Maxime Louis
Use of deep learning to maximize a score in a simplified Space Invaders game. (python)
Source
Space invaders AI based on deep learning (DeepMind style)
Similar Posts
The Era of Artificial Intelligence
Artificial intelligence: Why politicians don't say truth about AGI? [D]
What’s On The Other Side Of A Black Hole?
7 thoughts on “Space invaders AI based on deep learning (DeepMind style)”
Comments are closed.
Wait a minute… the algorithm slows down the whole game to calculate further maneuvers? That's kinda cheating. It should work asynchronously.
Cool. Love the background music, what is it?
if i want you to create a algorithm and training network for a game that can learn that way. can you do that? it can be a paid project.
Maybe one day an ai that can learn any game will be available and you'll be able to add it to a game and it'll play with you.Single palyer Co-Op games would probably be quite fascinating to go through, just to see how the ai figures them out.
I wonder if any of these stuff can be modified so that it actually looks like a human playing it. I mean there should be some penalty cost introduced for simply modifying a control combo, as a human would never start pressing a key for nothing if not necessary. Also there should be some reward given for not pressing anything, as it is more relaxing.
Hey Is there a github where i can check your code? what is the highest score?
Hello , is there any way for me to contact you regarding this video ?