemble of all models with residual error corrections and meta-explanation tracing sets. This work further advances quantum explainable AI in the areas of transparent, adaptive and high-accuracy fraud detection-all critical gaps in interpretability, data imbalance and behavioral drift on financial systems.

•Jun 1, 2026•0 citations

Fraud Detection using Quantum Machine Learning(QML)

Sahil Somaji Kamble, Suyog Sudhir Pawar, Tejas Babasaheb Veer +1 more

Fraud detection is a Crucial aspect of securing financial systems and protecting against fraudulent activities. This paper presents a new method for Detecting fraud, combining quantum-inspired feature engineering and neural network modeling, which achieves a classification accuracy of 92%, with a precision and recall of 0.92 for non-fraudulent transactions. Utilizing QuantumCircuit from Qiskit, the approach encodes input features into quantum states through angle encoding, improving data representation and capturing complex feature interactions such as entanglement between qubits. The method employs a Quantum Neural Network (QNN) for classifying records and effectively identifying fraudulent transactions. Through real-world dataset experiments, we demonstrate the efficacy of the proposed approach, achieving a precision of 0.65 and recall of 0.68 for fraudulent transactions, while maintaining a false positive rate below 10%. This research leverages advanced machine learning techniques, including balanced sampling strategies to address class imbalance and hyperparameter tuning with Grid-SearchCV for optimal model performance. The paper presents a comprehensive evaluation of the approach, highlighting its strengths in adapting to dynamic and complex environments. Quantum-inspired methods offer the potential to significantly enhance fraud detection systems. However, challenges such as computational complexity and resource requirements remain areas for further investigation. Future research will aim to scale these models for larger datasets and investigate their applicability in fields such as healthcare and e-commerce. In conclusion, our work demonstrates the promising use of quantum-enhanced machine learning for fraud detection. This innovative approach paves the way for developing advanced detection systems that can effectively respond to the evolving landscape of fraudulent activities. Future studies could concentrate on refining these techniques and exploring their potential applications in other areas to further harness the advantages of quantum computing in machine learning.

Global Conference on Communications and Information Technologies (GCCIT)•Oct 25, 2024•6 citations

Quantum Graph Neural Network for Financial Fraud Detection

Vajrapu Jyothirmayi

Financial fraud has become a major challenge with the rapid growth of online banking, digital payments,and e-commerce, causing significant financial losses. Traditional fraud detection methods often analysetransactions independently, making it difficult to identify complex patterns involving multiple entities. Graph Neural Networks (GNNs) address this limitation by representing financial transactions as graphs and learning relationships among accounts, users, and merchants. This work proposes an Adaptive Hybrid Quantum Graph Neural Network (AHQGNN) that combines graph-based learning with quantum machine learning. A classical GNN processes the transaction graph, classifying high-confidence predictions directly, while uncertain predictions are selectively analysed using a Variational Quantum Circuit (VQC). This adaptive strategy reduces unnecessary quantum computation while maintaining effective fraud detection. The framework is implemented using Python, PyTorch Geometric, and Qiskit, and evaluated using Accuracy, Precision, Recall, F1-Score, and ROC-AUC. The integration of GNNs with selective quantum classification provides an efficient and scalable framework for financial fraud detection in modern digital financial systems.

International Journal of Advanced Scientific Innovation•Nov 7, 2026•0 citations

HQRNN-FD: A Hybrid Quantum Recurrent Neural Network for Fraud Detection

Yaochong Li, Yi-Fan Zhang, Ruiqing Xu +2 more

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.

Entropy•Aug 27, 2025•11 citations

Hybrid Quantum-Classical Models and Optimization Techniques for Financial Fraud Detection

K. Deepa, Raunak Singh, Kondapuram Tharun

Every day machine learning in financial systems is increasing with financial institutions taking every steps in their power to detect and block counterfeit transactions. Today, modern electronic payment instruments and financial facilities are becoming more and more susceptible to new methods of fraud which are unobtrusively reproducing the activities of legitimate interests to the detriment of the economy which is losing incalculable sums of money. In order to overcome this the combination technique of quantum and classical methods of calculating the data is a great way to enhance the reliability and accuracy of the fraud detection. It is applied by means of Quantum Support Vector Machine, Variational Quantum Classifier, the classical algorithms of Support Vector Machine and K-Nearest Neighbor with an ensemble based on Quadratic Unconstrained Binary Optimization. These algorithms are used together to scan a stream of transactions in real time and classify what is legitimate behaviour and what is unfavourable behaviour. The work provides evidence to prove that hybrid quantum-classical techniques can improve on the performance of classical systems using improved recall, less false negatives and the speed of detection of new types of frauds. It is shown that the integration of quantum intelligence based upon classical methods of analysis is a powerful and fruitful means to enhance levels of secure financial practice and help combat large scale financial frauds in modern digital environments.

International Conference on IoT Based Control Networks and Intelligent Systems (ICICNIS)•Dec 15, 2025•0 citations
Abstract Fraud develops rapidly when transactions happen quickly through electronic pipelines. The traditional fraud detection methods, which are reliable but slow to adapt, cannot keep pace with fraudsters' rapid development of new schemes. They also have trouble adapting to complex relationships and creating explanations that regulators can understand. In this paper, we are proposing a multi-pronged attack on the fraud problem using hybrid quantum-classical techniques. QAFEN-CNN utilizes the concept of adaptive quantum feature entanglement, with volatility-based depth controls to achieve a separation increase of 8%. VQTM-LSTM uses parameterized quantum embeddings as input to each recurrent gate to achieve a 7.9% increase in long-horizon recall. QGCL-Net utilizes quantum enhanced graph construction to model transactional ecosystems, achieving a 12.3% increase in the quality of sub-graphs constructed. Lastly, QVAE-DD proposes a dual-decoder quantum variational autoencoder that generates both input representations and salient explanation representations. The last proposal, QMRL-XAI, utilizes quantum meta-reinforcement learning in combination with counter-factual reasoning to generate policy adaptations after only five episodes. All of these proposals together demonstrate how the use of quantum computing can provide improved accuracy, improve resistance to class imbalance, and produce explanations that can be understood by regulators — a fascinating outline for future applications of quantum computing in anti-fraud analytics.
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.
Financial transaction systems are becoming increasingly complex and large, in this scenario, fraud detection patterns have become more sophisticated than the conventional detection mechanisms by themselves. While machine learning models have been developed for operational purposes, very few of them are interpretable, sensitive to class imbalance or do not generalize to new fraudulent behaviors. Research on quantum machine learning (QML) for fraud detection is still nascent, with many efforts concentrating on classification without addressing explainability or even attempting to deal with graph-structured and sequential data comprehensively. This study aims to fill this gap by presenting a new explainable quantum machine learning architecture to detect fraud from financial transaction datasets. Three core quantum models are developed and validated: (1) Hybrid Quantum-Loss Adaptive Variational Autoencoder (HQLAVAE) for anomaly detection-has a fidelity aware, class-adaptive loss function that enhances reconstruction of typical transaction patterns and sensitively detects anomalies; (2) The third is Quantum Feature Relevance Graph Attribution (Q-FRGA): an explainable quantum graph neural network particularly capturing transactional relationships with respect to node-level fraud-attributed predictions from Hilbert-space-based relevance scoring; and (3) Quantum Environment-Adaptive Policy Refinement (Q-EAPR), a reinforcement learning model integrating distributional quantum Bellman updates with a novel tracking mechanism for policies to enable online processing of fraud behavior changes. Moreover, two cross-model innovations for enhanced explainability and model integration include Cross-Modal Quantum Counterfactual Explainer (CQCE), generating explainable perturbation-based rationales for single predictions as well as Quantum Residual Explainability-Integrated Ensemble Framework (Q-REIEF) using an ensemble of all models with residual error corrections and meta-explanation tracing sets. This work further advances quantum explainable AI in the areas of transparent, adaptive and high-accuracy fraud detection-all critical gaps in interpretability, data imbalance and behavioral drift on financial systems.
Fraud detection is a Crucial aspect of securing financial systems and protecting against fraudulent activities. This paper presents a new method for Detecting fraud, combining quantum-inspired feature engineering and neural network modeling, which achieves a classification accuracy of 92%, with a precision and recall of 0.92 for non-fraudulent transactions. Utilizing QuantumCircuit from Qiskit, the approach encodes input features into quantum states through angle encoding, improving data representation and capturing complex feature interactions such as entanglement between qubits. The method employs a Quantum Neural Network (QNN) for classifying records and effectively identifying fraudulent transactions. Through real-world dataset experiments, we demonstrate the efficacy of the proposed approach, achieving a precision of 0.65 and recall of 0.68 for fraudulent transactions, while maintaining a false positive rate below 10%. This research leverages advanced machine learning techniques, including balanced sampling strategies to address class imbalance and hyperparameter tuning with Grid-SearchCV for optimal model performance. The paper presents a comprehensive evaluation of the approach, highlighting its strengths in adapting to dynamic and complex environments. Quantum-inspired methods offer the potential to significantly enhance fraud detection systems. However, challenges such as computational complexity and resource requirements remain areas for further investigation. Future research will aim to scale these models for larger datasets and investigate their applicability in fields such as healthcare and e-commerce. In conclusion, our work demonstrates the promising use of quantum-enhanced machine learning for fraud detection. This innovative approach paves the way for developing advanced detection systems that can effectively respond to the evolving landscape of fraudulent activities. Future studies could concentrate on refining these techniques and exploring their potential applications in other areas to further harness the advantages of quantum computing in machine learning.
Financial fraud has become a major challenge with the rapid growth of online banking, digital payments,and e-commerce, causing significant financial losses. Traditional fraud detection methods often analysetransactions independently, making it difficult to identify complex patterns involving multiple entities. Graph Neural Networks (GNNs) address this limitation by representing financial transactions as graphs and learning relationships among accounts, users, and merchants. This work proposes an Adaptive Hybrid Quantum Graph Neural Network (AHQGNN) that combines graph-based learning with quantum machine learning. A classical GNN processes the transaction graph, classifying high-confidence predictions directly, while uncertain predictions are selectively analysed using a Variational Quantum Circuit (VQC). This adaptive strategy reduces unnecessary quantum computation while maintaining effective fraud detection. The framework is implemented using Python, PyTorch Geometric, and Qiskit, and evaluated using Accuracy, Precision, Recall, F1-Score, and ROC-AUC. The integration of GNNs with selective quantum classification provides an efficient and scalable framework for financial fraud detection in modern digital financial systems.
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.
Every day machine learning in financial systems is increasing with financial institutions taking every steps in their power to detect and block counterfeit transactions. Today, modern electronic payment instruments and financial facilities are becoming more and more susceptible to new methods of fraud which are unobtrusively reproducing the activities of legitimate interests to the detriment of the economy which is losing incalculable sums of money. In order to overcome this the combination technique of quantum and classical methods of calculating the data is a great way to enhance the reliability and accuracy of the fraud detection. It is applied by means of Quantum Support Vector Machine, Variational Quantum Classifier, the classical algorithms of Support Vector Machine and K-Nearest Neighbor with an ensemble based on Quadratic Unconstrained Binary Optimization. These algorithms are used together to scan a stream of transactions in real time and classify what is legitimate behaviour and what is unfavourable behaviour. The work provides evidence to prove that hybrid quantum-classical techniques can improve on the performance of classical systems using improved recall, less false negatives and the speed of detection of new types of frauds. It is shown that the integration of quantum intelligence based upon classical methods of analysis is a powerful and fruitful means to enhance levels of secure financial practice and help combat large scale financial frauds in modern digital environments.