visualizing the reinforcement training of a NPC through a maze.
In this mini-project I utilized the q-learning algorithim to create a visualization of what episodes look like and the learning process of an dataset going through said process.
Goals going into the project:
- Learn more about machine learning
- Come up with a fun way to use AI
The Agent/NPC attempts to escape the maze hundreds of times, learning better actions to take each new attempt through negative reinforcement. After letting the NPC try his hand multiple times you can generate the shortest path from the starting point to the end goal.
Steps:
- Download Roblox Studio
- Download pathfindingAi.zip
- Either open or drag and drop file into a Roblox Studio