Darwinia in reality

[BrianBlessed]

That is indeed impressive. To my memory it takes incredibly complex programming just to work out which stones are dead and which are alive.

In some ways I am glad I have no opponents to use my Go board against, far too complex. Tiddlywinks is better.

[martin]

an update! again...
I'm rebuilding the engine right now to work in 3D properly, once that's done I'll start incorporating an AI interface. Any takers on helping me code it?

I'm coding an AI for the game of Go at the moment. Its horrid... AI is NOT that easy to do. Have fun

I seem to recall two things:

1) I read an article somewhere about chess playing AIs saying that they weren't a challenge to build any more, so some scientists were trying GO; if you can manage it then well done, loads of AI pros are stuck on it

2) Hopefully for the last time now:

I AM NOT BUILDING AN AI - JUST AN ENVIRONMENT!!!

[Lowell]

an update! again...
I'm rebuilding the engine right now to work in 3D properly, once that's done I'll start incorporating an AI interface. Any takers on helping me code it?

I'm coding an AI for the game of Go at the moment. Its horrid... AI is NOT that easy to do. Have fun

I seem to recall two things:

1) I read an article somewhere about chess playing AIs saying that they weren't a challenge to build any more, so some scientists were trying GO; if you can manage it then well done, loads of AI pros are stuck on it

2) Hopefully for the last time now:

I AM NOT BUILDING AN AI - JUST AN ENVIRONMENT!!!

Tiberian Sun has an AI Editor...search them out. I have had it for some time now but haven't used it much. I have some giant ants and the aud sounds for them, they have icons and use normals, they also come with "rules" to use. Maybe with their landscape editor and their AI editor you can get the skill level you want. By the way I love the game GO...and Risk...

[martin]

thing is all game AIs are usually rule sets. And I want an open source environemnt people can customize to what they want.

[Puzzlemaker]

thing is all game AIs are usually rule sets. And I want an open source environemnt people can customize to what they want.

Learn about the brain, and base the AI off the way the brain works then.

[hitm4n]

You can't do AI without rules. If you can you'll be the first to give birth to true AI. Then the world is doomed... a la all the films with "real" AI in them.

[daset]

You can't do AI without rules. If you can you'll be the first to give birth to true AI. Then the world is doomed... a la all the films with "real" AI in them.

I seriously think that if the AI ever came to life, or if humans ever made something more intelligent than themselves, they would assimilate themselves into it and make the AI the new human. Just my opinion on things.

[Darksun]

Then the world is doomed... a la all the films with "real" AI in them.

What about Asimovian AI (the godawful film starring Will 'I like to piss on Asimov's gravestone while wearing my converse allstars and driving my audi' Smith obviously not included).

[elDiablo]

So people know, its Go as in this, in case some people are getting confused. And my project is NOT going to be making a decent AI any time soon. People have trying to make a Computer Go player that will beat (half-decent) humans for over 20 years. I have 8 months. NOT gonna happen.

martin: Yes, Chess AI's are now just a matter of searching as deep as possible. Deep Blue searched to a depth of 13, while it is theorised that Grand Master's think to a depth of 12. So yes, its pretty easy to make a Chess player as the evaluation function is pretty standard and works well. People are looking at Go as no one has found a good evaluation function yet. And there a lot more moves possible in Go then there are in Chess. Go look it up, and its horrid...

You can't do AI without rules. If you can you'll be the first to give birth to true AI. Then the world is doomed... a la all the films with "real" AI in them.

Yes you can. Thats a load of bollocks. Neural Nets learn without rules. Its called Soft AI techniques. Its fun!

[martin]

Yes you can. Thats a load of bollocks. Neural Nets learn without rules. Its called Soft AI techniques. Its fun!

good man! SO have you considered an ANN or Evolutionary algorithm for Go?

[Battlefrank]

Those crazy Japanese beat you to it.
http://www.es.jamstec.go.jp/esc/eng/ESC/index.html

ARG silly japanicans!

[xander]

So, how's it coming along, martin?

xander

[martin]

So, how's it coming along, martin?

xander

good lord xander! talk about necromancy! my last post was three years ago!

anyway, this topic was buried because I gave up on both projects, the environment and go. I gave up on the environment because it was beyond me at the time, I suspect that I could just about manage it now with my experience with DirectX and XNA to help me I gave up on go because... it was boring, after playing 500 games of GO and writing down every move to analyse later I reallied that go just isn't that great a game!

[Testrie]

with my experience with DirectX

so are you still planning on using DirextX 9 when 10 is out?

eyra

[xander]

I gave up on go because... it was boring, after playing 500 games of GO and writing down every move to analyse later I reallied that go just isn't that great a game!

Now I know you don't know what you are talking about. Go is an excellent game.

xander