steve uurtamo said:
> i wonder what is known about the set of unconditionally
> dead and unconditionally living groups. there must be
> something like a small and extremely fast mechanism for
> this. what is everyone using? i mean a mechanism that
> is independent of any fancy data structure tha
> How do you score Japanese
> games correctly in an automated way in the face of program disputes?
You let the disputant resume playing until the difficulty is resolved,
with his competitor given the option of responding whenever he thinks
necessary. Hence my suggestion as a way of resolving matte
Lars said:
> I had build an Monte-Carlo GO-Engine (GOMonCy) wich uses the Japanese
> scoring system. It reached a win rate against GnuGO 3.6 level 10 of
> stable 50%-52%. I used territorry-statistics about the Monte-Carlo
> outcomes. You get a probability for every field telling you who is the
> o
Jason House said:
> What about seki situations?
>
> On Nov 5, 2007 1:41 PM, <[EMAIL PROTECTED]> wrote:
>
>> It takes some tricky analysis to work out the Japanese score, due to
>> uncertainty about life/death; likewise it's not easy for a program to
>> recognize when moving is no longer to its adva
It takes some tricky analysis to work out the Japanese score, due to
uncertainty about life/death; likewise it's not easy for a program to
recognize when moving is no longer to its advantage.
How about bringing in a Monte Carlo routine after both players have
passed?--as a scoring referree, set to
1) The strongest players I've known personally have been extremely
articulate about what they were doing & why; one used to practice
explaining every move he made to his opponent.
2) I'm not that strong, myself--so you may take this with suitable
seasoning. I think the main advantage of personal g
Joshua Shriver said:
> Was reading a page about Go and came across this term. Anyone know what
> it means?
> Some googling yielded that it's some kind of tactic position. Though I
> might have misinterpreted it.
Essentially, a sharp move. If we used the word in chess, moves that formed
pins or kn
1) If we're sorting bitmaps into categories (for deciding on the next
move), the sorting will be most efficient when we can ask questions with
probability of 1/2 of "true" or "false," as in playing a sort of "Twenty
Questions."
[
These bitmaps wouldn't be necessarily maps of stones on the boar
> So even though you the playout agent has only 50% probability of
> playing correctly, the probability that after 2 plys the position is
> still won is 75%!
going toward a limit of 66.6% as the number of plies increases
-
This email was sent using AIS Web
Back to bitmaps with most of the bits at zero...
Let's take an extreme case, a bitmap with one out of 32 bits set. We're
assuming this map is a member of a class with a significant property which
we want to recognize.
If it were a perfectly random bitmap, the probability of it turning out
this wa
Working up subroutines to recognize patterns in bitmaps of a go game:
Most of the board will usually be zeros. The ones are significant, then,
but their significance largely depends on the big chunks of zeros between
them.
A technique used by animal brains & photo-enhancer software: edge
detectio
>> >Warning: Cynical Definition...
>> >
>> > My definition of AI is any algorithm that is new in computer
>> science.
Too much of an assumption here: That we should be designing & using
_algorithms_!
A good piece of what puts a problem/program in the realm of AI is the need
to use "heurist
>it
> has functions to turn a 0..N coord into x and y coords, functions to
> rotate points, functions to return the coord a certain distance away
> (e.g. ikken to the North, keima to the South-East).
Okay, I've got functions that detect a certain relation between
stones--and can be linked together
George Dahl said:
> FANN (http://leenissen.dk/fann/) is a great neural network library
> written in C. I don't know much about books on *programming* neural
> networks specifically, but I know of many great books on neural
> networks. I am a big fan of Bishop's Neural Networks for Pattern
> Recog
Jason House said:
> Thanks for the documentation. I have a few questions.
>
> Looking at only the four neighbors to detect eye-like points seems like
> it could leave many false eyes and allow captures of dangling chains.
> Is there any mechanism to avoid this problem in the play of the bot?
Eye
For computer purposes, this is the problem:
Territorial scoring is more human-convenient, can be done without filling
the dame or removing dead stones.
But it all depends on knowing which groups are live, which dead, which in
seki. If there's a disagreement, it needs to be settled by resuming the
steve uurtamo said:
>> I have read dozens of times that computer-Go is the next big
>> challenge. But in fact it is a completly amateuristic field where even
>> the most basic things are missing.
>
> one thing that it seems to have plenty of is chess programmers who are
> shocked and surprised tha
Peter Drake said:
> I think Steve meant that the move /should have been used as/ a ko
The burning of ko threats is just a narrow example of what I was talking
about: what's called "aji keshi." That is, the exchange of a threat for a
forced response, prematurely eliminating the potential for more e
steve uurtamo said:
>>There is one other issue I have seen that is similar. Sometimes
>> Lazarus will play a move that doesn't hurt nor help it's position. It's
>> not a wasted move because the opponent must respond or else lose.
>
> this sounds a good bit like a ko threat, which is tricky to dis
Peter Drake said:
> It took me a long time to get around my mental block and accept the
> advice of everyone here, but your intuition is correct: superko is so
> rare, and so expensive to detect, that you should NOT check for it on
> every move.
Depending on how a program works, you may well need
> It is clear that in
> professional play 2 handicap stones is overwhelming.
Kageyama mentioned a student who had been playing him at a small handicap
and winning. The student didn't think he could lose a game and nine
stones. So they played a nine stone game; Kageyama kicked his butt and
says the
>
> mr. yang uses the ideas of short and long
> extensions and high-low combinations in the
> beginning. (a short extension being 1 or spaces
> and a long being ideally 5 spaces). this tends to be eficient.
There is the classic Chinese rule of thumb on how far one can comfortably
extend along an e
Has anyone written anything reasonably accessible re
designing a neural net where the pulses themselves carry information (ie a
32-bit piece of bitmap, perhaps)
and/or
that information itself is used in determining whether a cell fires?
Forrest Curo
-
T
Pseudoliberties, as someone here explained recently, are a count of how
many adjacent empty spaces a program would find around a chain of stones
if it didn't bother to correct for how many times the same space gets
counted from different directions.
example
0 0 . .
X X 0 .
. X 0 .
. 0 . . The X's
[If this is redundant, please excuse me. I'm wondering if I ran into some
kind of filter the last time I sent this.(?)]
Pseudoliberties, as someone here explained recently, are a count of how
many adjacent empty spaces a program would find around a chain of stones
if it didn't bother to correct f
> Your idea is useful if it can be show to be superior in some
> way to other move generation techniques. It may be superior
> in speed or some other metric.
>
Main superiority I can think of: It becomes easy to combine conditions
like (example) "All points with one black piece above and one whit
26 matches
Mail list logo
