Generative AI Foundations – Diffusion, Attention, and Creative Systems
Lesson 2: Diffusion in Plain Language
Lesson Objectives
By the end of this lesson, students should be able to:
- Explain diffusion without math-heavy language.
- Describe why prompts guide but do not fully control outputs.
- Recognize why small changes can produce different images.
Lesson Content
Diffusion-style image systems can be understood with a simple idea: learn how to turn noise into structure. During training, the system studies examples and learns how images become noisy and how to reverse that process. During generation, it starts from noise and gradually shapes it toward something that matches the prompt and controls.
This is why two runs of the same prompt can produce different results. The process starts from random variation and follows learned patterns. Prompt wording, seed, guidance strength, style instructions, image references, and tool settings can all change the path.
For students, the practical lesson is to prompt and review iteratively. You rarely get a final image in one attempt. You adjust subject, composition, lighting, style, exclusions, and review criteria.
Enroll to continue this lesson.
The preview above shows the lesson objectives and opening lesson content. Enroll to view the full lesson, complete the practice work, and take the lesson quiz.