Synthetic data
Synthetic data: Artificially generated data from models, simulations or rules, used to augment or replace real observations. They are useful for rare scenarios, privacy and systematic controls, but inherit generator errors. Validation on real data and provenance tracking are essential.
In italiano: Dati SinteticiLegge III · Come una rete impara dai datiThey are useful for rare scenarios, privacy and systematic controls, but inherit generator errors. Validation on real data and provenance tracking are essential.
Examples
- Simulated industrial images
- Dialogues generated for rare intents
- Tabular records without real personal data