Re: Game Playing: Chess vs. Go

[Subbarao Kambhampati <rao@asu.edu>]

Yes--GO is still quite beyond computers (because of the
massive branching factor 
involved even a small look ahead turns out to be infeasible apparently).

I am going to mention it as part of "state of the art"
discussion at the end of the
adversarial search lectures (which is likely on next Tuesday)

Rao


At 10:06 PM 9/25/2003, you wrote:


> Dr. Kambhampati,
>
> First I wanted to say that today's lecture was the most enjoyable
> yet.  I liked 
> the talk on chess and game playing, but I was a little disappointed that
> you 
> didn't mention a new goal for AI game agent: building a go (also called
> weiqi 
> in chinese) playing program.  The book mentions it in passing, and
> it's kind of 
> curious that only the Japanese seem interested in funding a serious
> effort into 
> Go AI research.  IMHO, the simplicity and the elegance of the rules
> of Go, make 
> for a much more interesting and challenging topic than chess (not to
> disparage 
> chess fans).
>
> There's an article that I found that talks about it from an AI
> perspective.  
> The section of Go AI complexity seems to complement our current
> discussion 
> fairly well.
>
> http://ai-depot.com/LogicGames/Go.html
>
> Here is a websit on Computer Go.  Apparently, the strongest Go
> playing program 
> out there is ranked at 6 kyu, the equivalent to an experienced
> amateur.  
> According to the author, it accomplishes this by using a pattern
> database.
>
> http://www.reiss.demon.co.uk/webgo/compgo.htm
>
> Regards,
> Yicheng