Artificial World Building

Description

Have you ever wondered how tree structure is formed? What about crystals? And what about more complex stuff? How living beings and societies work? If you would like to improve your understanding of the world a little bit more through a practical point of view this is the course for you.

I hope to explore the following questions in this class…

• What are the logic rules behind natural phenomenons?
• How can I simulate and modify them using a computer?

Structure

• Lessons: There is 1 session per week, 2 hour long
• Lessons are recorded and shared with students in case they missed the class or in case they want to review the class
• There is a forum to ask and share anything at anytime

Tools

• Python and Google Colab
• Blender

Schedule

Week 1: How things grow?

We will learn about how simple rules can generate complex figures using different procedural techniques with Python

1. Tree shapes using L-Tree systems
2. 2D cellular automatas for spots, stripes and other patterns
3. 3D cellular automatas for crystals

Week 2: How things evolve?

We will learn about genetic algorithms through a practical examples using python. We will see how to modify the genes of a cellular automata and define a fitness rule to define which shapes pass its genes to the next generation.

Week 3: How things learn?

We will learn about reinforcement learning the algorithm behind how humans and robot vacuum cleaners work. We will train our computer to learn how to move the body of an artificial being.

Week 4: How things interact?

We will learn how to model simple societies inspired in SimCity and bird flocks

Week 5: How things communicate?

We will learn how language can emerge as a need to solve a problem that requires cooperation. We will train state of the art language models like GPT-2.

Prerequisites

• Basic programming skills are useful but not required