Solves severe class imbalance for critical events like fraud or rare diseases. Generative models learn the complex characteristics from few examples and create vast numbers of realistic variations, providing robust training sets that dramatically improve a model's detection capabilities.