ML Jumper

Click Here to Play If Game Doesn’t Load

ML Jumper is a simple educational game to teach student about the basics of using machine learning (ML) to train artificial intelligent agent (IA). The fundamental concept being taught here is Linear Regression, especially with a focus on the correlation between obstacle speed and the player’s target jumping distance. To make the lesson more fun in class, the game support an “AI Battle” mode, where students to pit their IA against each other. Simply enter the slope (m) and intercept (b) values calculated from the trainning process.

The game currently only support simple Linear Regression using obstacle speed to calculate target jumping distance. In the future, more advanced model, such as multiple linear regression, using the obstacle speed, obstacle height, distance between obstacles, etc. as the independing variables.