Re:Beating the World’s Best at Super Smash Bros. with Deep R...
by dyselon on Thursday February 23, 2017 @01:22AM
(#13712048)
Can I ask what the feature set looked like? I always kind of wanted to do this with the Skullgirls AI, but never had the time while we were developing it. As a developer, I obviously had full access to the game state, but I'm still not really sure what the best way to represent that state to a neural network is.
Re:Beating the World’s Best at Super Smash Bros. with Deep R...
by vladfi1 on Friday February 24, 2017 @03:03AM
(#13721904)
It was just basic stuff like player positions, velocities, and animation states.
Re:Beating the World’s Best at Super Smash Bros. with Deep R...
by Blackthorn on Thursday February 23, 2017 @09:17AM
(#13713997)
Getting them from RAM instead of the screen doesn't give you an advantage on (for example) DI or ledge teching?
Plus, our bot doesn't have any clue about projectiles. We don't know where they live in memory, so the network doesn't get to know about them at all.