Rules, simulators, or models create examples with chosen properties, labels, or edge cases. It can expand coverage or protect some sensitive sources, but generated artifacts and biases can be learned by downstream models.
Synthetic data is information generated or simulated rather than directly collected from real-world events or people.
Rules, simulators, or models create examples with chosen properties, labels, or edge cases. It can expand coverage or protect some sensitive sources, but generated artifacts and biases can be learned by downstream models.