Expand description
GPU kernels for accounting network analysis.
This module contains Rust DSL implementations that can be transpiled to CUDA or WGSL for GPU execution.
§Kernel Categories
- Journal Transformation - Methods A-E for converting entries to flows
- Network Analysis - Suspense detection, GAAP violations, fraud patterns
- Temporal Analysis - Seasonality, trends, behavioral anomalies
§Method Distribution (Ivertowski et al., 2024)
| Method | Description | Frequency |
|---|---|---|
| A | 1-to-1 debit→credit | 60.68% |
| B | n-to-n bijective | 16.63% |
| C | n-to-m partition | 11.00% |
| D | Higher aggregate | 11.00% |
| E | Decomposition | 0.69% |
Re-exports§
pub use analysis::*;pub use temporal::*;pub use transformation::*;
Modules§
- analysis
- Network analysis kernels for fraud detection and GAAP compliance.
- temporal
- Temporal analysis kernels for time-series anomaly detection.
- transformation
- Journal entry transformation kernels (Methods A-E).