Intelligent Anti-Money-Laundering Platform Based on Dynamic Graph Neural Networks
Bibliographic record
Abstract
This project, titled “ZhiDun ZDC: An Intelligent Abnormal Transaction Detection and Risk Warning Platform Based on Dynamic Graph Neural Networks,” focuses on applying advanced graph learning techniques to anti–money laundering (AML) in the financial sector. The study proposes a Dynamic Graph Neural Network (Dynamic GNN) framework that models banking transaction networks as evolving graphs. Unlike traditional rule-based systems, the model dynamically adjusts the information propagation weights between accounts based on anomaly scores, enabling it to focus on suspicious entities and transaction paths. A dual-modality self-supervised learning approach is designed to jointly reconstruct both the network structure and transaction attributes, allowing the system to detect anomalies without requiring labeled data. To handle large-scale financial graphs, the research introduces a hierarchical graph training strategy using the Metis partitioning algorithm combined with K-means++ sampling, achieving high scalability and efficiency. Experimental evaluations on multiple datasets—including the Elliptic++ financial transaction dataset—demonstrate that ZDC achieves superior performance (AUC ≈ 0.93) compared with existing graph-based anomaly detection methods. Beyond technical innovation, the project explores its practical application within commercial banking, tailoring the model to real-world scenarios such as cross-border payments, layered transfers, and suspicious fund flows. The system outputs interpretable “risk paths” and can be integrated into banks’ existing AML and compliance systems for real-time monitoring and decision support.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".