Create controlled fixtures while checking relational validity, rare-case coverage, real-world utility and disclosure risk.
Learning path
- Synthetic-data contracts: purpose, units and fidelity
- Synthetic relational data: keys, chronology and integrity
- Rare-case simulation: scenario coverage and domain gap
- Synthetic-data utility: downstream tests on an untouched holdout
- Synthetic-data privacy: copies, nearest neighbors and disclosure
- Synthetic-data releases: seeds, manifests and rejection gates
- Project: simulate a receipt pipeline and prove its limits
Connected foundations
Use these data and evaluation contracts before the project.