AI learns to play Mario Kart 64 by itself

In a blog post released by Kevin Hugues, he created an artificial neural network to learn how to best complete tracks in Mario Kart 64. In order to do this he used a program called TensorFlow, which is an open-source program used to create machine intelligence (kinda like an AI) based on calculations upon a logic flow cycle. So logically, he called the project ‘TensorKart’.

Have a read through the post that Kevin made, it’s a bit confusing if you don’t know what he’s talking about (like I am). However, I found the latter half to be a bit more relatable since that’s when he stops talking about how it was set up, and more about the relation between playing the game and having the AI learn to play the game.

Mario Kart AI video

Articles across the web

Fullstack developer Kevin Hugues made an AI to learn how to play Mario Kart 64 by using a TensorFlow machine intelligence. let's have a look.
Article published on N64 Squid


Subscribe to the mailing list

Follow N64 Squid

  • RSS Feed
  • Tumblr

Popular posts

Random featured posts

Leave a Reply

Your Name (required)

Your Email (required)


Your Message