|
|
|
astropin
SFN Regular
USA
970 Posts |
 Posted - 09/02/2004 : 12:43:46
|
Valiant Dancer & I drifted off into a discussion on AI in another thread so I thought I would start a new thread just for it. Basically I feel AI is inevitable. Valiant claims that artificial sentient life can never exist. See the preceding discussion here:
http://www.skepticfriends.org/forum/topic.asp?TOPIC_ID=3144
|
I would rather face a cold reality than delude myself with comforting fantasies.
You are free to believe what you want to believe and I am free to ridicule you for it.
Atheism:
The result of an unbiased and rational search for the truth.
Infinitus est numerus stultorum |
|
|
Valiant Dancer
Forum Goalie
USA
4826 Posts |
Posted - 09/02/2004 : 13:05:16 [Permalink]
|
quote:
Originally posted by astropin
Valiant Dancer & I drifted off into a discussion on AI in another thread so I thought I would start a new thread just for it. Basically I feel AI is inevitable. Valiant claims that artificial sentient life can never exist. See the preceding discussion here:
http://www.skepticfriends.org/forum/topic.asp?TOPIC_ID=3144
I believe that machines can mimic sentience as far as they can be told about it, however, human behavior is made up of some unconscious thought that it is not logical to believe that humans will ever be able to produce a machine capable of it. You cannot program for things within the human subconscious which is not known about.
Computers wade through a long stream of binary logic gates. (Thanks, Alan Touring) Humans do not. The whole concepts of emotion, intuition, and instinctual reactions are so poorly understood that it cannot be coded for. AI in the measure of a completely artifical sentient being is such a complex undertaking that I do not think that humans will ever know everything necessary to make one. Science is an ever expanding knowledgebase. It is literally impossible for us to "learn it all" as a species. It is this nature of science which I have to invariably point to and say the void of the things we do not fully understand will never be filled. |
Cthulhu/Asmodeus when you're tired of voting for the lesser of two evils
Brother Cutlass of Reasoned Discussion |
 |
|
|
BigPapaSmurf
SFN Die Hard
3192 Posts |
Posted - 09/02/2004 : 13:23:56 [Permalink]
|
Perheps the AIM (Artificial int machine) could have an independent system which has limited info-sharing capabilities to represent a subconcious. (details to be determined)
Clearly a true AIM would require a new form of programming and various other new techniques. The newest anti-kasparov computer was taught a new way to think about chess, instead of thinking about every possible move and being bogged down on the third move ahead, the new machine can eliminate large groups of bad moves and not consider them. Of course it is still limited to programming but trial and error with this type of logic could have great potential.
I think we can create an AIM but to mimic it after a faulty machine like the human brain would be a mistake. We would just be asking for them to take over if we make them stupid like us. |
"...things I have neither seen nor experienced nor heard tell of from anybody else; things, what is more, that do not in fact exist and could not ever exist at all. So my readers must not believe a word I say." -Lucian on his book True History
"...They accept such things on faith alone, without any evidence. So if a fraudulent and cunning person who knows how to take advantage of a situation comes among them, he can make himself rich in a short time." -Lucian critical of early Christians c.166 AD From his book, De Morte Peregrini |
|
|
|
Skyhawk
New Member
33 Posts |
Posted - 09/02/2004 : 15:48:54 [Permalink]
|
I'd just like to point out that I watched a NASA lecture discussing the advancement of computer technology. The speaker gave graphs of the rate of MIPS (million instructions per second)increasing per year and also other forms of advancements. As it turns out, the graph are all exponential and it extrapolates that in a VERY near future there will be computers that are faster than human thought. Recently, I believe IBM is making a new supercomputer that is as fast as human thought. This is a good first step to reaching AI.
I think saying that computers are still using gates can be a misconception. We have ways to manipulate logic gates so that they have more "maybe" decisions, or more analog decisions. One example is fuzzy logic. There is also heavy research into desiging neural structures through computer science. Its extremely abstract. There is also research into genetic algorithms, etc. where it can be used to solve complex problems by making the program act like a virus. Also, some of the best virus makers are trying to manipulate computer viruses like actual viruses.
There are also supercomputers that use natural selection to solve complex problems. There was an amatuer computer engineer that wanted to patent more simplified versions of some circuit designs. The way he created these circuit designs was to have the computer run in a "natural selection" environment to pick out the best circuits and then "evolve" those circuits even further. Recently, scientists used this process to create perfect Formula One cars for certain conditions.
Instead of designing a brain to function as a human, the better approach is to work with the pieces of the brain (like brain cells) thast more simplified and have multiple pieces networked to mimic the brain. If neural research continues, with computers as fast as the human brain, and by implementing neuralogy in a natural selection process; I think in the distant future we will be able to "evolve" a brain. Thats my 2 cents on that.
|
|
|
|
Ricky
SFN Die Hard
USA
4907 Posts |
Posted - 09/02/2004 : 17:13:16 [Permalink]
|
I would just like to offer a warning: Don't try to predict the future and expect to be accurate. You really can't. We look at history and we find that they are almost always wrong. Also, don't try to say something is impossible. It was said that the speed barrier would never be broke. Guess what?
At the rate we are going, we will never get to AI. Why? Because right now, AI programming is focused on handeling certain situations. The faster the computer, the more situations it can handle, the better the "AI". At this rate, because there are really an infinite amount of situations, you would approach AI, but never really reach it.
To reach true AI, one would have to have a computer that can learn. We have the ability to mimic that, for example a computer learns your name asking you in a dialog box and you giving it. The computer can store that and use it every time it wants to refer to you. However, to be true AI, the computer would have to actually understand what a person is, and how that name applies to that specific person. Now a computer would have to learn this, not just get an object from a programmer to store data about a person.
There is one thing I think we must break through to reach AI, the language barrier. Every word is defined by other words, in a sense, when you speak one word, you are really using the entire language. For example:
"The cat is on the mat."
Well, what does "the" mean?
The: Used before singular or plural nouns and noun phrases that denote particular, specified persons or things
Ok, so what does "Used", "before", "singular", "or"...... mean? Because we can only define words based upon other words, we can't quite tell the computer what one word means. I think we need to study how a child learns to speak, and figure out if this would be able to be done on a computer.
As of right now, we are far off from that. |
Why continue? Because we must. Because we have the call. Because it is nobler to fight for rationality without winning than to give up in the face of continued defeats. Because whatever true progress humanity makes is through the rationality of the occasional individual and because any one individual we may win for the cause may do more for humanity than a hundred thousand who hug their superstitions to their breast.
- Isaac Asimov |
|
|
|
Dude
SFN Die Hard
USA
6891 Posts |
Posted - 09/02/2004 : 17:40:37 [Permalink]
|
Who's to say that an AI would have to mimic humans exactly? What we posses is the ability to simultaneously examine multiple possibilities and reach a conclusion.
If quantum computer science ever bears out the hype, we'll have computers capable of the same.
Whatever the result, it should be interesting. |
Ignorance is preferable to error; and he is less remote from the truth who believes nothing, than he who believes what is wrong.
-- Thomas Jefferson
"god :: the last refuge of a man with no answers and no argument." - G. Carlin
Hope, n.
The handmaiden of desperation; the opiate of despair; the illegible signpost on the road to perdition. ~~ da filth |
|
|
|
|
Dave W.
Info Junkie
USA
26022 Posts |
Posted - 09/02/2004 : 19:07:51 [Permalink]
|
Ricky wrote:quote:
Ok, so what does "Used", "before", "singular", "or"...... mean? Because we can only define words based upon other words, we can't quite tell the computer what one word means. I think we need to study how a child learns to speak, and figure out if this would be able to be done on a computer.
As of right now, we are far off from that.
Actually, Natural Language Processing (or Parsing) is doing quite well as a field. The problem is not that words are only defined by other words, it's the ambiguity in natural language that's the stumbling block, along with having to keep track of a heck of a lot of context. Figuring out pronouns has got to be a pain, algorithmically, as even humans (and rabbits and ducks) get it wrong on a regular basis.
For example, in the previous sentence, to what does 'it' refer?
Check out Gödel, Escher, Bach: An Eternal Golden Braid for more on this subject than the average person can tolerate. And since that book has been around since 1979, the state-of-the-art in language processing has got to have improved. It was already quite impressive back then, but it's been 25 years.
A nifty exchange is found in the chapter titled "SHRDLU, Toy of Man's Designing." SHRDLU is a program which deals with a table covered in various blocks, and the user can interact with it. Here are a couple pieces of the dialog, without the running commentary:1. Eta Oin [the human]: pick up the big red block.
SHRDLU: OK.
2. Eta Oin: grasp the pyramid.
SHRDLU: I DON'T UNDERSTAND WHICH PYRAMID YOU MEAN.
3. Eta Oin: find a block which is taller than the one you are holding and put it in the box.
SHRDLU: BY "IT", I ASSUME YOU MEAN THE BLOCK WHICH IS TALLER THAN THE ONE I AM HOLDING.
SHRDLU: OK.
4. Eta Oin: what does the box contain?
SHRDLU: THE BLUE PYRAMID AND THE BLUE BLOCK.
5. Eta Oin: what is the pyramid supported by?
SHRDLU: THE BOX.
6. Eta Oin: how many blocks are not in the box?
SHRDLU: FOUR OF THEM.
7. Eta Oin: is at least one of them narrower than the one which i told you to pick up?
SHRDLU: YES, THE RED CUBE.
8. Eta Oin: is it supported?
SHRDLU: YES, BY THE TABLE.
9. Eta Oin: can the table pick up blocks?
SHRDLU: NO.
10. Eta Oin: can a pyramid be supported by a block?
SHRDLU: YES.
11. Eta Oin: can a pyramid support a pyramid?
SHRDLU: I DON'T KNOW.
12. Eta Oin: stack up two pyramids.
SHRDLU: I CAN'T.
...
17. Eta Oin: will you please stack up both of the red blocks and either a green cude or a pyramid?
SHRDLU: OK.
...
23. Eta Oin: had you touched any pyramid before you put the green one on the little cube?
SHRDLU: YES, THE GREEN ONE.
24. Eta Oin: when did you pick it up?
SHRDLU: WHILE I WAS STACKING UP THE RED CUBE, A LARGE RED BLOCK AND A LARGE GREEN ONE.
25. Eta Oin: why?
SHRDLU: TO GET RID OF IT.
26. Eta Oin: why did you do that?
SHRDLU: TO CLEAN OFF THE RED CUBE.
27. Eta Oin: why did you clear off that cube?
SHRDLU: TO PUT IT ON A LARGE GREEN CUBE.
28. Eta Oin: why did you do that?
SHRDLU: TO STACK UP THE RED CUBE, A LARGE RED BLOCK, AND A LARGE GREEN CUBE.
29. Eta Oin: why did you do that?
SHRDLU: BECAUSE YOU ASKED ME TO. Shows they were getting the hang of pronouns ('it', 'one', 'them', etc.), at least in a strictly-defined setting, quite a while back.
And as far as I can tell, creating a computer capable of AI will be largely independent of the type of hardware it's using. Given a large enough pile of today's machines, and a boatload of network hardware, I'm betting it could be done. People are looking to various forms of quantum processing, optical gates and fuzzy logic largely for the tremendous leaps in processing power available. These things can all be mimicked today with enough boring old binary logic gates. In fact, they are being mimicked today in order to design the "fancy" stuff. |
- Dave W. (Private Msg, EMail)
Evidently, I rock!
Why not question something for a change?
Visit Dave's Psoriasis Info, too. |
|
|
|
astropin
SFN Regular
USA
970 Posts |
Posted - 09/02/2004 : 22:01:26 [Permalink]
|
quote:
Originally posted by Valiant Dancer
I believe that machines can mimic sentience as far as they can be told about it, however, human behavior is made up of some unconscious thought that it is not logical to believe that humans will ever be able to produce a machine capable of it. You cannot program for things within the human subconscious which is not known about.
Well in my opinion if some future computer could "mimic" sentience perfectly than you can simply drop the word mimic. As far as this argument goes who's to say that humans don't mimic sentience?
Again...who said we need to mimic human (biologic) intelligence. There could be many forms of intelligence and I think electronic (not positronic) "brains" will someday far surpass human intelligence. (May they look fondly upon their creators)
quote:
Computers wade through a long stream of binary logic gates. (Thanks, Alan Touring) Humans do not. The whole concepts of emotion, intuition, and instinctual reactions are so poorly understood that it cannot be coded for. AI in the measure of a completely artificial sentient being is such a complex undertaking that I do not think that humans will ever know everything necessary to make one. Science is an ever expanding knowledgebase. It is literally impossible for us to "learn it all" as a species. It is this nature of science which I have to invariably point to and say the void of the things we do not fully understand will never be filled.
I see no need to fully understand how our brains function (although I still think that someday we will) in order to surpass it. Emotions, intuition and instincts in particular may not be necessary at all. Although a little empathy might be good for us.
|
I would rather face a cold reality than delude myself with comforting fantasies.
You are free to believe what you want to believe and I am free to ridicule you for it.
Atheism:
The result of an unbiased and rational search for the truth.
Infinitus est numerus stultorum |
|
|
|
astropin
SFN Regular
USA
970 Posts |
Posted - 09/02/2004 : 22:21:34 [Permalink]
|
quote:
Originally posted by Ricky
Also, don't try to say something is impossible.
quote:
At the rate we are going, we will never get to AI.
Come on Rick. You say don't say something is impossible and then in the very next paragraph state we will never get to AI at the rate we are going? The rate we have been going has been exponential!
quote:
To reach true AI, one would have to have a computer that can learn. We have the ability to mimic that, for example a computer learns your name asking you in a dialog box and you giving it.
I think getting a computer to learn and recall is actually the least of our problems. More difficult is giving a computer the five senses that meet or exceed our own. I think mobility may play an important role as well as seeing, hearing, smelling and tasting. The senses may not be required for AI (Like the example of HAL from 2001) but I think it would help as far as relating to biologic animals.
|
I would rather face a cold reality than delude myself with comforting fantasies.
You are free to believe what you want to believe and I am free to ridicule you for it.
Atheism:
The result of an unbiased and rational search for the truth.
Infinitus est numerus stultorum |
|
|
|
Maverick
Skeptic Friend
Sweden
385 Posts |
Posted - 09/03/2004 : 11:25:56 [Permalink]
|
Yes in my opinion AI is possible. When I say AI it includes many things not just humanoid robots that act in an intelligent way or HAL-type computers, but also systems that are selflearning to a certain degree, or that can make decisions based on knowledge and relations between facts etc.
Where to draw the line? When is a system sentient or aware? Is the plant aware of the surrounding conditions such as water and sunlight, since it responds to it? Is a computer aware of anything, since it does take input through senses such as the keyboard and microphone? In the case with the computer, in normal cases anyway, it isn't "aware" of anything else than the electric impulses it gets from the input devices. In a stretched way it's aware, but not more than an ant. This computer is also not selfconscious. But is it possible to build a machine that has an intelligence and selfawareness that is similar to ours? Yes, I'm sure that it is. I'm not at all sure that such a system can be done in only the way it's been done, but that it is possible in other ways as well. It should be possible to emulate one general pattern (the human brain and the way it behaves) with another pattern (the computer). Also, then it would be entirely possible to create other, different intelligences, perhaps specialized for certain tasks. I also see that it should be possible to build selfreplicating robots, so that intelligence could evolve within a population of robots.
It seems to me that many see human intelligence as the only intelligence on this planet, and that a true AI has to be exactly like the human intelligence, and I disagree with both. |
"Life is but a momentary glimpse of the wonder of this astonishing universe, and it is sad to see so many dreaming it away on spiritual fantasy." -- Carl Sagan |
|
|
|
Dude
SFN Die Hard
USA
6891 Posts |
Posted - 09/03/2004 : 21:19:05 [Permalink]
|
quote:
who said we need to mimic human (biologic) intelligence. There could be many forms of intelligence and I think electronic (not positronic) "brains" will someday far surpass human intelligence. (May they look fondly upon their creators)
There is alot of work being done investigating the possible merger (hopefully not Borg style) of mechanical and biological systems.
Very interesting stuff. |
Ignorance is preferable to error; and he is less remote from the truth who believes nothing, than he who believes what is wrong.
-- Thomas Jefferson
"god :: the last refuge of a man with no answers and no argument." - G. Carlin
Hope, n.
The handmaiden of desperation; the opiate of despair; the illegible signpost on the road to perdition. ~~ da filth |
|
|
|
|
Ricky
SFN Die Hard
USA
4907 Posts |
Posted - 09/03/2004 : 21:39:20 [Permalink]
|
quote:
Originally posted by astropin
quote:
Originally posted by Ricky
Also, don't try to say something is impossible.
quote:
At the rate we are going, we will never get to AI.
Come on Rick. You say don't say something is impossible and then in the very next paragraph state we will never get to AI at the rate we are going? The rate we have been going has been exponential!
Astropin, since when does "At the rate we are going, we will never get to AI." and "Also, don't try to say something is impossible." ever mean the same thing? If you look back to my reasoning of saying this, you will find that I said that it is because they method we are using. Right now, we are telling the computers how to respond to certain situations. Because there are really an infinite amount of situations, we would never reach true AI. We need to change our method. Then in the next paragraph, I suggested what that method needs to be changed to. We need to tell computers how to learn things on their own. Thats the only way I can see us ever reaching true AI.
I never said it was impossibe. If you look at some other posts, I make the arguement that nothing is really impossible, if your willing to accept absurdities. |
Why continue? Because we must. Because we have the call. Because it is nobler to fight for rationality without winning than to give up in the face of continued defeats. Because whatever true progress humanity makes is through the rationality of the occasional individual and because any one individual we may win for the cause may do more for humanity than a hundred thousand who hug their superstitions to their breast.
- Isaac Asimov |
|
|
|
Valiant Dancer
Forum Goalie
USA
4826 Posts |
Posted - 09/07/2004 : 07:00:06 [Permalink]
|
quote:
Originally posted by astropin
quote:
Originally posted by Valiant Dancer
I believe that machines can mimic sentience as far as they can be told about it, however, human behavior is made up of some unconscious thought that it is not logical to believe that humans will ever be able to produce a machine capable of it. You cannot program for things within the human subconscious which is not known about.
Well in my opinion if some future computer could "mimic" sentience perfectly than you can simply drop the word mimic. As far as this argument goes who's to say that humans don't mimic sentience?
Given the limited brain power of humans and intrinsic problems with human interaction, I do not believe that this is possible. Mimicing something as complex as human abstract thought and pushing past programming to answers that the computer has not been told about goes beyond the basic simple function of computers.
quote:
Again...who said we need to mimic human (biologic) intelligence. There could be many forms of intelligence and I think electronic (not positronic) "brains" will someday far surpass human intelligence. (May they look fondly upon their creators)
While computers can process decision trees extremely quickly already, they still lack the ability to go past their programming. (Although the same could also be said of most humans) New superclusters now talk about processing in the BIPS (Billion Instructions per Second or one thousand million to our European posters) range and some new technology is being researched to increase it into the TIPS (Trillion Instructions Per Second, or one million million or billion to European posters) range.
quote:
quote:
Computers wade through a long stream of binary logic gates. (Thanks, Alan Touring) Humans do not. The whole concepts of emotion, intuition, and instinctual reactions are so poorly understood that it cannot be coded for. AI in the measure of a completely artificial sentient being is such a complex undertaking that I do not think that humans will ever know everything necessary to make one. Science is an ever expanding knowledgebase. It is literally impossible for us to "learn it all" as a species. It is this nature of science which I have to invariably point to and say the void of the things we do not fully understand will never be filled.
I see no need to fully understand how our brains function (although I still think that someday we will) in order to surpass it. Emotions, intuition and instincts in particular may not be necessary at all. Although a little empathy might be good for us.
We already surpass it in some ways. Primarily in the mathematic computational areas. However, that which makes us human which is the ability to engage in abstract thought, does not exist in computers and represents such a non-concrete specification which is impossible to program for. |
Cthulhu/Asmodeus when you're tired of voting for the lesser of two evils
Brother Cutlass of Reasoned Discussion |
|
|
|
chaloobi
SFN Regular
1620 Posts |
Posted - 09/07/2004 : 07:04:54 [Permalink]
|
| The words "inevitable" and "impossible" are two that one would be wise to stay away from when discussing the future. I think the latter is the one more dangerous. I feel very confident in predicting neither is an accurate descriptor of the potential for AI. |
-Chaloobi
|
|
|
|
Ricky
SFN Die Hard
USA
4907 Posts |
|
|
Maverick
Skeptic Friend
Sweden
385 Posts |
Posted - 09/07/2004 : 09:29:00 [Permalink]
|
quote:
Originally posted by Ricky
I 2nd your post chaloobi.
Looking at the computers we have now, I could reasonably say that AI with this technology is almost impossible. However, what I ask is, why not change the way computers work? There is always a possibility, probably even a probability, that we will completely redesign the computer so that it doesn't just store data, but it learns it.
Until this happens, I feel as if we will never reach AI. We may be approaching it, but we will never be able to reach it.
Well we have AI applications already, so in a sense we do have it. The question is what do we actually mean when we say AI? |
"Life is but a momentary glimpse of the wonder of this astonishing universe, and it is sad to see so many dreaming it away on spiritual fantasy." -- Carl Sagan |
|
|
 |
|
|
|
Topic Locked
