Lunar Lander is a workshop developed by AWS to educate participants about reinforcement learning and how it can be used to create successful video games.
The purpose of Lunar Lander is to land a rocket (and a successful video game) with reinforcement learning. In this workshop, you will use reinforcement learning (RL) to train a lunar lander vehicle in a Box2D simulation environment to land itself on the surface of the moon. The agent will be trained using Amazon SageMaker RL. You will learn how to optimize agents and compete against other participants to see who can create the best RL model.
With this workshop, we are inviting you to ask yourself:
- What is REINFORCEMENT LEARNING and why do I want to use it in my game?
- How does reinforcement learning compare to other types of MACHINE LEARNING?
- How can I use AMAZON SAGEMAKER RL to train agents using reinforcement learning?
- In what ways will this work for my USE CASE and how can I integrate RL into my own game?
You will compete to optimize an agent that can land accurately with minimal fuel usage. May the best team win!