Generative AI Notes

Structured study notes for Generative AI.

The goal is to build the subject incrementally from first principles:

1. Probabilistic Machine Learning
2. Latent Variable Models
3. Variational Methods
4. Diffusion Models
5. Generative Adversarial Networks
6. Auto-Regressive Language Models
7. State-Space Models
8. Flow-Based Models

How to Use These Notes

Each topic lives in its own Markdown file. This keeps the notes modular, searchable, and easy to expand.

Recommended workflow:

1. Add handwritten notes, screenshots, or lecture bullets.
2. Convert them into Markdown.
3. Expand each concept using:
   • Definition
   • Mathematical formulation
   • Training objective
   • Sampling/inference process
   • Strengths and weaknesses
   • Interview questions
   • Implementation notes
4. Commit changes frequently.

Structure

.
├── README.md
├── syllabus.md
├── glossary.md
├── concept-map.md
├── 01-probabilistic-ml/
├── 02-variational-methods/
├── 03-adversarial-models/
├── 04-autoregressive-models/
├── 05-state-space-models/
├── 06-flow-based-models/
├── math/
├── papers/
├── code/
└── questions/

Current Scope

These notes are based on the initial handwritten Generative AI topic outline:

• Introduction to Probabilistic ML
• Variational Methods and Latent Variable Models
  • Gaussian Mixture Models
  • Variational Autoencoders
  • Diffusion Models / DDPMs
• Adversarial Family
  • Generative Adversarial Networks
• Auto-Regressive Models
  • Transformer-Based Language Models
• State-Space Models
• Flow-Based Models