Advanced AI

[Demon.killer]

I believe that a unbeatable chess AI is not a AI at all but Teaching a algorithm to play a even a crapy chess game is a really advance AI . what do u all think ?

I gonna a Start programming a algorithm like that - but I don't think i am gonna go any where :P

[rogster001]

well i don't really agree, becuse it would be easy just to program a computer to play any old game of chess, i could do it now and i'm rubbish!
You would only have to
1 select a piece from your set at random,
2 make a legal move in some random direction
3 do same for opposing pieces or input a user move by opposing pieces.
4 repeat.
This is just like somebody that does know the rules for legal moves but does not know anything about winning the game, playing while wearing a blindfold

[CornedBee]

But that's not a learning algorithm then, is it?

The definition of AI is rather vague.

[anon]

I think people initially thought that a computer winning a good human player in chess would be a major breakthrough in the field of AI. Then it turned out that you get a stronger and faster computer player if you just let the computer do what it is best at - mindless number crunching.

As to your idea, how are you planning to do that?


You would only have to
1 select a piece from your set at random,
2 make a legal move in some random direction


How about
1. select a random legal move (consulting with a database, though)
2. store the position and the move in a database: if this move leads to a major worsening of the position (next moves or a number of moves later), don't use it again in the same situation; if it improves the position, always make this move in this position (as long as the opponent doesn't devise a winning strategy).

Now this might give it an appearance that the AI is actually learning.

(What if you let two AIs play each other like that - would they become exceptional players? :p)

[CornedBee]

Possibly, but since they're starting at random, the number of possible positions is so absurdly high that it's not feasible to store them.

[691175002]

An AI that simply replies with an identical canned response based on the positions of the pieces cannot be called intelligent, and would probably not be very effective either.

[@nthony]

I agree with the bee: when does a learning "algorithm" stop being just an algorithm and start becoming intelligent? Our notion of AI is too vague and undefined. I believe it stems from our belligerent miscategorization of it as being "artificial"; in contrast, when an algorithm uses calculus techniques to solve a system of equations, is this deemed to be "artificial mathematics"?

[anon]

An AI that simply replies with an identical canned response based on the positions of the pieces cannot be called intelligent, and would probably not be very effective either.

My suggestion was not quite serious, but if it would pick from a collection of more or less equal moves, and the evaluation of positions would change over time?

I wonder how that would work for something simpler, like Tic-Tac-Toe.

[Demon.killer]

making a cpu to think and act like a human would be nearly impossible i.e . for now at least.

So best way is - calculating every possible move and making a move that has higher chance of winning and lower chance of losing. and also record ever step.
if the cpu loses then it stores that in the HD and then when the next game is played It refers to that file and makes sure that it doesn't do that mistake again ( which would be the hardest part ).

I am gonna give it a shot ;)


As for as humans go - we develop our skills by trying to copy other humans action. i.e. if u laugh before a baby it laughs too. I have noticed it every time but at the the same time if u raise ur voice a bit the baby gets scared with cannot be archived by simple if staements ;).
humans rock ! :D

[abachler]

The term artifical applies to AI in that it necessarily distibnguishes it from intelligence that is not created by man, i.e. natural intelligence. Since mathematics are entirely artificial it is not necessary or useful to apply the modifier artificial to that field. Mathematics does not exist in nature. Mathematics is merely a system with which we can describe nature.

[anon]

I don't however believe that the term "unbeatable AI" is quite correct. Chess programs are technically beatable alright, since chess cannot be solved to the end (yet). And for human players, chess programs let you set the playing strenght - noone likes to lose all the time. I've even seen a program with a particular mode where it would make intentionally weak moves occasionally, let you know of that and then tell you after your move whether you managed to take advantage of it.

A lot of intelligence comes from position evaluation. The decision how good or bad a position is involves a lot more than material count.

As for playing by the book (computers use databases for openings) it is the same thing with human players. A strong player needs to know a huge number of openings. If your opponent knows exactly what they are doing in the opening phase and you don't, you are at a great disadvantage. Unless you are extremely talented, you'll most likely never achieve a position where you might hope to gain the upper hand by intelligence only.

In addition a strong player needs to be familiar with a large number of tactical patterns and strategical positions. This is, however, where some fuzzy pattern matching is required because you won't neccessarily get the exact same positions described in the books.

[rogster001]

Teaching a algorithm to play a even a crapy chess game is a really advance AI . what do u all think ?



All i was saying to start with was in relation to this quote, my point being its a doddle to make the computer play legal moves with ABSOLUTELY NO strategem coded in,
thus this is not in any way even an attempt at AI yet definitely fits the description of 'even a crappy game of chess' above. it certainly would be the lousiest game ever,
Yet if you were looking at a chess board it certainly would appear someone was 'playing', just they would be wild moves making no sense to the other player, hey maybe would get a few new openings out of it tho haha!
it would be interesting to pit a mid level computer player against this random move selector, though for the random player stalemate would be the best it could hope for, a checkmate would be...well like winning the intergalactic lottery draw with a hundred number sequence from 10.000 balls or summat haha!

[Demon.killer]

Lol I have no passion for playing chess my self :P . there fore thats what I do. I just make a most random ever possible. the rule I keep is not to let the king be alone. of course ppl with strategy own me badly but if the other player was poor like me it will take a long time to End ;)

anon: I didn't meant that to be taken literally :S

Btw I started the Chess AI today :)
its gonna a written in a Mix of C, ASM(mostly) , vb.net (Lazy to create a GUI in C)

[anon]

I remember playing chess with the TV console ("Chessmaster"?) which had several handicapped playing styles for the computer. If I'm not mistaken, "drunken" was one of the personalities and that basically followed rogster's strategy. As this is very similar to how complete beginners play, then you might say it was the most human-like AI in the program. Except it never got sober.

[@nthony]

The term artifical applies to AI in that it necessarily distibnguishes it from intelligence that is not created by man, i.e. natural intelligence.As natural as genetically enhanced offspring, or shall we call them "artificial babies" now too? Currently this is an easy distinguishment to make, but in the (possibly near) future it will become increasingly harder, so the better we ween ourselves off it today and stop encouraging its use, the better our ability to grasp the concept of intelligence (and not just the kind that pertains solely to a mass of organic tissue, as many homo-centric philosphers will have you believe).
Granted, we are no where near this point, and perhaps the term "artificial" reminds us of this, but being all of scientific minds, we know that a Turing machine does not have to exist before we can begin to discuss its nature.
Mathematics does not exist in nature.fractals, magic numbers, golden ratios..., just because we give nature a name and symbols doesn't make it any less natural or any more artificial ( a rose by any other name ...) :D

[abachler]

As natural as genetically enhanced offspring, or shall we call them "artificial babies" now too?
While this is a technically correct term, I think the use of GM offspring is the preferred specification since they are not (yet) completely artificial. In contrast, current AI is completely of human manufacture. Aside from the fact that current GM babies are rather limited to gene therapy, which does not enter the genome, and to offspring that most people would not call babies.

Just because a term makes you uncomfortable does not make it incorrect. You are trying to force two classifications under a single broad term when they in fact have very little in common. What we refer to as artificial intelligence has no more in common with human or animal intelligence than the fact that it is a decision making process. The similarities end there. It is not self aware on any level whatsoever. It is ONLY an algorithm that models optimal behavior. Current and forseeable methods of AI have absolutely no potential to achieve conciousness by any but the broadest and most useless definition.

[@nthony]

It is ONLY an algorithm that models optimal behavior.And what exactly is the human thought process then, if not exactly that? So if we can deem a human's algorithm "intelligence" yet a computer's "artificial" at which point exactly does the algorithm become just that? When it is run on neurons instead of semi-conductors? You are attempting to split an abstract category such as intelligence merely on whether or not its implemented in flesh or flash.
And just for the record, I currently have in my possesion an AI algorithm that will not only match but exceed the "optimizing" ability of a human... I just need the hardware to run it ;)

[oogabooga]

The term "artificial intelligence" is meaningless. "Computational intelligence" is better since it indicates the central assumption being made, that intelligence is computable.

Of course, people brainwashed by popular pandering junk like "The Emperor's New Mind" (for example) will remain in a perpetual state of denial.

[abachler]

So if we can deem a human's algorithm "intelligence" yet a computer's "artificial" at which point exactly does the algorithm become just that?

At the point where it is created by the hand of a natural being, that is what detrmines its state of being artificial. It is an artifact. That makes it artificial.

Artificial (http://en.wiktionary.org/wiki/artificial)

The adjective artificial in artificial intelligence refers to the primary (first) definition, not the connotative definition.

[Mario F.]

And what exactly is the human thought process then, if not exactly that? So if we can deem a human's algorithm "intelligence" yet a computer's "artificial" at which point exactly does the algorithm become just that? When it is run on neurons instead of semi-conductors?

Just because the term AI is vague, it doesn't mean we can't make assumptions about what it is not. What is probably difficult is defining what it is.

AI is not a cup of tea. AI is not a chicken. AI is not the human brain capabilities. Consequently, I think it is pretty safe to assume that computers aren't intelligent.

At least not for now.

[bradszy]

You guys make my head asplode.

[abachler]

In which way -

1. A mind expanding experience without limit

2. Deconstructive expansion due to built up pressure.

[Brad0407]

Might as well define artificial intelligence in 4 different ways:

Humanly:
1. Thinking like humans. - The thought process to arrive at a conclusion should be like a human's thought process. The conclusion may be different. Ex. if the thought process is picking a random number between 1 and 5, the conclusion could be different but both used the same thought process

2. Acting like humans - Thought process is not as important. It is important that the intelligent "agent" arrives at the same conclusion as a human.

Rationally:
3. Thinking rationally - the thought process must be ration or in otherwords try to produce the best conclusion based on the given information. This generally means that the agent uses logic and reasoning to get to the conclusion.

4. Acting rationally - The agent must arrive at the rational conclusion. The thought process to get to the most rational conclusion doesn't matter.

All these definitions have pros and cons. All have areas of thinking where they are most helpful. In this situation, I think the 3rd definition is most applicable because the O.P. is concerned not with whether the computer makes the correct moves, but that it learns from them. Other people might like the 1st definition because they like to compare the algorithm to how humans think.

[daltore]

Personally, I think we can define "intelligence" as the being performing actions that it's original programming did not tell it to do. So, if you program a robot with a camera to get from point A to point B without hitting anything, you put it on a track with some obstacles, and it weaves in and out of the objects, that's autonomy, not intelligence. However, if that robot decides to get off the track and go around to the other side, that's intelligence. Feel free to debate me on this.

[Brad0407]

Personally, I think we can define "intelligence" as the being performing actions that it's original programming did not tell it to do. So, if you program a robot with a camera to get from point A to point B without hitting anything, you put it on a track with some obstacles, and it weaves in and out of the objects, that's autonomy, not intelligence. However, if that robot decides to get off the track and go around to the other side, that's intelligence. Feel free to debate me on this.
I think what you're describing is free will. And I also think that there's a difference between free will. If a chess player plays a great game of chess, we would call him intelligent. It doesn't matter whether he chose to play chess or was forced to play. On the other hand, I know some people that do some pretty dumb things on their own free will.

Also, I think it takes at least a small amount of intelligence to avoid obstacles. It has to make "decisions" about its path to get from A to B. The more intelligent it is the better decisions it will make.

And if it's programmed goal is to get from A to B then we could say it "wants" to get to B. One definition of free will is a second order desire. Which is to say whether it wants to want to get to B. If the robot has second order desires (or goals), then it could chose to not want to want to get to B. It could chose to want to want to go off course. In other words, it could want to set a new goal. It may or may not be capable of changing its own goals. Of course, this type of desire would have to be programmed into the robot.

Which leads me to my final challenge to your argument. Robots never disobey their programming. So a robot could never perform an action that its original programming did not tell it to do. And by that definition, no computer could ever be intelligent. I hope that's not the case because someday, robots will outperform humans in every aspect of our "intelligence". And I won't be able to say that I can do anything that can't be done by a "stupid" robot except walk off course and fail.

[abachler]

The concept youa re both dancing around is called emergent behavior. Discussing things like free will with regard to AI is like discussing the origin of the universe when you haven't mastered fire yet. But of course historically that is exactly what humans do, so my hope is that sentient AI will definately not think like humans.

AI is not sentient, it has no free will, it's creation and destruction carries no greater moral dilemna than buying a new lawnmower or throwing the old one away. If you don't believe me then I suggest you actually read books on the subject, and not the holier than thou philosphy of AI crap, the real ones that get into the math and science of AI. You will quickly realize that there is no 'intelligence' involved.

[CornedBee]

Perhaps you could discuss free will in humans before bothering with free will in machines. Or you could argue whether humans aren't a sort of machine, too.

[Brad0407]

To abachler:
I think everyone agrees that AI has no free will and some people think that humans don't have free will (hard determinism). What are you arguing? I'm sorry, your post just seems off topic and your logic is scattered. Last time I checked, we were talking about the definition of intelligence. What is your definition of it?

To: CornedBee
My point was that robots don't need to have free will to be intelligent. Sounds to me like you think humans might not have free-will either(which is definitely possible). So if humans don't have free-will but humans are definitely intelligent, then would you agree that free-will isn't necessary for intelligence?

[abachler]

The OP was arguing against the term aritifical as applied to artifical intelligence, not whether machines are intelligent. Now you seem to want to argue a different point, perhaps you should start a new thread. In any case, unless your definition of intelligence is so broad as to include anything that has mass and takes up space, AI is not intelligence. The term started as a misnomer. Unfortunately it has become prevalent and changing it now just seems a trite exercise in political correctness. I don't have to define intelligence in order to say that machines are not intelligent any more than I have to define a computer to say that a rock isn't a computer. For anyone that works with AI professionally (as I do) it is so blatantly obvious that AI is not intelligence that when fanboy's argue about free will etc etc blah blah ad nausium, it just makes me question the future of the human gene pool. Its liek tryign to argue that a calulator is intelligent because it can mimic human behavior by adding two numbers together.

[Brad0407]

Calm down there, buddy. I see your point. There are 2 main objections that I have but I can see where you're coming from. It's not a big deal anyways.

[robwhit]

The OP was arguing against the term aritifical as applied to artifical intelligence, not whether machines are intelligent.I think that was a different thread.

[abachler]

I think that was a different thread.

<shame>

You are correct, I misremembered the thread name. Please accept my apologies.

</shame>

It is pointless to argue this matter, as when AI is actually intelligent, it will be able to make its own argument, until then its a mute point.

[rogster001]

Perhaps you could discuss free will in humans before bothering with free will in machines. Or you could argue whether humans aren't a sort of machine, too.

A good point, Prof Richard Dawkins of Oxford University wrote 'The Selfish Gene' which is more or less about that very idea, including the whole animal kingdom also, living things are shown to be 'survival machines' built by genes, their forms arising from survival strategy evolved through environmental pressures.
it is a most thought provoking work by a big academic hitter, and controversial at publication. the author also writes code including the evolutionary model 'The Blind Watchmaker'.

[abachler]

That text is so outdated, Im surprised people still read it. But then there is another far older text that is full of outdated info that people still read, so I guess its human nature to believe in what you want, not in what is.

[Mario F.]

That text is so outdated, Im surprised people still read it. But then there is another far older text that is full of outdated info that people still read, so I guess its human nature to believe in what you want, not in what is.

You really are incredible. I don't want to seem I'm after everything you say.
But jebus! Almost everything you say is so incredibly stupid, the selfish gene in me forces me to reply.

You have the slightest idea how incredibly ignorant that statement of yours is? I mean, c'mon! By the same account Darwin's theory of evolution is a fossil everybody should burn in the fireplace.

The book in question is to this date one of the most read books in the subject. It keeps providing grounds for lots of debate and it's, despite one agreeing or not with its contents, a center piece in the contribution to the human knowledge of evolution.

Shutup!

[rogster001]

That text is so outdated, Im surprised people still read it. But then there is another far older text that is full of outdated info that people still read, so I guess its human nature to believe in what you want, not in what is.

I have to agree, what a foolish comment that was, and reflecting MarioF's comment the author himself freely admits to publishing the work 'warts and all' in the foreword of late editions (If abalacher had read it perhaps he would know that) and heavily references entries with new work and research and thoughts on the opinions, which gives a great view on how the authors views have 'evolved' or remained static on some of his own arguments

[abachler]

You are mistaken. I read the book not long after it was originally published.