Finding a needle in a haystack: A machine learning framework for anomaly detection in payment systems
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
We propose a flexible machine learning (ML) framework for real-time transaction monitoring in high-value payment systems (HVPS), which are central to a country’s financial infrastructure and integral to financial stability. This framework can be used by system operators and overseers to detect anomalous transactions, which—if caused by a cyber attack or an operational outage and left undetected—could have serious implications for the HVPS, its participants and the financial system more broadly. Given the high volume of payments settled each day and the scarcity of actual anomalous transactions in HVPS, detecting anomalies resembles finding a needle in a haystack. Therefore, our framework employs a layered approach to manage the high volume of payments and isolate potential anomalies. In the first layer, a supervised ML algorithm is used to identify and separate ‘typical’ payments from ‘unusual’ payments. In the second layer, only the ‘unusual’ payments are run through an unsupervised ML algorithm for anomaly detection. We test this framework using artificially manipulated transactions and payments data from the Canadian HVPS. The ML algorithm employed in the first layer achieves a detection rate of 93 %, marking a significant improvement over commonly-used econometric models. The ML algorithm used in the second layer marks the artificially manipulated transactions as nearly twice as suspicious as the original transactions, proving its effectiveness.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it