BankGAN: A Generative Model for Synthetic FinancialTransactions
Bibliographic record
Abstract
The digital age has equipped financial institutions with vast amounts of data. Privacy concerns have posed challenges to harnessing this data's full potential. Generation of synthetic data is one of the most promising solutions for allowing analysis of the patterns and trends contained in this data without compromising privacy. Although initial methods for generating synthetic data were basic, emerging generative models have expanded the possibilities. However, generating synthetic data for unique datasets, like bank transaction sequences, remains challenging. These sequences exhibit complex variability driven by the various customer transaction behaviors, distinguishing them from the more predictable patterns in other data types. We propose BankGAN, an innovative conditional tabular GAN architecture designed specifically for synthesizing bank transaction sequences that exhibit non-uniform date patterns. We show that BankGAN outperforms a recurrent neural network (RNN)-based model in achieving superior statistical resemblance to real data. Moreover, it excels at replicating features of periodic transactions, surpassing both the RNN and transformer-based models. BankGAN distinguishes itself by generating privacy-preserving synthetic data without compromising data quality—a stark contrast to the existing models where adding privacy-preserving guarantees typically degrades performance.
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 | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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".