+Diffusion Models are a class of [Generative Models](/wiki/generative_models) that learn to create new data by reversing a gradual "noising" process. They begin with pure noise and, guided by learned [Neural Networks](/wiki/neural_networks), iteratively refine it into a coherent, high-fidelity sample, like an image or audio clip. This elegant process allows them to synthesize diverse and remarkably realistic outputs.
+## See also
+- [Deep Learning](/wiki/deep_learning)
+- [Machine Learning](/wiki/machine_learning)
+- [Image Generation](/wiki/image_generation)
... 1 more lines