Definition
A class of generative models that produce high-quality data (like images) by iteratively reversing a process that adds random noise to the training data.
In short: A class of generative models that produce high-quality data (like images) by iteratively reversing a process that adds random noise to the training data.