Hybrid LSTM-Graph Neural Framework for Robust Financial Fraud Detection and Adversarial Resilience
Quick Answer
The proposed FraudShield AI combines LSTM networks with Graph Topological Features for enhanced financial fraud detection, outperforming Logistic Regression and XGBoost in Precision, Recall, and F1-Score on the PaySim dataset.
Quick Take
It effectively addresses data imbalance and improves resilience against smurfing attacks through dynamic thresholding.
Key Points
- FraudShield AI integrates LSTM with Graph Topological Features for robust detection.
- Achieves superior performance on PaySim dataset compared to Logistic Regression and XGBoost.
- Utilizes Focal Loss to tackle the extreme data imbalance of 0.13% fraud rate.
- Introduces dynamic thresholding to enhance resilience against low-value smurfing attacks.
- Ablation study confirms the importance of both temporal and topological components.
DeepSignal Analysis
What happened
The paper presents FraudShield AI, a hybrid framework that combines LSTM networks with Graph Topological Features for financial fraud detection. It addresses the challenge of extreme data imbalance and enhances resilience against smurfing attacks. The model outperforms traditional methods like Logistic Regression and XGBoost on the PaySim dataset.
Key evidence
- FraudShield AI integrates LSTM networks with Graph Topological Features to capture both temporal sequences and structural relationships in financial data.
- The proposed model achieved superior results in Precision, Recall, and F1-Score compared to Logistic Regression and XGBoost on the PaySim dataset.
- A dynamic thresholding mechanism was introduced to improve the model's resilience against low-value smurfing attacks, addressing the issue of data imbalance.
Why it matters
Financial fraud detection is critical for institutions facing sophisticated money laundering tactics. The proposed hybrid approach not only improves detection accuracy but also adapts to evolving fraudulent strategies. By shifting the focus from isolated transactions to network-level analysis, it offers a more comprehensive understanding of fraud patterns, which is essential in a landscape characterized by low fraud rates and high complexity.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Financial institutions face significant challenges in detecting sophisticated money laundering patterns, such as smurfing and layering, due to extreme data imbalance (0.13% fraud rate) and evolving adversarial evasion tactics. This paper proposes FraudShield AI, a hybrid framework that integrates Long Short-Term Memory (LSTM) networks with hand-crafted Graph Topological Features to capture both temporal sequences and structural relational context. By engineering network-centric features including PageRank Centrality, In-Degree dynamics, and a custom Flow Ratio, the system shifts the detection paradigm from isolated transaction analysis to network-level forensics. A Focal Loss objective is used to address class imbalance, and a dynamic thresholding mechanism is introduced to improve resilience against low-value smurfing attacks. Experimental evaluation on the PaySim dataset shows that the proposed hybrid model substantially outperforms Logistic Regression and XGBoost baselines in Precision, Recall, and F1-Score, particularly on hard-to-detect micro-transaction fraud patterns. An ablation study confirms the complementary contribution of both the temporal and topological components.
| Comments: | 6 pages, 12 figures |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2607.19350 [cs.AI] |
| (or arXiv:2607.19350v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2607.19350 arXiv-issued DOI via DataCite |
Submission history
From: Mariam Ali [view email]
[v1]
Sat, 2 May 2026 02:10:20 UTC (599 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →HOBA: Hierarchical On-Policy Bidding Agents for Adaptive Online Advertising
HOBA (Hierarchical On-policy Bidding Agents) is a novel hierarchical reinforcement learning framework that enhances online advertising bidding systems by improving adaptability and reducing hyperparameter tuning costs. It utilizes a for hyperparameter inference, a SARSA agent for expert model selection, and a dynamic expert pool for bid execution, achieving a +3.6% increase in target cost during large-scale deployment and outperforming state-of-the-art baselines on AuctionNet.