Schedule

• What are generative models?
• Applications of generative models
• A brief history of generative modeling, from VAEs to GANs to Diffusion.
• What are diffusion models?

Suggested Readings:

• Diffusion models from scratch
• smalldiffusion codebase
• Elucidating the Design Space of Diffusion-Based Generative Models
• Analyzing and Improving the Training Dynamics of Diffusion Models

• Introduce different interpretations and perspectives on diffusion
• Derivation of DDPM and DDIM, SDE reversal and probability flow ODE.
• Fokker-Planck and connections to physics.

Suggested Readings:

• What are diffusion models?
• SDE interpretation blog post
• Step-by-step diffusion: an elementary tutorial
• In-Depth Mathematics of Diffusion: Stochastic Interpolants

• Conditional diffusion training and inference.
• Classifier and classifier-free guidance.
• Using diffusion models to solve inverse problems.

Suggested Readings:

• Guidance: a cheat code for diffusion models
• Classifier-free guidance paper
• Latent diffusion models
• Guiding a diffusion model with a bad version of itself

• What is the distillation of diffusion models? Fast, few-step generation with diffusion models.
• New architectures and one-step distillation: Consistency Models and MeanFlow.

Suggested Readings:

• Rectified Flow
• Consistency models
• Meanflow

• Diffusion Policy vs Action-Chunking Transformers
• Adapting diffusion architectures to new applications.

Suggested Readings:

• DreamFusion: Text-to-3D using 2D Diffusion
• ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation

• Why do diffusion models generalize?
• Inductive biases of diffusion models: training and architecture.
• How to train better and focus on noise levels that matter most.

Suggested Readings:

• Interpreting and improving diffusion models from an optimization perspective
• Generalization in diffusion models arises from geometry-adaptive harmonic representations
• Closed-Form Diffusion Models
• Nuclear Norm Regularization for Deep Learning
• An analytic theory of creativity in convolutional diffusion models