Generative Modeling#
What are Generative Models?#
Generative models are learning something intuitively easy but in practce is an expensive process. Lets pretend we have an image \(x\) and the collection of ALL possible images is \(X\). Then our \(x\) is a single sample from the population \(X\). The purpose of a generative model is to learn the probability of \(x\), written as the \(p(x)\). If we learn this distribution, then we can sample from it to generate new samples. How we learn \(p(x)\) changes based on which method we use.
Acknowledgments#
Initial version: Mark Neubauer
© Copyright 2025