Following our study of Generative Adversarial Networks, this chapter introduces Diffusion Models, a distinct and effective approach to generative modeling. These models have demonstrated remarkable success in generating high-quality data, especially images, by progressively adding noise to data points and then learning the reverse denoising process.
This chapter covers the following key areas:
You will gain an understanding of the theory behind diffusion models and the practical considerations for implementing and improving them effectively.
4.1 Mathematical Foundations: Stochastic Differential Equations
4.2 Denoising Diffusion Probabilistic Models (DDPM)
4.3 Score-Based Generative Modeling
4.4 Improved Techniques: DDIM and Variance Schedules
4.5 Classifier Guidance and Classifier-Free Guidance
4.6 Architectural Considerations for Diffusion Models (U-Net)
4.7 Hands-on Practical: Implementing a Basic DDPM
© 2025 ApX Machine Learning