Deep unsupervised anomaly detection in high-frequency markets
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
Inspired by recent advances in the deep learning literature, this article introduces a novel hybrid anomaly detection framework specifically designed for limit order book (LOB) data. A modified Transformer autoencoder architecture is proposed to learn rich temporal LOB subsequence representations, which eases the separability of normal and fraudulent time series. A dissimilarity function is then learned in the representation space to characterize normal LOB behavior, enabling the detection of any anomalous subsequences out-of-sample. We also develop a complete trade-based manipulation simulation methodology able to generate a variety of scenarios derived from actual trade–based fraud cases. The complete framework is tested on LOB data of five NASDAQ stocks in which we randomly insert synthetic quote stuffing, layering, and pump-and-dump manipulations. We show that the proposed asset-independent approach achieves new state-of-the-art fraud detection performance, without requiring any prior knowledge of manipulation patterns.
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 | 0.003 |
| Open science | 0.002 | 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