Goes beyond simple transformations by using generative models to understand the underlying data distribution. It creates entirely new, high-fidelity data instances that are statistically congruent with the original dataset, dramatically expanding small datasets to improve model generalization and reduce overfitting.