1 Definition and classification of generative models
Generative models are an unsupervised learning method. The definition is that given a bunch of training data produced by the true distribution, our model learns from it, and then produces new samples that approximate the true distribution.
Generative models are divided into explicit and implicit generative models:
Why generative models are important:
Generating samples, coloring problems, reinforcement learning applications, implicit representation inference, etc.