AI and Machine Learning are the hot topics this decade, not only in industry, but also in computer science education, as they contain many exciting and interesting new concepts. However, many educators think these topics are “magical”, too complex, and that they can’t be taught in classrooms. But Machine Learning does not have to be crazy magic! Using the Snap!-environment, which is perfect for creating games with, we can use Machine Learning to teach an AI how to beat us at our favorite games. We use Reinforcement Learning implemented entirely in Snap! itself, without relying on external libraries, JavaScript-wrappers, or servers. Because of this, the learning algorithm is fast, intuitive, simple, and can easily be visualized for easier understanding. It also means that our approach can be transferred to many other game settings, simulations, problems, etc. Explore the magic with us!
Created February 13, 2020
Last updated February 13, 2020
Published February 13, 2020