MarI/O – Machine Learning for Video Games • Hacker News Board

Re: MarI/O – Machine Learning for Video Games

#11

Post by jeanlucas » Sun, Jun 14, 2015, 4:39 PM UTC

nip wrote: post #6

synthmeat wrote: post #4

I have reasonable suspicion that it wouldn't be nearly as easy to train without "spinning" move which basically gives invulnerability against all enemies sans their bullets, iirc.

I see this videoclip as a great introduction to neural networks and genetic algorithms (something I didn't know existed in ML) rather than a simple "computer completes videogame level". You're probably right nonetheless (I can't remember the mechanics of this mario game) :)

Actually for Genetic Algorithms and/or Evolutionary computation it is not. :P

Re: MarI/O – Machine Learning for Video Games

#12

Post by 140am » Sun, Jun 14, 2015, 4:46 PM UTC

Veery nice visual introduction to the basics of neural networks

Re: MarI/O – Machine Learning for Video Games

#13

Post by deepnet » Sun, Jun 14, 2015, 4:59 PM UTC

Bizhawk is a SNES emulator for Game Casters.

Seth, the video author is a Super Mario World Speedrunner and Twitchcaster. http://www.twitch.tv/sethbling

Bizhawk has Lua Scripting built in : http://tasvideos.org/Bizhawk/Features.html

Seth implemented Neuro-Evolution in Lua and http://pastebin.com/ZZmSNaHX

now machine learning casting is a thing...

Re: MarI/O – Machine Learning for Video Games

#14

Post by M4v3R » Sun, Jun 14, 2015, 5:22 PM UTC

synthmeat wrote: post #4

I have reasonable suspicion that it wouldn't be nearly as easy to train without "spinning" move which basically gives invulnerability against all enemies sans their bullets, iirc.

A similar work was done for Super Mario on NES (along with other NES games): https://www.youtube.com/watch?v=xOCurBYI_gY

Re: MarI/O – Machine Learning for Video Games

#15

Post by scottfr » Sun, Jun 14, 2015, 6:33 PM UTC

This is very cool, but it isn't really "playing" the game. Instead it's "memorizing" how to complete a specific level through trial and error.

If you took this neural network and tried it on another Mario level, it would fail miserably.

Re: MarI/O – Machine Learning for Video Games

#16

Post by comex » Sun, Jun 14, 2015, 6:40 PM UTC

scottfr wrote: post #15

This is very cool, but it isn't really "playing" the game. Instead it's "memorizing" how to complete a specific level through trial and error. If you took this neural network and tried it on another Mario level, it would fail miserably.

Well, for some definition of fail miserably:

> Yeah, I think it wouldn't generalize super well, but I did try it on another easy level and it was able to get halfway through it.

- the author, at http://www.reddit.com/r/videos/comments/39qel5/top_super_mar...

Re: MarI/O – Machine Learning for Video Games

#17

Post by deepnet » Sun, Jun 14, 2015, 6:51 PM UTC

scottfr wrote: post #15

This is very cool, but it isn't really "playing" the game. Instead it's "memorizing" how to complete a specific level through trial and error. If you took this neural network and tried it on another Mario level, it would fail miserably.

The author has said that a net trained on one level can get about halfway through another level. http://www.reddit.com/r/videos/comments/39qel5/top_super_mar...

Super Mario World is very deterministic.

The Author is streaming various learning tasks on his twitch. MarI/O has been trained on 3 levels so far, for 24 hours each.

Re: MarI/O – Machine Learning for Video Games

#18

Post by JonnieCache » Sun, Jun 14, 2015, 7:01 PM UTC

Here's the author taking the world record (now superseded) at glitching mario world into the credits to "win" the game in 3:07:

https://www.youtube.com/watch?v=HxFh1CJOrTU

This is on a real console, not an emulator. They overflow the snes into the controller registers to achieve arbitrary code execution by hand. They do this by placing koopa shells at pixel-perfect locations to spell out ASM, overflow to the controller registers, and press a specific combination of buttons to jump to the sprite table and their literal shellcode.

At the start of this year there was a crazy few weeks where different speedrunners were competing to execute the "credits warp" in the quickest and most efficient way. It started at around 6 minutes, now it's down to around 2 I think, most of which is the game's intro. Fun stuff. SethBling was the first person to ever execute it on a real console.

Here's a video explaining it, from a different speedrunner: https://www.youtube.com/watch?v=vAHXK2wut_I

Re: MarI/O – Machine Learning for Video Games

#19

Post by cheald » Sun, Jun 14, 2015, 9:30 PM UTC

JonnieCache wrote: post #18

Here's the author taking the world record (now superseded) at glitching mario world into the credits to "win" the game in 3:07: https://www.youtube.com/watch?v=HxFh1CJOrTU This is on a real console, not an emulator. They overflow the snes into the controller registers to achieve arbitrary code execution by hand. They do this by placing koopa shells at pixel-perfect locations to spell out ASM, overflow to the controll…

"literal shellcode" made me chortle. What a delightfully wonderful hack.

Re: MarI/O – Machine Learning for Video Games

#20

Post by uxp100 » Sun, Jun 14, 2015, 10:22 PM UTC

I apologize, this is pedantic, but I think people often overstate the relationship between neural networks and the brain. I'm far from an expert, but saying neural networks are based on the brain is a stretch. They're inspired by a theory of how the brain works that is not wholly in line with what most scientists believe.

It's so tempting to say, this is based on the brain, when really it's more like, this was inspired by one psychologists model of the brain.