| |
| |
| |
|
Page: 1 2 3 4 5 6
Comments:
<0> good idea :)
<0> you do that
<0> i wonder though heh
<1> i think it's early morning in israel at the moment, though
<0> so we need to move it again?
<0> i don't think it would be that good if they actually participated though
<1> oh, i meant currently
<1> i would love it if they showed up
<0> the discussion would be about now
<0> hmm, actually nevermind, it would probably be okay
<0> since we'll prepare ourselves
<1> ehh, if they don't show up, they don't show up
<0> was thinking read&go
<2> chessguy: it's not early morning, it's just past mignight there as far as I know
<2> because last I checked we were on the same timezone. could be confused tho
<1> oh ok
<1> well, i'll tell them 8 PM GMT, they can figure it out from that
<2> that'd make it 10PM for them (but yeah, you do that)
<2> (I mean, right *now* it's a little past mignight. :-)
<0> Israel Jerusalem (UTC+2) 8/30/06 11:05 PM
<0> that would be GMT?1
<0> +1
<2> .. depends on daylight saving?
<0> including the status of that
<2> ok..
<1> hmm, i'm not sure what all to tell them
<0> GMT is UTC0 without daylights and UTC+1 with daylights, it's UTC+2 currently while GMT is UTC+1
<1> i hate time zones
<0> that we'll completely misunderstand their article and discus it with weird arguments?
<1> lol
<2> and send resulting flame war log
<2> "this is YOUR fault" "write better article next time" :-p
<0> perhaps excuse and warn in advance that in case they would show up, people might not actually respect them and their arguments
<0> stressing that it's a discussion in difference to a debate
<3> How do you plan to moderate a debate in an unmoderated channel?
<0> and perhaps challenge them to your own little battle-off while you're at it
<0> discussion
<2> Ceran: it's going to be moderated
<3> Defty: Where? Here?
<1> ok, email sent
<2> Ceran: yes
<0> in the worst case scenario, we move it to a separate channel
<3> Defty: I didn't think this channel had any ops. Who will be moderating?
<2> Ceran: someone who's given ops as needed by chanserv
<1> i hope DanF_DrC is nice to them
<3> heh, in other words, no one
<2> 22:56 -!- mode/#ai [+o AJC] by ChanServ
<0> 9/9 at 9 pm, good for our rallying, defty
<2> huhu.. why not.
<0> cause if you flip them over, you get 666, which are three 2 * 3, i.e. two primes
<1> Deep Blue, and its ospring Deeper Blue, rely mainly on brute-force methods to gain an advantage over the opponent, by traversing as deeply as possible the game tree [5]. Although these programs have achieved amazing performance levels, Noam Chomsky [6] has criticized this aspect of game-playing research as being "about as interesting as the fact that a bulldozer can lift more than some weight lift
<1> er."
<2> three perfect numbers in a row :-o
<0> i meant it's catchy in order for us to get friends to join
<3> cwenner: what in the world are you talking about?
<1> Deep Blue, and its o
<1> lol
<0> ceran: the 999->666 et.c. was a joke
<1> i like that chomsky quote
<0> do you agree with it?
<1> absolutely
<3> indeed
<1> oh wait
<3> chessguy: That paper looks interesting, that his method was able to beat experts and other world-cl*** chess programs.
<1> wow, how did i not see this before
<1> they are doing something radically different than i am
<1> well, perhaps not radically different
<1> but certainly different
<3> What are they using as their genome?
<1> a large set of genes
<3> chessguy: a-duh :P I meant what part of the program do the genes represent?
<1> for functions, they're using thinks like an If function, and boolean functions
<1> they represent parts of an evaluator
<0> i find it interesting though, since the (IMO faulty) parallel would imply machines top us in that area as well. it is interesting that bulldozers are stronger than us, the principles behind it has contributed greatly
<1> which attempts to ***ign a value to a given board position
<3> This might be applicable to Go. Much of Go is like the endgame in chess.
<1> for terminals, they're using things which look at the board and ask questions about it
<0> i don't agree with that ceran, it's nothing like the endgame, the middle game would be the most similar but even that isn't good.
<1> for example, am i in check?
<0> there's no way you can generate anything close to end-game tablebases or ***ociated methods for go
<1> as well as random numbers
<0> there is pattern matching, but that is completely different
<0> what i want to say is: why is that, how do you mean
<3> cwenner: I mean they're similar in the sense that there's a lot of empty space, resulting in deep game trees leading to the nearest tangible evaluation point.
<0> okay, for those tricky configurations i agree, if we consider the "search-strategy" in go
<0> sorry to interrupt
<0> chessguy, didn't they list the games they had applied it to? was go included
<1> in this article? no
<0> the other page you had
<3> cwenner: just backgammon, robocode, and the endgame in chess
<1> you mean this one: http://www.genetic-programming.org/hc2005/sipper.txt
<1> this really is incredible. they make all the points that i've been trying to make to you guys for months
<3> Yet they make sense ;)
<1> amazingly :)
<3> I understand now what you meant by using the board evaluator to choose the next moves.
<1> well, that approach is very different from mine
<1> they're scoring successor positions and then choosing the move that leads to the best successor position
<1> i would claim that such a scoring mechanism is very unnatural
<3> nothing about AI is natural. that's why it's called *artificial* intelligence ;)
<1> well, if we're modelling natural intelligence, we want something that looks natural :)
<1> what do we mean by AI, unles it's in terms of something we naturally see
<0> personally, i see AI as a fuzzy subset of algorithms
<3> chessguy: True, but the goal of AI is to achieve and surp*** human intelligence, not necessarily mimic it's natural form. Galileo tried to mimic the shape of birds to achieve artificial flight, but it turned out unnatural forms, such as fixed and rotary wing craft, that proved more successful.
<3> ok, his ornithopter turned out to work once it had a viable power supply, but that's besides the point ;)
<3> And I meant Da Vinci, not Galileo...
<1> stuid windows
<1> *stupid
<0> we need a log
<0> "(23:48:49) Ceran: chessguy: True, but the goal of AI is to achieve and surp*** human intelligence, not necessarily mimic it's natural form. Galileo tried to mimic the shape of birds to achieve artificial flight, but it turned out unnatural forms, such as fixed and rotary wing craft, that proved more successful."
<3> cwenner: we do have a log. info clog
<0> i've googled for it, could you link me?
<0> info clog didn't give anything shallow either, is it a term or a name?
<4> [clog] (n=nef@bespin.org): CLOG (Channel LOGger) http://tunes.org/~nef/logs/
<4> whois was the command
<5> question : Do those irc bots like ElizaXX (XX is a 2 digit number) actually exhibit any form of AI ??
<6> no
<5> i see
<6> sorry
<5> heh. people dont talk much in here?
<6> yeah im sorry
<6> well good night
<5> ?
<6> sorry to not be more active
<6> I got cl*** tomorrow
<6> night
<5> night
<5> [10:42] <6> sorry to not be more active
<5> [10:42] <6> I got cl*** tomorrow
<5> [10:42] <6> night
<5> [10:42] <5> night
<5> :p
<6> what
<5> i'm sorry.
<7> hi, How to solve n-queen problem with A*?
<4> how do you solve n-queen with search?
<7> how?
<4> yes, what do you have so far...
<7> i don't understand you....
<7> you mean from where I start?
<4> well, to solve n-queen with A*, you first have to be able to solve the same problem with a normal search
<4> can you do that?
<7> I know the basic solution with divide and reminders...
<4> so, at some stage while looping, you pick a next "board configuration" to search
<7> yes
<4> now, you need a measure of how far you've got
<4> i'm presuming that can be the number of queens already lined up, as a floating point number
<4> not lined up, different problem, queens already placed validly without overlap
<7> wait...
<7> first we need to agree on h(n)
<7> how we buid it?
<7> build
Return to
#ai
or
Go to some related
logs:
#debian
uninstall ara.ini
#ubuntu
#linux
#css
md5sum encript
rsync no encryption
t
(111) Connection refused ASP.NET
#python
|
|