Summary: I coded one project EVERY WEEK for a YEAR
Games
-
Platformer Game:
- First project with a gliding mechanic.
- Used HTML canvas to draw basic shapes.
- Despite its flaws, plans to revisit it in the future.
-
Polar Flappy Bird Game:
- Used polar coordinates for mechanics.
- Made with simple shapes and math (e.g., spiral pattern using
R = θ). - Ran well on mobile.
-
3D Game with a Rasterization-based Renderer:
- Learned rasterization from “2 Minute Papers” lectures.
- Featured enemies and teleportation mechanics.
- Used triangles and linear algebra to create the 3D objects.
- Despite heavy math usage, still used HTML canvas rectangles for pixels.
-
Advanced Games:
- Further games included:
- HTML canvas enhancements.
- A text-based dungeon roguelike.
- Created a custom chess variant, inspired by historical and unique chess pieces:
- Example pieces: Chameleon, Hyaena, Wizard, Flying Cat, etc.
- Added multiple game modes, such as Chameleon Chess, where pieces change type every turn.
- Further games included:
Math Projects
-
Complex Numbers and Euler’s Formula:
- Explored rotation and beauty in complex numbers.
- Created a spirograph to demonstrate rotation effects.
-
Mandelbrot Set and Julia Set Visualizations:
- Built a Mandelbrot plotter and a point tracker.
- Used WebGL to improve rendering.
- Visualized Julia sets with mouse-controlled parameters.
-
Cellular Automata:
- Developed simulations based on simple rules.
- Example: Conway’s Game of Life.
- Created a GPU-accelerated version for handling multiple cells.
- Developed projects like Boids, simulating bird flocking, and Langton’s Ant with customizable patterns.
-
Other Math Projects:
- Created graphing tools for 2D/3D equations.
- Developed animations for relaxing weeks.
- Experimented with wave function collapse algorithms and a multiplayer version of Conway’s Game of Life.
AI Projects
-
Flappy Bird with Genetic Algorithms:
- Used genetic algorithms (simpler than other ML techniques) to teach AI how to play Flappy Bird.
- Improved the AI over generations by copying top-performing solutions.
-
GPT-3-based Adventure Game:
- Created an adventure game using OpenAI’s GPT-3 model.
-
Perceptron-based XOR Operation:
- Built a neural network to solve the XOR problem, a fundamental AI task requiring multi-layered neural networks.
-
Generative Adversarial Network (GAN):
- Built a GAN to generate magic cards.
- Trained the AI using TensorFlow and Google Colab to create semi-realistic cards.