Artificial Intelligence

[Illuminattile]

Do you ever think about artificial intelligence? It's usually the stuff of science fiction and conspiracy theory, but do you ever wonder how far we could take it?

IBM built a computer capable of beating the reigning World Chess Champion, a computer capable of learning and adapting its strategy based on experience and its opponent. Once defeated, Garry Kasparov became obsessed with how intelligently and humanly the machine played, accusing IBM of cheating.

Can we really build a computer capable of human intelligence, so-called "Strong" AI?

If "strong" AI is possible, and we create computers that can truly reason, does that make it a superior species? What would this mean for society? If computers were capable of thought, could they not replace humans in almost any job? Would the world be better off if computers - machines of objectivity, reason and logic - ruled it?

What about art? Is it possible to program a computer to understand what it is we enjoy about a picture or a piece of music and then create it?

How many question marks can one person use in a post? Will any of these questions be answered?

Anyway, your thoughts?

 

[The.Menace]

I don't think there'll be a real AI. To play a chess game is one story, but real AI in real life situations is a different story. I can't think of a way to make a computer understand feelings etc etc.

 

[Duke]

I don't think there'll be a real AI. To play a chess game is one story, but real AI in real life situations is a different story. I can't think of a way to make a computer understand feelings etc etc.

Not yet at least.

But basically you're right. We can build an AI capable of beating humans at chess. Okay, but that same machine can't do the dishes (random example)

What i'm saying is that, at least for now, a really good AI will also be very limited in what it can do.


But the questions you asked, Illumintatille, are probably very relevant in, say, 50 or 100 years from now.

Computers came a long way in a short period of time...and they're far from finished.

 

[CalcuoCuchicheo]

IBM built a computer capable of beating the reigning World Chess Champion, a computer capable of learning and adapting its strategy based on experience and its opponent. Once defeated, Garry Kasparov became obsessed with how intelligently and humanly the machine played, accusing IBM of cheating.

I know you acknowledged the suspicion surrounded this "win", but I don't think that balances out the emboldened. If the accusations are true, it was a human victory.

 

[Illuminattile]

I know you acknowledged the suspicion surrounded this "win", but I don't think that balances out the emboldened. If the accusations are true, it was a human victory.

To be honest, I think Kasparov is just paranoid. Bear in mind that Deep Blue beat him in one game in 1996, but lost the overall match. It was only after Deeper Blue beat Kasparov over a six game match in 1997 that he accused them of cheating. The improvements between Deep Blue and Deeper Blue were minimal.

The fact that IBM programmers updated the computer between games to better counteract Kasparov's playing style might be seen by some as unfair (it was well within the rules), but this was eight years ago. Deep Thought - and all other chess computers like the three that comprehensively beat the team Topalov, Ponomariov and Karjakin in 2004 - are based primarily on "brute force" rather than strategy.

Chess is just an example of where machines can operate more quickly and infallibly than humans.

 

[bigmack]

Do you ever think about artificial intelligence? It's usually the stuff of science fiction and conspiracy theory, but do you ever wonder how far we could take it?

IBM built a computer capable of beating the reigning World Chess Champion, a computer capable of learning and adapting its strategy based on experience and its opponent. Once defeated, Garry Kasparov became obsessed with how intelligently and humanly the machine played, accusing IBM of cheating.

Can we really build a computer capable of human intelligence, so-called "Strong" AI?

If "strong" AI is possible, and we create computers that can truly reason, does that make it a superior species? What would this mean for society? If computers were capable of thought, could they not replace humans in almost any job? Would the world be better off if computers - machines of objectivity, reason and logic - ruled it?

What about art? Is it possible to program a computer to understand what it is we enjoy about a picture or a piece of music and then create it?

How many question marks can one person use in a post? Will any of these questions be answered?

Anyway, your thoughts?

good thread.

I believe in the future AI will be very advanced. Yes I've heard that in the future it's possible for computers to have feelings.

When computers can have feelings,can think on their own and make decisions,then that will be the day humans become obsolete.

Is it a good thing? No

What if every single job gets automated by computers,then we get computers that fix the other computers that break down. Then people wont have jobs anymore. Then everything will basically be free,u can get anything you want for free. when that happens the world will fall apart. life wont be fun anymore.would u want to chill and drink and smoke up and fuck around for the rest of your life 24hrs a day 7/week? whats the purpose of life then?

no more schools,no more jobs,the kids will end up retarded.no one will reproduce anymore,or if they do theyll probably kill their kids. then the human will become extinct and the world will be run and owned by robots.

then when the resources finish,thats the end of the world

 

[Rukas]

To be honest, I think Kasparov is just paranoid. Bear in mind that Deep Blue beat him in one game in 1996, but lost the overall match. It was only after Deeper Blue beat Kasparov over a six game match in 1997 that he accused them of cheating. The improvements between Deep Blue and Deeper Blue were minimal.

The fact that IBM programmers updated the computer between games to better counteract Kasparov's playing style might be seen by some as unfair (it was well within the rules), but this was eight years ago. Deep Thought - and all other chess computers like the three that comprehensively beat the team Topalov, Ponomariov and Karjakin in 2004 - are based primarily on "brute force" rather than strategy.

Chess is just an example of where machines can operate more quickly and infallibly than humans.

This IS NOT AI though. This is simply a program that can crunch huge amounts of data at once. For example, each move has a possible outcome, which has a possible outcome, which has a possible outcome, etc etc. The computer calculates all of these, and chooses the best one to counteract and win the current match.

This isnt AI, Deep Blue will never learn to play scrabble, or even want to, it will only ever be able to play chess, because it is a chess simulator, not true AI.


As for the possibility of AI, sure its possible. Its totally possible, all it would take he creating a code (program) that could re code and adapt itself infinitely. Now, thats easier said than done, coding is difficult and there are a lot of variables that if introduced could kill the program. But, if you managed to write a programming language that could reprogram itself and adapt itself and "evolve," AI is totally possible. Once this code is written you dont even need to program anything into it at all. You program the essentials, the ability to program and upgrade and evolve itself, and leave it running. And like a new born it will learn. It will see that it cant do certain things, like a baby learns not to touch hot things, and then it wont program these things in, and it will learn its short comings in other aspects and program in counter balances.

It is possible, and it is coming.

Fundamentally if you are religious or believe in some sort of creator, then human life is artificial intelligence.

 

[Illuminattile]

This IS NOT AI though. This is simply a program that can crunch huge amounts of data at once. For example, each move has a possible outcome, which has a possible outcome, which has a possible outcome, etc etc. The computer calculates all of these, and chooses the best one to counteract and win the current match.

This isnt AI, Deep Blue will never learn to play scrabble, or even want to, it will only ever be able to play chess, because it is a chess simulator, not true AI.

This IS AI. Playing chess is essentially a problem solving exercise, and as such requires a level of intelligence. Chess isn't an objective game. You cannot look at two possible moves and absolutely decide which is better. Deep Blue analysed thousands of chess games to understand the strategy of chess; how important it is to control the centre of the board, how important a safe king is etc.) to essentially learn which of all the possible moves was best.

People tend to think of AI in terms of things that haven't been done yet. They don't see AI as something which exists and is in practical use. Computer chess is AI, voice recognition software is AI, even online diagnosis websites are AI. They are all limited examples of artificial intelligence, but they ARE artificial intelligence.

 

[Rukas]

This IS AI. Playing chess is essentially a problem solving exercise, and as such requires a level of intelligence. Chess isn't an objective game. You cannot look at two possible moves and absolutely decide which is better. Deep Blue analysed thousands of chess games to understand the strategy of chess; how important it is to control the centre of the board, how important a safe king is etc.) to essentially learn which of all the possible moves was best.

People tend to think of AI in terms of things that haven't been done yet. They don't see AI as something which exists and is in practical use. Computer chess is AI, voice recognition software is AI, even online diagnosis websites are AI. They are all limited examples of artificial intelligence, but they ARE artificial intelligence.

I disagree. Those programs are limited to a set of data boundaries provided to them, they only follow one overall algorithm.

You’re trying to cheat what this thread is about, I see too much of this happening in WOW and its one of the reasons I don’t post anymore. Instead of everyone discussing the thread subject everyone tries to outsmart one another.

By definition you are correct, but that’s a very limited sneaky way to look at this topic in a means to “out smart” others. Why make this thread if you’re only going to shut down any discussion that doesn’t follow what you already think?

“The ability of a computer or other machine to perform those activities that are normally thought to require intelligence.”

That is an accepted definition for what artificial intelligence is, but by that definition, a calculator is AI as it performs an activity that normally requires intelligence. A watch is AI as it keeps time, which normally requires intelligence. In the same way, albeit more complex, Deep Blue plays chess.

However, that is not what true AI that is discussed in “science fiction” is about, and you know that.

Artificial intelligence is self thinking and evolving programs, not algorithms that have been programmed to do a specific job, whether add numbers or play chess.

True artificial intelligence must be self aware and have the ability to learn, think and evolve; Deep blue is not self aware. Deep Blue can not learn, it can only interpret data it is provided with. Deep Blue can not think, it can only calculate based on written algorithms and data it is provided with. Deep Blue can not evolve, it will never play anything but chess.

 

[The.Menace]

But, if you managed to write a programming language that could reprogram itself and adapt itself and "evolve,

well I don't believe that this is possible and this is what it comes down in the end: Believe. Noone knows yet how far AI can go, it comes down to believe.

 

[Illuminattile]

I disagree. Those programs are limited to a set of data boundaries provided to them, they only follow one overall algorithm.

You’re trying to cheat what this thread is about, I see too much of this happening in WOW and its one of the reasons I don’t post anymore. Instead of everyone discussing the thread subject everyone tries to outsmart one another.

By definition you are correct, but that’s a very limited sneaky way to look at this topic in a means to “out smart” others. Why make this thread if you’re only going to shut down any discussion that doesn’t follow what you already think?

“The ability of a computer or other machine to perform those activities that are normally thought to require intelligence.”

That is an accepted definition for what artificial intelligence is, but by that definition, a calculator is AI as it performs an activity that normally requires intelligence. A watch is AI as it keeps time, which normally requires intelligence. In the same way, albeit more complex, Deep Blue plays chess.

However, that is not what true AI that is discussed in “science fiction” is about, and you know that.

Artificial intelligence is self thinking and evolving programs, not algorithms that have been programmed to do a specific job, whether add numbers or play chess.

True artificial intelligence must be self aware and have the ability to learn, think and evolve; Deep blue is not self aware. Deep Blue can not learn, it can only interpret data it is provided with. Deep Blue can not think, it can only calculate based on written algorithms and data it is provided with. Deep Blue can not evolve, it will never play anything but chess.

The type of AI in chess computers and voice recognition software (soft AI) can pave the way for strong AI which can do the things you're talking about. I was simply using Deep Blue of an example of where Artificial Intelligence is today. If the level of AI remained there, it would make very little impact in society. Deep Blue can learn, though. It remembers moves it makes and the results, so that it doesn't make the same mistakes again. Obviously this is a very simple form of AI, and no it's not capable of "thinking" for itself, evolving or learning other skills. Human-level intelligence is the primary aim of AI but machines like Deep Blue are steps towards the main aim.

All machines have to be initially programmed, given information and parameters to work within. It's the same with humans. A human being doesn't inherently know how to play chess, it has to be told the rules and know some basic strategy. If someone were to program Deep Blue with the rules and strategies of Scrabble, it could probably play that.

 

[bigmack]

well I don't believe that this is possible and this is what it comes down in the end: Believe. Noone knows yet how far AI can go, it comes down to believe.

dont be too sure that its impossible. anything is possible these days. i heard pretty soon you can download and save your brain on a computer.maybe they can use that kind of technology for the AI,who knows.

 

[Bender]

I'm all for AI as long as their IQ is kept below that of humans- although it might be good to have an 'Einstein' computer that you could ask it a question and it would know the answer, like the genie I had on my pc.

AI robots would mean the end of boring jobs, and more time for our leisure activities. Until then it is just science fiction, and something Stephen King writes about, but he writes horror so ...

 

[Kosen]

they cant even create computer programs that generate TRUE random numbers.. how the fuck are they goin to come up with true AI ?

 

[The.Menace]

hahaha, GREAT point lol ..... hahaha, great.

 

[Rukas]

The type of AI in chess computers and voice recognition software (soft AI) can pave the way for strong AI which can do the things you're talking about. I was simply using Deep Blue of an example of where Artificial Intelligence is today. If the level of AI remained there, it would make very little impact in society. Deep Blue can learn, though. It remembers moves it makes and the results, so that it doesn't make the same mistakes again. Obviously this is a very simple form of AI, and no it's not capable of "thinking" for itself, evolving or learning other skills. Human-level intelligence is the primary aim of AI but machines like Deep Blue are steps towards the main aim.

All machines have to be initially programmed, given information and parameters to work within. It's the same with humans. A human being doesn't inherently know how to play chess, it has to be told the rules and know some basic strategy. If someone were to program Deep Blue with the rules and strategies of Scrabble, it could probably play that.

Sure it could play scrabble is someone programed the instructions in, but it could never learn it on its own.

they cant even create computer programs that generate TRUE random numbers.. how the fuck are they goin to come up with true AI ?

Can you please explain what you mean?

 

[Default MX]

Well, the way I see AI is that it's preprogramed, but chooses the best outcome for a certain situation. One example of this is the computer player in Starcraft/WarcraftIII, depending on what units you make, or what units you place where, the computer compensates, and attempts to counterattack.

Now, most of this is programed in a Rock Paper Scissiors manner, as in "Unit X Defeats Unit Y while Unit C Defeats Unit X."


It really varies on you'r definition of AI.

 

[Kosen]

Can you please explain what you mean?

The way computers generate "random" numbers right now isnt really random. Its just a long piece of data that spits out numbers, so theoritically, at the end of the list of data, the cycle of random numbers would repeat. Many slot machines are exploited because of this, as there are trained gamblers who know how to spot what 'random' number the computer will generate next. One of my professors had a friend who did this, and he won something like 3 jackpots in 30 minutes( he was arrestted).

Thus, these numbers arent truly random, rather they are pre programmed. Alot of money has been spent on trying to develop a truly random number generator since the 1960's, but nothing has come close.

 

[Rukas]

The way computers generate "random" numbers right now isnt really random. Its just a long piece of data that spits out numbers, so theoritically, at the end of the list of data, the cycle of random numbers would repeat. Many slot machines are exploited because of this, as there are trained gamblers who know how to spot what 'random' number the computer will generate next. One of my professors had a friend who did this, and he won something like 3 jackpots in 30 minutes( he was arrestted).

Thus, these numbers arent truly random, rather they are pre programmed. Alot of money has been spent on trying to develop a truly random number generator since the 1960's, but nothing has come close.

I thought thats what you meant but wasnt sure. That is true, any code written to generate random numbers generates numbers at random from a given data set, or array of numbers. If the number is not present in that array, it wont be selected.

But, thats what the human brain does too. Think of a number, sure its at random, but its already in your head, you cannot randomly think of a number if you dont know it already.

 

[Illuminattile]

Since there's an unlimited number of numbers it's practically impossible for you to program a computer to pick one of them at random. Computers don't have enough memory.