Super Mario Maker levels built by software using machine learning

Programmers have used machine learning to teach machines how to build Super Mario Marker levels without any involvement from humans. Programmer Adam Geitgey and his team taught a computer to create Super Mario Maker levels by using the principles of machine learning wherein they fed their algorithm with previous Super Mario Bros levels and after lots of training the computer spitted out playable Super Mario Maker levels without any human involvement. According to Geitgey, Super Mario levels are effectively objects placed in a grid as shown below: Each of these objects can be represented in a simple character i.e. a blank space can be represented by ‘-‘; solid block can be represented by ‘=’; breakable brick by ‘#’; coin block by ‘?’; and so on and so forth. When the…


Link to Full Article: Super Mario Maker levels built by software using machine learning