Paper: arXiv 2507.22908
Authors: Abhishek Sawaika, Swetang Krishna, Tushar Tomar, Durga Pritam Suggisetti, Aditi Lal, Tanmaya Shrivastav, Nouhaila Innan, Muhammad Shafique
Abstract
Rapid growth of digital transactions has led to a surge in fraudulent activities, challenging traditional detection methods in the financial sector. To tackle this problem, we introduce a specialised federated learning framework that uniquely combines a quantum-enhanced Long Short-Term Memory (LSTM) model with advanced privacy preserving techniques. By integrating quantum layers into the LSTM architecture, our approach adeptly captures complex cross-transactional patters, resulting in an approximate 5% performance improvement across key evaluation metrics compared to conventional models. Central to our framework is “FedRansel”, a novel method designed to defend against poisoning and inference attacks, thereby reducing model degradation and inference accuracy by 4-8%, compared to standard differential privacy mechanisms. This pseudo-centralised setup with a Quantum LSTM model, enhances fraud detection accuracy and reinforces the security and confidentiality of sensitive financial data.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 5.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced mathematical concepts, including quantum machine learning layers and differential privacy formulations, indicating high mathematical complexity. While the abstract presents specific quantitative results (e.g., 5% improvement, 4-8% reduction in degradation), the excerpt lacks detailed implementation code or backtest specifics, placing empirical rigor in the moderate range.
Research Flowchart
flowchart TD
A["Research Goal: Enhance financial fraud detection with privacy-preserving federated learning and quantum AI."] --> B["Methodology: Hybrid Quantum-Enhanced LSTM in a Federated Framework"]
B --> C["Input: Distributed Financial Transaction Data"]
C --> D["Process: Quantum LSTM Feature Extraction & Training"]
D --> E["Process: FedRansel Privacy Preservation<br>(Defense vs. Poisoning/Inference Attacks)"]
E --> F{"Outcome: Evaluation Metrics"}
F --> G["Result 1: ~5% Improvement in Fraud Detection Accuracy"]
F --> H["Result 2: 4-8% Reduction in Model Degradation/Inference Risk"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a federated learning framework that combines a quantum-enhanced LSTM model with a novel privacy mechanism called FedRansel.
Source sentence
we introduce a specialised federated learning framework that uniquely combines a quantum-enhanced Long Short-Term Memory (LSTM) model with advanced privacy preserving techniques.
Assumes 3
- Each client receives an i.i.d. sampled local dataset from a common underlying distribution. data
Source sentence
Each local dataset Dk is i.i.d.-sampled from the common underlying distribution PD.
- The overall dataset contains 20,000 samples per dataset. data
Source sentence
We use N = 20 000 samples for each dataset.
- Experiments are simulated using the default qubit simulator from PennyLane. model
Source sentence
Experiments are simulated using the default qubit simulator from PennyLane
Data
- Financial fraud detection dataset other unknown
Source sentence
We use N = 20 000 samples for each dataset.
Costs
Not applicable
Finds 4
- The quantum-enhanced LSTM yields an approximate 5% performance improvement across key evaluation metrics compared to conventional models. performance
Source sentence
resulting in an approximate 5% performance improvement across key evaluation metrics compared to conventional models.
- FedRansel reduces model degradation and inference accuracy loss by 4-8% relative to standard differential privacy mechanisms. performance
Source sentence
reducing model degradation and inference accuracy by 4-8%, compared to standard differential privacy mechanisms.
- The QLSTM outperforms the classical LSTM by about 2% AUC, 5% Accuracy, and 10% Recall for Dataset 1, and by 3% AUC, 6% Accuracy, and 4% Recall for Dataset 2. performance
Source sentence
The QLSTM outperforms the classical LSTM by approximately 2% in AUC, 5% in Accuracy, and 10% in Recall for Dataset 1, and by 3% in AUC, 6% in Accuracy, and 4% in Recall for Dataset 2.
- Adding the FedRansel privacy module causes only a negligible drop of about 1% in performance across both datasets. performance
Source sentence
The addition of the FedRansel privacy-preserving module results in a negligible drop of approximately 1% in performance across both datasets
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 24 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (76)
- Fully decentralized federated learning (2018)
- Enhancing quantum support vector machines through variational kernel training (2023)
- FedQNN: Federated Learning using Quantum Neural Networks (2024)
- Optimizing Low-Energy Carbon IIoT Systems With Quantum Algorithms: Performance Evaluation and Noise Robustness (2025)
- A variational quantum perceptron with Grover’s algorithm for efficient classification (2024)
- QNN-VRCS: A Quantum Neural Network for Vehicle Road Cooperation Systems (2025)
- Decentralized federated learning: A survey and perspective (2024)
- Horizontal federated learning (2020)
and 68 more
- Vertical federated learning: Concepts, advances, and challenges (2024)
- QFNN-FFD: Quantum Federated Neural Network for Financial Fraud Detection — → 2404.02595
- FedTrans: Efficient Federated Learning via Multi-Model Transformation (2024)
- Multi-model federated learning (2022)
- A secure federated transfer learning framework (2020)
- Financial fraud detection applying data mining techniques: A comprehensive review from 2009 to 2019 (2021)
- Advanced financial fraud detection using GNN-CL model — → 2407.06529
- Model aggregation techniques in federated learning: A comprehensive survey (2024)
- Reviewing federated learning aggregation algorithms; strategies, contributions, limitations and future perspectives (2023)
- Local model poisoning attacks to $\ $Byzantine-Robust$\ $ federated learning (2020)
- Threats, attacks and defenses to federated learning: issues, taxonomy and perspectives (2022)
- Federated learning with differential privacy: Algorithms and performance analysis (2020)
- Deep learning with differential privacy (2016)
- Evaluating gradient inversion attacks and defenses in federated learning (2021)
- Federated quantum machine learning with differential privacy (2024)
- Poisoning attacks in federated learning: A survey (2023)
- Threats to federated learning: A survey (2020) — link
- Membership Inference Attacks and Defenses in Federated Learning: A Survey (2024)
- Membership inference attacks on machine learning: A survey (2022)
- Local and central differential privacy for robustness and privacy in federated learning (2020) — link
- Can you really backdoor federated learning? (2019) — link
- Learning differentially private recurrent language models (2017) — link
- Comprehensive privacy analysis of deep learning: Passive and active white-box inference attacks against centralized and federated learning (2019)
- Membership Inference Attacks Against Machine Learning Models (2017) — link
- Quantum federated learning with quantum networks (2024)
- Quantum federated learning with decentralized data (2022)
- Quantumfed: A federated learning framework for collaborative quantum training (2021)
- Federated quantum machine learning (2021)
- Quantum federated learning with quantum data (2022)
- Foundations of quantum federated learning over classical and quantum networks (2023)
- Quantum long short-term memory (2022)
- Quantum long short-term memory (QLSTM) vs. classical LSTM in time series forecasting: a comparative study in solar power forecasting (2024)
- Federated quantum long short-term memory (fedqlstm) (2024)
- FedQNN: Federated Learning using Quantum Neural Networks (2024) — link
- Financial fraud detection using quantum graph neural networks (2024)
- Fraud detection bank dataset 20K records binary — link
- Synthetic Financial Datasets For Fraud Detection — link
- Applied federated learning: Architectural design for robust and efficient learning in privacy aware settings (2022) — link
- Optuna: A Next-generation Hyperparameter Optimization Framework (2019)
- Support vector method for novelty detection (1999)
- Long short-term memory (1997)
- QGAPHEnsemble: Combining Hybrid QLSTM Network Ensemble via Adaptive Weighting for Short Term Weather Forecasting (2025) — link
- Communication-efficient learning of deep networks from decentralized data (2017)
- The regression analysis of binary sequences (1958)
- A History of Financial Fraud: America’s First Financial Fraudsters (2025)
- Quantum Technology Use Cases as Fuel for Value in Finance (2025)
- Quantum federated learning through blind quantum computing (2021)
- Non-IID quantum federated learning with one-shot communication complexity (2023)
- Quantum federated learning through blind quantum computing (2021)
- In Search of Quantum Advantage: Estimating the Number of Shots in Quantum Kernel Methods (2024) — link
- Federated Quantum Machine Learning with Differential Privacy (2023)
- Variational quantum algorithms (2021)
- Quantum support vector machine for classifying noisy data (2024)
- A survey on homomorphic encryption schemes: Theory and implementation (2018)
- A financial fraud detection model based on LSTM deep learning technique (2020)
- Enhanced credit card fraud detection based on attention mechanism and LSTM deep model (2021)
- Intelligent financial fraud detection: a comprehensive review (2016)
- Financial fraud: A literature review (2019)
- Digital finance and corporate financial fraud (2023)
- Financial fraud detection using graph neural networks: A systematic review (2024)
- Financial fraud detection based on machine learning: a systematic literature review (2022)
- Financial fraud detection: a comparative study of quantum machine learning models (2024)
- Federated Learning for Credit Card Fraud Detection: Key Fundamentals and Emerging Trends (2024)
- Responsible and effective federated learning in financial services: A comprehensive survey (2023)
- Quantum circuit learning (2018)
- Quantum support vector machine for big data classification (2014)
- Face patterns analysis and recognition system based on Quantum Neural Network QNN (2022)
- QuantumLeap: Hybrid quantum neural network for financial predictions (2022)
2 of 76 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.