The more I study on the subject the more I get the feeling that this is just an air buble to some extend.
The scientific discipline and engineering enterprise of AI has been characterized as “the attempt to discover and implement the computational means” to make machines “behave in ways that would be called intelligent if a human were so behaving” (John McCarthy), or to make them do things that “would require intelligence if done by men" (Marvin Minsky).
So in 1956 McCarthy invented the name: Artificial Intelligence. If he would have coined the name "Human Behavior Resembling Performance"........ But just think of it. We then would discuss HBRP....what abbreviation!
What I mean is that from that day on people used metaphors in describing the operating of computers. And we do it all day long nowadays. My computer searches the internet, watches for a key pressed, is on guard against virus attacks, and so on.
All our language about computers is in many respects a form of metaphorical usage of language. So, philosophically AI is a metaphor for HBRP for me from now on.
Due to the use of a certain language there is so much more suggested than there really is and scientifically justified. Let me give you an example.
Winograd (1972) developed a natural-language understanding program, which simulates a robot arm which can move a set of variously shaped blocks,
and allows a person to engage in a dialogue with the computer—asking questions, making statements, and issuing commands about this simple world of movable blocks.
This simple world of blocks then is named "microworld" and read the article in the Internet Encyclopedia of Philosophy (a good one!) and there anno 2009 you find the term 'microworld' as a standard concept.
But what is the suggestion in this term? To put is simple.....the more of these 'microworlds' we create and interlink with eachother the closer we'll get to the whole world of the Artificial Intelligence: not specific AI, but a general artificial intelligence.
To give this idea of microworlds an extra connotation, Winograd writes: 'We are concerned with developing a formalism, or "representation", with which to describe ... knowledge. We seek the "atoms" and "particles" of which it is built, and the "forces'' that act on it. '
It is true that in physics we can study relatively simple and isolated systems and then make the next step, based on the assumption that the whole universe is a coherent, law governed system.
My point here is that the language used in the AI community is to some extend rather suggestive, like these microworlds suggest that they eventually are a part of the big world.
When I come home from work I can say, "Well, my job is really a totally different world from my family". Microworlds?! Yes absolutely, but these subsystems we distinguish are part of the total system, which I call my life. And all relations between these subsystems are clear and distinct.
That is what the AI miniworlds suggest too, to be atoms and particles with forces acting on them as elements of one big coherent world. A metaphor.
Don't get me wrong. A lot is achieved in HBRP. In 1965, Daniel Bobrow wrote one of the first Rule-Based Expert Systems. It was called "STUDENT" and it was able to solve a variety of high-school algebra "word problems".
I love expert systems. I love PROLOG, the programming language which means Programming Logic. I love programming (LSL scripting if you like :-) In this financial crisis banks rely more on their expert systems than on human judgement.
At least that was in our news....you input all data and press the button. The answer is RED, ORANGE or GREEN. Formerly ORANGE caused the employee to take a closer look at the correctness of the data to certify that the ORANGE was justified. Now only GREEN entitles you to a bankloan....the expertsystem said so. Hurray...safe banking ^_^
Our world is filled with computer systems which do and even do better what humans did. The moment I am typing this, the wordprocessor reads every word and warns me for typos. Terrific but it is not AI but HBRP....(^_^)
Within limited contexts there is a lot achieved in making the computer do, what humans do too. Even in such a way that computers even do it better now and then (when they dont crash), but the question was: Can my computer think?
We can look into three ways of reasoning:
A.
1. Thought is some kind of computation.
2. Digital computers can perform all possible computations.
therefore,
3. Digital computers can think.
B.
1. Thought is some kind of conscious experience.
2. Machines can't have conscious experiences.
therefore,
3. Machines can't think.
C.
1. Thoughts are specific biological brain processes.
2. Artificial computers can't have biological brain processes.
therefore,
3. Artificial computers can't think.
My conclusion of this moment is that I am not finished with this subject and you may have to endure a third lecture on HBRP :-)
The Discussion
[13:25] Daruma Boa: yes please
[13:25] Paula Dix: lol good
[13:25] Alarice Beaumont: sounds good to me :-)
[13:25] Daruma Boa: perhaps we should learn more about computers first^
[13:25] Gemma Cleanslate: oh goodness
[13:25] Daruma Boa: lsl course or html at the beginning
[13:26] Daruma Boa: ^^
[13:26] Paula Dix: and about thinking
[13:26] Daruma Boa: *GIGGLES* :)~~~~
[13:26] Laila Schuman: have we defined the word "think".....
[13:26] herman Bergson: i could give an LSL course indeed
[13:27] herman Bergson: Have we defined the word "think"
[13:27] herman Bergson: That, Laila is the whole point...
[13:27] Laila Schuman: so many philosophers have asked.... how do we know things
[13:27] Laila Schuman: sensory experience is usually sited
[13:27] herman Bergson: All writers on HBRP begin with saying...welll...let's not start debating about the definition of thinking
[13:28] Gemma Cleanslate: lol
[13:28] Laila Schuman: that is a cop out
[13:28] Daruma Boa: so easy way for them
[13:28] Paula Dix: i see, their idea is to make a human simulation, not real inteligence
[13:29] herman Bergson: as I pointed out in my former lecture..Turing started with the question "Can machines think?" and immediately changed it in another one...dropping the concept of thinking
[13:29] Laila Schuman: if you ask... do they think... then you should not take out the word think as something you cannot discuss
[13:29] Samuel Okelly: i think there is an essential question we need to attempt to answer, and that is based around the clarifying "free will" and what it means to have a freedom of action
[13:29] Daruma Boa: true laila
[13:30] herman Bergson: Not so much Samulel...
[13:30] Laila Schuman: i agree wtih Samuel... free will is very important
[13:30] herman Bergson: Read the article in the IEP....
[13:30] Daruma Boa: for thinking?
[13:30] herman Bergson: it states that a deterministic machine and free will can go together..
[13:30] herman Bergson: lol...I forgot the arguments
[13:31] Laila Schuman: free will? or just a mathematical probability
[13:31] itsme Frederix: Herman in all cases you state: "thought=" followed by "computers can or cannot" but there is also the possibility computers can generate an environment capable off ...
[13:31] Samuel Okelly: surely an understanding of “free will” (as opposed to a deterministic view) is essential in being able to determine the limits or possibility of AI?
[13:32] Laila Schuman: i think..therefore i am
[13:32] herman Bergson: There is much more to state the limits
[13:32] itsme Frederix: Do we asume free will is a property of Intelligence?
[13:32] Laila Schuman: AM
[13:32] Laila Schuman: concious of self
[13:32] Paula Dix: it would be like they can have free will between certain limits??
[13:33] herman Bergson: I think we have to make a difference between sof and hard AI, as is proposed in the IEP article
[13:33] Tess Aristocrat: limits is the operative word.
[13:33] Tess Aristocrat: we place the limits
[13:33] itsme Frederix: most computers are very conscious about their behaviour p.e. ... sorry an error occurred
[13:33] herman Bergson: What philosophically is interesting is the contentions of the hard AI..
[13:33] Peaps Homewood: Independant thinking is the intelligence
[13:33] Paula Dix: yes, in some sense there is no free will, there are always limits
[13:34] Samuel Okelly: if our actions are “determined” then that will clearly lead to different conclusion regarding AI than if our actions are governed by independent free thinking
[13:34] herman Bergson: the idea that computers eventually can think.....
[13:34] herman Bergson: The free will issue isnt such a big thing
[13:34] herman Bergson: if computers are physical objects following physical laws...so are human bodies
[13:35] Laila Schuman: it is a manipulation of binary... that has tons of possible outcomes/responses... but it is still electricity and binary
[13:35] Laila Schuman: i disagree Herman
[13:35] itsme Frederix: What are the result of thinking (rather then what is the proces) and then judge the computer results
[13:35] herman Bergson: That is the behaviorist approach Itsme
[13:36] itsme Frederix: forget binary thats just a implementation'
[13:36] Laila Schuman: i don't think so itsme
[13:36] Laila Schuman: it is programing
[13:36] Tess Aristocrat: the computer's answers will always reflect, in some way, the person that programmed it.
[13:36] herman Bergson: But that leads to the idea that everything that behaves like humans is human
[13:36] Laila Schuman: pure and simple
[13:36] Laila Schuman: programing
[13:36] itsme Frederix: Herman call the approach as you want - it gives room for an answer
[13:36] Samuel Okelly: I disagree herman because if we are reduced to preprogrammed beings then a comparison with computers is obvious
[13:36] Samuel Okelly: however if we are autonomous beings capable of free will then it makes comparisons with AI far more difficult
[13:36] Paula Dix: we praise people with math habilities... in this sense computers are smarter than us
[13:37] Laila Schuman: computer is a tool...like a calculator
[13:37] Laila Schuman: not smarter
[13:37] Laila Schuman: the programer is the smart one
[13:37] Jangle McElroy: The first things called computers were human, it was a job to compute.
[13:37] itsme Frederix: in war humans are tools also (sorry)
[13:37] Peaps Homewood: computers follow rules , defined by their creator , the creators thoughts , ideas
[13:37] itsme Frederix: and what about production / fabric
[13:38] Laila Schuman: itsme.... please stay on the subject
[13:38] herman Bergson: The fundamental question still is..can my comutor THINK
[13:38] Tess Aristocrat: yes, even if they program the computer to do 'what ifs' etc..it is limited to its input
[13:38] itsme Frederix: I just argued your analogy Laila
[13:38] Laila Schuman: can your record player think? it sings
[13:38] Peaps Homewood: who programmes the computer to think ?
[13:38] Peaps Homewood: Humans
[13:38] herman Bergson: Is THINKING a purely human capability or can you implement it in a machine
[13:39] itsme Frederix: please define thinking
[13:39] Laila Schuman: it is programing... not thinking
[13:39] Samuel Okelly: can we agree an understanding of THINKING without refernce to free-will?
[13:39] Paula Dix: i guess computers are still too primitive, stupid, but that will hopefully change with time
[13:39] herman Bergson: Well..thinking is a mental state generated by a brain
[13:39] Tess Aristocrat: to a degree you can implement it- its answer will always be in black and white- no grey areas
[13:40] Peaps Homewood: true Tess
[13:40] Laila Schuman: a map has tons of possible routes to get to the same place... but it does not think..it too is a tool
[13:40] Peaps Homewood: It doesn't think outside the box
[13:40] itsme Frederix: thats very basic, and defines the material you need - reducing it to a narrow definition
[13:40] Tess Aristocrat: I suppose it could give an I don't know answer too
[13:40] Laila Schuman: and it has the right answers too
[13:41] herman Bergson: If it is programmed to, Tess
[13:41] Paula Dix: Yes Tess, but there are researches about computers with more complex structure
[13:41] Jangle McElroy: Is we assume other animals can think, how low do you have to go down the hierarchy of the animal world before you find a creature less able to think than a computer. And yet we'd still say insects can think ,because they are able to overcome changing situations in a complex world and learn from each other.
[13:41] Paula Dix: Jangle ive seen some expert said computers are level with lizzards
[13:41] itsme Frederix: Herman your definition is about a "how" but what is the result .. not just a mental state Iguess
[13:42] Laila Schuman: hitchhiker's guide to the galaxy... i believe the computer came up with an anwer to the question... what is the answer to life and the universe and everything...
[13:42] Jangle McElroy: So if we agree a lizard can think, so can computers I guess
[13:42] Peaps Homewood: animals are conditioned , by lifes experience , computers are programmed
[13:42] Laila Schuman: answer was... 37
[13:42] Paula Dix: 42!
[13:42] Laila Schuman: or some number i can't remember
[13:42] Paula Dix: lol
[13:42] Jangle McElroy: computers are conditioned, it;s the same thing though a different process
[13:43] Laila Schuman: 42 thanks Paula
[13:43] Gemma Cleanslate: this is so interesting
[13:43] itsme Frederix: (just a remark, in every computer there is that little tiny men - we cant see - )
[13:43] Paula Dix: lol
[13:43] Tess Aristocrat: haha
[13:43] herman Bergson: To get back to Itsme..thinking is a mental state
[13:43] itsme Frederix: like the first chessmachine (with a dwarf in it)
[13:43] Tess Aristocrat: yes, we call them gremlins..that's who botched the job, not us!!!
[13:44] herman Bergson: ]Exactly :-)
[13:44] Paula Dix: itsme, thats for real? someone did that??
[13:44] itsme Frederix: Herman - a mental state - what should I see then ? what is a mental state - how can I measure it
[13:44] herman Bergson: that chessmachine..yes it happened
[13:44] Paula Dix: lol how good the dwarf was on chess??
[13:44] Tess Aristocrat: hehehe..
[13:45] herman Bergson: it is a state of self awareness and the content of what you are awareof
[13:45] herman Bergson: a state of consciousness
[13:46] herman Bergson: and that is a different state than a computer is in when I run an expert system in it
[13:46] itsme Frederix: well a computer (at least a mainframe) is programmed to be aware of it state & content
[13:46] Paula Dix: My friend has a book here called ME (from Mechanical Entity) about a computer software that becomes alive, very interesting
[13:46] herman Bergson: it juggles with values...no idea what they mean..
[13:46] Paula Dix: the key point was this one, it suddenly was aware of itself
[13:46] Peaps Homewood: still a programme
[13:47] Tess Aristocrat: free of emotions; fear, anger, sadness
[13:47] Samuel Okelly: how do we measure the level or degree of self-awareness in others?
[13:47] herman Bergson: Yes..computers that become selfaware is a great issue in SF and movies
[13:47] Paula Dix: in the end of the book the programmer release it on the net and said to it "be good" and the program goes out thinking "what he means with that?"
[13:47] Tess Aristocrat: emotions that might have a say one way or another depending on the mood
[13:47] herman Bergson: THat is a special chapter Samuel....Other Minds...
[13:48] itsme Frederix: like HAL in 2001 space (hall is 1 char shifted IBM)
[13:48] herman Bergson: How do we know the other (human or machine ) hasa mind...
[13:48] Peaps Homewood: limited by its state and content
[13:49] itsme Frederix: Herman that might be the key question
[13:49] herman Bergson: But we only have to deal withthat when my computer says..Hmm I am switched on..I exist :-)
[13:49] Paula Dix: first some capacity of communication is needed
[13:49] Samuel Okelly: i agree and i would suggest that we need to resolve or at least arrive at a position on this before we can progress further
[13:49] herman Bergson: Ok Samuel....you got a deal
[13:49] itsme Frederix: Sam goodpoint
[13:50] herman Bergson: Next lecture will be on the issue of Other Minds. or "How do I discover the mind in my computer?"
[13:50] Paula Dix: nice
[13:50] Paula Dix: e
[13:50] Laila Schuman: that assumes there is one
[13:50] herman Bergson: It is a good point...for if my computer thinks..it has a mind..or not?
[13:51] herman Bergson: or does it think without a mind?
[13:51] Laila Schuman: you still have not defined think
[13:51] Jangle McElroy: Is there a section on discovering the mind in SL that decides it won't work correctly so often
[13:51] itsme Frederix: answering a question with a proclamation
[13:51] herman Bergson: The AI people dont talk about that
[13:51] Paula Dix: lol good point, to think requires a mind? bacteria can think?
[13:51] Tess Aristocrat: a computer is a reflection..like a mirror
[13:51] Laila Schuman: yes... Tess
[13:52] Laila Schuman: i agree
[13:52] Tess Aristocrat: however, it's a great database and notebook for answers
[13:52] itsme Frederix: Laila Herman did
[13:52] herman Bergson: Did what Itsme?
[13:53] Paula Dix: defned thinking
[13:53] Laila Schuman: i did not see it defined
[13:53] Paula Dix: process of a mind
[13:53] itsme Frederix: did define thought Herman
[13:53] Tess Aristocrat: nods..yes ..haha
[13:53] Laila Schuman: oops... what is mind then
[13:53] Paula Dix: process of a brain?? :)
[13:54] Peaps Homewood: independant thought
[13:54] Paula Dix: lol
[13:54] Laila Schuman: philosophers love semantics don't they?
[13:54] herman Bergson: Well..I guess ....since we will have a lecture on other minds, we might take into account our own too :-)
[13:54] Jangle McElroy: f we can be in 2 minds, computer's surely can be in at least 1
[13:55] Laila Schuman: eeewwwww
[13:55] Qwark Allen: what you mean jangle?
[13:55] herman Bergson: what about parrallel computing..is that digital schizophrenia?
[13:55] Gemma Cleanslate: see you nextt week :-)
[13:55] Gemma Cleanslate: have to go
[13:55] herman Bergson: Bye Gemma..:-)
[13:55] Qwark Allen: darn ! me to
[13:55] Paula Dix: brains do parallel processing, dont they?
[13:56] Qwark Allen: cya tuesday
[13:56] Laila Schuman: baiee...sorry
[13:56] Laila Schuman: gotta go
[13:56] itsme Frederix: and discrete?
[13:56] Daruma Boa: ciao
[13:56] Paula Dix: byE!
[13:56] Ze Novikov: off to RL ty herman bb everyone se you all soon
[13:56] Jangle McElroy: Could sentinece bring the ability to err (think twice), rather than just compute and action (like a computer wieghing up the equasion and acting as its programme dictates)
[13:56] herman Bergson: Our time is up..
[13:56] herman Bergson: lots of computers are going down :-)
[13:56] itsme Frederix: think I'm going to crash in this discussion - brain damage - overflow
[13:56] Paula Dix: lol
[13:56] herman Bergson: Thank you all for your participation
[13:56] Jangle McElroy: Bye everyone
[13:57] Jangle McElroy: Thanks HErman
[13:57] Samuel Okelly: thank you herman :) cheerio for now every1 :)
[13:57] Tess Aristocrat: thank you Herman :)
[13:57] Paula Dix: wonderful... cant wait for next episode!!
[13:57] herman Bergson: To be continued next week Paula
[13:57] Paula Dix: lol yes
[13:57] Paula Dix: wich is perfect since i have to run and would hate to lose it :)))
[13:58] Paula Dix: bye!
[13:58] herman Bergson: bye paula
Showing posts with label Artificial Intelligence. Show all posts
Showing posts with label Artificial Intelligence. Show all posts
Sunday, June 7, 2009
Wednesday, June 3, 2009
6a Can my computer think???
The preparations of this lecture caused in me so much stress that I decided to see my therapist, a charming lady with beautiful eyes. Maybe you know her too. Her name is ELIZA ^_^
ELIZA: Hello.
ELIZA: How do you do. Please state your problem.
Me : I must find an answer to an important question
ELIZA: You say you must find an answer to an important question ?
Me : yes
ELIZA: You seem to be quite positive.
Me : Can computers think?
ELIZA: Do computers worry you ?
Me : Absolutely
ELIZA: I'm not sure I understand you fully.
Me : The idea that computers can think scares me
ELIZA: Why do you mention computers ?
Me : Because a thinking computer scares me
ELIZA: What do you think machines have to do with your problem ?
Me : Don't you listen?
ELIZA: We were discussing you -- not me.
Me : I am afraid of thinking computers
ELIZA: Don't you think computers can help people ?
Me : Maybe
ELIZA: You don't seem quite certain.
Me : no
ELIZA: Are you saying no just to be negative?
Me : I say no because I am not certain
ELIZA: Is it because you are not certain that you came to me ?
Me : yes
ELIZA: You are sure.
Me : I am
ELIZA: How long have you been ?
Me : all the time
ELIZA: Please go on.
Me : Are you a computer?
ELIZA: What about machines worrys you ?
Me : machines have no soul
ELIZA: You are being a bit negative.
Me : it is true
ELIZA: What does that suggest to you ?
Me : It suggests that machines are not human
ELIZA: Did you think they might not be not human ?
Me : They are not human
ELIZA: Would you like it if they were not not human ?
Me : Oh yes
ELIZA: I see.
See: http://www.chayden.net/eliza/Eliza.html
She understood and took the burden from my shoulders. I could go on with my preparations and reading.
I am a great fan of Weizenbaum. He died in 2008 in Berlin. He was the creator of the computerprogram ELIZA, which acts as an empathic psychologist. Conversations with her can be really hilarious.
Some people really took ELIZA serious and openen their heart for her, discussing intimate personal problems. But it all began with Alan Turing, who wrote in 1950 a paper named "Computing Machinery and Intelligence".
He begins his article thus: "I propose to consider the question "Can machines think?" This should begin with definitions of the meaning of the terms 'machine' and 'think'. The definitions might be framed so as to reflect so far as possible the normal use of the words, but this attitude is dangerous."
Why is this attitude dangerous? Simple, there is no proper definition for "thinking" availabe. But then it goes wrong, when he says "Instead of attempting such a definition I shall replace the question by another, which is closely related to it and is expressed in relatively unambiguous words."
Turing comes with the appealing question "Can machines think?" A question that drives me into the arms of ELIZA, but there could a hidden agenda too ;-) and then suggest to replace it with another question. What in the world may mean CLOSELY related?
Anyway, his new question is something like "Are there imaginable digital computers which would do well in the [Turing test]"? The picture here (also found at http://en.wikipedia.org/wiki/Turing_test#cite_note-2) shows you the set up of the test.
So this is closely related "thinking"? The fact that a computer can show the same conversational behavior as a human being? The Turing test is based on the assumption that human beings can judge a machine's intelligence by comparing its behaviour with human behaviour.
There it is....intelligence. There neither is a proper definition for intelligence. What kind of intelligence? Didn't we learn about emotional intelligence by Goodman's book, or social intelligence?
I think, that Turing was trapped in his historical context. Just imagine, though gigantic machines, they had the first electric machines which were able to perform complex calculations, just like a human can.
And then we have our imagination.......if this is possible, there definitely is more to come. When I am not mistaken, the computer Turing describes in his test, wasn't even technically possible in 1950, only theoretically.
The Turing Test has become a classic in the history of artificial intelligence and Weizenbaum's ELIZA would be a candidate to pass the test. But philosophically it doesn't bring much if it is regarding the question "Can my computer think?"
But we live in a time period in which people more and more tend to believe that computers can think. Did it never occur to you? You call the office and say "You sent me a letter telling me that I haven't paid my bill."
"Yes we did, sir" " But I have paid my bill, weeks ago." "I am sorry Sir, but the computer says that you have not paid your bill, so we sent you that reminder." "But your computer is wrong!" "I am sorry Sir, it is says it here on my screen: you have not yet paid."
It can be even more bizarre. Here in the Netherlands we have a TV commercial like this; Young couple in bar. She just met him. Is he the perfect partner? Send an SMS message with your name and his name to number XXXX and the computer gives you the answer in no time.
Girl does so. Score 3% !!! As a conclusion she throws the content of her glass in the face of the young man.
Or this one: Send your name and your wifes name to number XXXX and the computer gives you in no time the perfect name for your baby.
Something is wrong here with man and computer. So it wont be overkill to spend another lecture on the philosophy of Artificial Intelligence: AI
John McCarthy, who coined the term in 1956, defines it as "the science and engineering of making intelligent machines." But is intelligence the product of the interaction of all kinds of materials, put together in a factory, or is it product of the fact that we are dealing LIVING matter?
The Discussion
[13:23] herman Bergson: If you have any questions or remarks.....
[13:23] Scope Cleaver: Floor's open?
[13:23] Gemma Cleanslate: already?
[13:23] herman Bergson: go ahead :-)
[13:24] Laila Schuman: i find it necessary to get that definition of intelligence... just as we need a definition of education... depends upon whom is speaking
[13:24] Scope Cleaver: Well I think the Turing test is still very relevant today and better than anything we have so far
[13:24] Scope Cleaver: That definition of intelligence I think is a bit tricky
[13:24] Ludwig John: sometimes I suppose that my writing programm is intelligent: I write some letters and it completes the whole word
[13:24] Laila Schuman: yes
[13:24] herman Bergson: Well in the world of AI it has not so much meaning
[13:24] Laila Schuman: agreed
[13:24] Scope Cleaver: If I introduce you to someone and ask you, do you find that person intelligent, you don't go into definitions of intelligence
[13:24] Laila Schuman: no but i have opinions
[13:25] Scope Cleaver: intelligence is about human performance
[13:25] Paula Dix: ive read a part of a Dennet book where he talk about turing test, very interesting
[13:25] Scope Cleaver: Machine intelligence is about how well it emulates that
[13:25] Laila Schuman: and a lot depends upon whether they or i are significantly more or less intelligent than the other
[13:25] herman Bergson: yes if you define intelligence in a behavioristic or functional way
[13:25] Scope Cleaver: Thats the generally received view yes
[13:26] herman Bergson: Yes...so we call emulation intelligent behavior
[13:26] Scope Cleaver: Note that it's one thing to have an intelligent thinking machine
[13:26] Scope Cleaver: that doesnt mean it's conscious or that even has to be
[13:26] herman Bergson: But the reat question was Can machines think
[13:26] Laila Schuman: finding the result of a math problem or giving a reply based on grammar...does not mean one doing it can apply it
[13:26] Scope Cleaver: Yes they can
[13:27] Gemma Cleanslate: and will they ever really be able to
[13:27] Scope Cleaver: Unless we're willing to debate whether anything *can* think, machines certainly can.
[13:27] Paula Dix: When ive seen this would be the theme for today, i put this question to the people here at home, and Fat said its unfair to judge computers in the state they are, since they are so limited
[13:27] herman Bergson: If so ...it is only a matter of definition
[13:27] Paula Dix: He say they are intelligent, but something in the level of a lizzard :)))
[13:27] Scope Cleaver: Yes you can debate semantics of course
[13:27] Scope Cleaver: But like I said it's interesting that people dont' do that when you ask them to judge if another person is intelilgent.
[13:28] Jangle McElroy: I guess it might be easier to suggest machines can calculate based on rules, where as man can create without rules.
[13:28] Scope Cleaver: they seem to have a sensible sense of what intelligent means when asked about someone else but then discuss and debate semantics when it's a machine.
[13:28] herman Bergson: Another person you say....not another computer,,,that is the difference
[13:28] Gemma Cleanslate: that makes sense to me jangle
[13:28] Paula Dix: Scope but we see all the time people passing judgments because someone was born here or there and so on, its the same thing
[13:28] Scope Cleaver: But thats the beauty of the Turing test
[13:28] Scope Cleaver: it doesn't tell you at the onset which is which
[13:29] Scope Cleaver: You only interact with the other entity via teletext
[13:29] herman Bergson: Yes indeed....
[13:29] Scope Cleaver: But whether you pass judgment or not is completely immaterial to whether an entity is or isn't intelligent.
[13:29] herman Bergson: But Turing had to change his question...and had to drop the fundamental concept of thinking
[13:30] Scope Cleaver: What matters is, is there a test we can devise that would realiably tell us if an entity is intelligent
[13:30] Paula Dix: oh, you dont know what we brazilians think about portugueses!!!! :))))
[13:30] herman Bergson: Yes Scope..from a pragmatic point of view..if it works it works...whether it is a computer or a man doing the job
[13:30] Ze Novikov: lol
[13:30] Scope Cleaver: Right
[13:31] herman Bergson: But the question Can computers think suggest much more...
[13:31] Scope Cleaver: It's a very complex question, but I think the interesting part reside in our formulating the question the right way, somehow the turing test helps us do that
[13:31] herman Bergson: it is close to the idea..Have computers a mind?
[13:31] Scope Cleaver: it's has flaws of course but it's overall a good test
[13:31] Scope Cleaver: hehe yes indeed but the word "mind" opens a big can of worms semantically speaking
[13:31] Scope Cleaver: For instance you'd get a different definition for everyone in this room
[13:31] herman Bergson: A test of what...? That human behavior can be emulated by machines?
[13:31] herman Bergson: We have factories ful of robots
[13:32] Scope Cleaver: A test for intelligence
[13:32] Scope Cleaver: Which is plain performance of cognitive functions
[13:32] herman Bergson: No....for it only is about emulating huma conversatin
[13:32] Scope Cleaver: memory, reasoning, inference, and so on
[13:32] Scope Cleaver: the human conversation part is just a tool
[13:32] herman Bergson: And I would define intelligence as the ability to have a converstaion
[13:32] Paula Dix: I would say the moment computers have capacity similar to a human brain the question will be solved
[13:33] Scope Cleaver: I belive it' s more than that but the point is, a conversation can help you determine if an entity is intelligent
[13:33] Scope Cleaver: we'd have to establish necessary and sufficient conditions
[13:33] herman Bergson: memory, reasoning, inference yes....but imagination, intentionality, emotions..no
[13:33] Scope Cleaver: Thats a scientic question herman
[13:33] herman Bergson: also part of us as intelligent beings
[13:34] herman Bergson: It is an observation
[13:34] Scope Cleaver: For instance it isn't clear how *we* aren't machines that in itself is very debetable
[13:34] Paula Dix: but we are machines
[13:34] Jangle McElroy: Intelligence tests often rely on observation and calculation. Making a calculator with a large memory capable of simulating intelligence.
[13:34] Scope Cleaver: Like I said it's debatable
[13:34] Scope Cleaver: The general received view is that we are.
[13:35] Paula Dix: the divisions, the labels, are all artificial
[13:35] Scope Cleaver: it depends what your materialistic position is
[13:35] herman Bergson: The Turing Test is just a very narrow view on what thinking entails,I would say
[13:35] Scope Cleaver: It's just a tool
[13:35] Paula Dix: i bet tons of humans wouldnt pass turing test
[13:35] herman Bergson: but completely outdated...
[13:35] Scope Cleaver: You see it has to have something ont he table that *both* humans and machine could do
[13:36] Scope Cleaver: communicating through text is one
[13:36] Laila Schuman: is an education the ability to pass tests... or is there something within the educated person that has much more going on... whether they pass tests or not
[13:36] Scope Cleaver: there is a clear reason why he choose to have text via teletext instead of voice for instance or acting
[13:36] Ze Novikov: in the end it is about sequencial series of data or non sequential...
[13:36] Paula Dix: dennett mentions a situation where people couldnt pick the machine on the test
[13:37] Scope Cleaver: Yes paula when that happens
[13:37] Scope Cleaver: when you results are 50/50 you can safely say you have true AI
[13:37] Laila Schuman: can a computer tell a joke that is not fed to it
[13:37] Scope Cleaver: It could why not?
[13:37] Gemma Cleanslate: to me it demonstrates that a machine can be intelligent but does it really have intelligence if it does not have feelings too
[13:37] Paula Dix: they made a program to simulate paranoia, and psychiatrists had to say who was the machine between real paranoics
[13:37] herman Bergson: It could?
[13:37] Scope Cleaver: Can a bunch of gray matter cells tell a joke thats not fed to it? haha
[13:37] Paula Dix: they couldnt, because they couldnt go really deep on the questions for fear of causing damage to the persons
[13:38] Paula Dix: so they all got it wrong
[13:38] herman Bergson: Yes it can Scope...you can think of a new joke..a play of words and so on
[13:38] Scope Cleaver: in that case yes, I think a machine could do that as well ㋡
[13:38] Paula Dix: lol somewhere the jokes where invented
[13:38] Scope Cleaver: Exactly my point
[13:39] Scope Cleaver: So there is cognitive work being done to come up with a joke
[13:39] herman Bergson: Yes..jokes are invented..but not rule based, but by deviating from rules
[13:39] Ludwig John: can a computer write a novel or a poem?
[13:39] Scope Cleaver: Something an AI could do
[13:39] Jangle McElroy: Tell a joke? Or create a joke? Creation is perhaps closer to free will and intelligence than calculation. It's a bit like the tests that show computers can write music - it often sounds fomulaic rather than freeform and flowing.
[13:39] Scope Cleaver: Ludwig yes
[13:39] Scope Cleaver: and in fact, theoretically it doesn't even need to be intelligent to do it
[13:39] Scope Cleaver: it could do it by fluke hehe ㋡
[13:39] herman Bergson: Yes Jangle....that is the case indeed
[13:40] Paula Dix: but its again the same point, computers "brains" are still too small
[13:40] Paula Dix: they do music or anything with this small capacity
[13:40] Scope Cleaver: Someone mentioned Dennett earlier I think he has that monkey typing a play by fluke as an example
[13:40] Paula Dix: surely not like us, but that will change with time
[13:40] Paula Dix: No that was Dawkins :))
[13:41] Scope Cleaver: We're a bunch of decades away but I don't think the question of when is that interesting
[13:41] herman Bergson: We'll definitely dig into those issues the next lecture..:-)
[13:41] herman Bergson: Well the bet is set to 2029 now
[13:41] Paula Dix: i wonder, someone here think computers cant thing?
[13:41] Qwark Allen: yes
[13:41] Scope Cleaver: i think thats way early hehe
[13:41] Qwark Allen: close to that hermman
[13:41] Gemma Cleanslate: i do not think they can tghink
[13:41] Daruma Boa: i think they think^
[13:42] Qwark Allen: they will
[13:42] Daruma Boa: my mac does^^
[13:42] Qwark Allen: in time
[13:42] Scope Cleaver: I wont' pass *my* turing test by 2029 not to be presumptuous lol
[13:42] herman Bergson: No...in my opinion computer cant think...
[13:42] Paula Dix: my vote is for yes too :))
[13:42] Gemma Cleanslate: i think they only think what we tell them to think
[13:42] Gemma Cleanslate: so far anyway
[13:42] Jangle McElroy: :)
[13:42] herman Bergson: the perfor rulebased actions
[13:42] Daruma Boa: no, these days there r programms, which learn
[13:42] Paula Dix: but herman our brains also dont work like that?
[13:42] Ze Novikov: every evening I fight with my computer .." Now time to turn you off and go to bed....."
[13:43] Scope Cleaver: What is of interest is that *once* they are as intelligent as we are... they won't stay that miserable for long
[13:43] herman Bergson: no....the idea that the brain is a computer is a nice metaphor but a flaw
[13:43] Laila Schuman: lol
[13:43] Paula Dix: lol imagine when computers are bigger and start to have personalities?? :)))
[13:43] Gemma Cleanslate: or much smaller and do that
[13:43] Daruma Boa: right scope^^
[13:43] Paula Dix: why herman??
[13:43] Scope Cleaver: I think the substrate question is tangential
[13:43] herman Bergson: No. Paula....the size of a machine doesnt count....
[13:43] Scope Cleaver: what intelligence is implement *in* doesn't really matter
[13:43] herman Bergson: To begin with it is Life that counts..
[13:44] herman Bergson: Our mental states are caused by the fact that we are alive
[13:44] Scope Cleaver: herman do you belive there is a quality to life such as an "elan vital"?
[13:44] Jangle McElroy: If computers are ever capable of sentient independent thought, I wonder if they will be like the Orangutan? The human tribes people of Borneo say Orangutans only don't speak because someone would give them a broom and order them to sweep up if they could communicate.
[13:44] Scope Cleaver: A something above the atoms that makes our being?
[13:44] herman Bergson: When you put all chemicals together that consitute life....it still hasnt happen to create living material
[13:45] herman Bergson: No Scope....
[13:45] herman Bergson: It must be those atoms, I think...
[13:45] Scope Cleaver: So what is the magic property?
[13:45] Scope Cleaver: and why machines can't have it?
[13:45] Paula Dix: we are going toward biological computers... will them be alive?
[13:45] herman Bergson: If you knew that we could create living matter in the laboratory
[13:46] Scope Cleaver: I belive thats flawed but thats my opinion.
[13:46] herman Bergson: We may be going to biochemical computers..that's the difference
[13:46] Paula Dix: i think like scope, no difference if its carbon or metal
[13:47] Paula Dix: or silica :))))
[13:47] herman Bergson: Ok.....good material for the next lecture...:-)
[13:47] Scope Cleaver: You'll often find difference in the metaphysics of being is rooted in deep beliefs...
[13:47] Paula Dix: lol yes, thats fun
[13:47] Scope Cleaver: About other stuff ㋡
[13:48] Scope Cleaver: We hope Eliza didn't cause you too many turmoils herman
[13:48] herman Bergson: Well she is a great help :-)
[13:48] Scope Cleaver: Could you believe how many people she helped.
[13:48] Scope Cleaver: Maybe more than any single *human* therapist hehe
[13:48] Gemma Cleanslate: she is a loop of nothing lol
[13:49] herman Bergson: Some conversations...especially with an improved version can have a positive behavioral effect, I think
[13:49] Gemma Cleanslate: yes lol
[13:49] Gemma Cleanslate: is fun
[13:49] Scope Cleaver: Thanks for the stimulating lecture
[13:49] herman Bergson: You should visit Elbot....
[13:49] Daruma Boa: here?
[13:49] Paula Dix: whats elbot?
[13:49] herman Bergson: He is fun to talk with
[13:49] Paula Dix: oh, another program?
[13:49] herman Bergson: Just google on ELBOT
[13:49] Paula Dix: ok
[13:50] herman Bergson: He is another chatbot
[13:50] Daruma Boa: ok will do
[13:50] herman Bergson: Won the Loebner prize
[13:50] Scope Cleaver: *nods*
[13:50] Paula Dix: dennet mention another funnny experiment, a program made to think it was some american politician
[13:50] Scope Cleaver: We'll start to worry when it's more fun to talk to a bot than our spouses
[13:50] Paula Dix: (I think it was american...)
[13:50] Samuel Okelly: i found elbot like having a conversation with a child who wont take its eyes of the TV to talk to you ;-)
[13:50] Scope Cleaver: lol Samuel
[13:51] herman Bergson: Yes Scope.....then we have s lost scope indeed
[13:51] Scope Cleaver: Ai is very childlike which is why it isn't fully Ai yet
[13:51] Paula Dix: exact
[13:51] Scope Cleaver: But hey it beat Kasparov at chess, first step to everything
[13:51] herman Bergson: I dont like that argument.....
[13:51] Paula Dix: talking TV, loved the peanuts cartoon!
[13:51] Daruma Boa: why herman?
[13:51] herman Bergson: the argument of poverty...
[13:51] Scope Cleaver: I meant to say it's in developmental stage
[13:52] Scope Cleaver: Learning is a very big component to AI
[13:52] herman Bergson: Well ..it is the same as the belief that eventually Science will solve all problems
[13:52] Daruma Boa: why is it pover?
[13:52] Scope Cleaver: the reasearchers are still baning their head in implementing better learning algorithms in AI
[13:52] Paula Dix: yes, its a belief
[13:52] Paula Dix: should i add "but a right one"? :)))))
[13:53] herman Bergson: It is poverty as it states..oh I am a poor guy now, but wait..I will get rich :-)
[13:53] herman Bergson: When we evaluate AI we have to evaluate it as it IS.
[13:53] herman Bergson: Not on what it promisses to become
[13:53] Scope Cleaver: But thats the beauty of science it can build upon it's own knowledge and discovery
[13:53] Scope Cleaver: yes I agree on that point
[13:54] Paula Dix: yes, its true, but we think it can do this or that and this is a belief
[13:54] Scope Cleaver: But whats important about the "what ir promises to become" part is that we can use that to start now, to think about the ethical problems we will face.
[13:54] Samuel Okelly: science often argues from what may be
[13:54] herman Bergson: Yes...that is what we humans can....
[13:54] herman Bergson: Have hope and expectation...how to program that in a computer?
[13:55] Daruma Boa: mh, i must leave. #
[13:55] Paula Dix: agree, the belief is necessary to orient the way to work
[13:55] Gemma Cleanslate: exactloy
[13:55] Gemma Cleanslate: at least now anyway
[13:55] Paula Dix: bye daruma
[13:55] Gemma Cleanslate: Bye
[13:55] herman Bergson: Bye Daruma
[13:55] Daruma Boa: will see u thursday i hope.
[13:55] Jangle McElroy: Bye Daruma
[13:55] Rodney Handrick: Bye Gemma
[13:55] Ze Novikov: sequence vs. non-sequence at the end of the day still a machine...
[13:55] herman Bergson: I'll be here...deo volente ^_^
[13:56] Rodney Handrick: Bye Daruma
[13:56] Scope Cleaver: What do you mean Ze?
[13:56] Gemma Cleanslate: Bye
[13:56] Gemma Cleanslate: oh i was saying to daruma lollol
[13:56] Scope Cleaver: I don't get the sequence vs. non-sequence, do you mean in the comp-sci way?
[13:56] Ze Novikov: yep
[13:56] Scope Cleaver: i.e. machines can operate in parallel
[13:56] Scope Cleaver: and massively parallel
[13:56] Qwark Allen: i know
[13:57] Ze Novikov: does not change anything
[13:57] Scope Cleaver: So whats your point?
[13:57] herman Bergson: Fuzzy logic, stochatic procedures..you can program it all
[13:57] Ze Novikov: sequence
[13:57] Ze Novikov: non sequence
[13:57] Ze Novikov: that is not thinking
[13:57] Scope Cleaver: Well how do you suppose the brain works?
[13:57] Ze Novikov: no more than you
[13:57] herman Bergson: Ze. computers don't think..that still is my thesis
[13:57] Ze Novikov: i am explaining the machine
[13:58] Samuel Okelly: has the notion of free-will been abandoned?
[13:58] Jangle McElroy: Perhaps read Idoru by William Gibson if you enjoy storues about computers that are more human than human. It is enjoyable.
[13:58] Scope Cleaver: I agree with herman, maybe they do on a very primitive level
[13:58] Paula Dix: Ze, couldnt that be the first steps toward real thinking?
[13:58] herman Bergson: That is such an issue Samuel
[13:58] Gemma Cleanslate: oh gosh
[13:58] herman Bergson: Not even on a primitive level....
[13:58] Ze Novikov: nope look at the way they work
[13:59] herman Bergson: the working of a computer can not be called thinking
[13:59] Scope Cleaver: I beg to differ
[13:59] Ze Novikov: i agree herman
[13:59] herman Bergson: Yes we should Scope ^_^
[13:59] Gemma Cleanslate: me too herman
[13:59] Ze Novikov: it is not thinking
[13:59] Paula Dix: when they have a power of processing equivalent to a human brain it will be thinking
[13:59] Scope Cleaver: Everytime we come up with a definition of thinking and a computer emulates it. we struggle to find something else ㋡
[13:59] Ze Novikov: nope it does not
[13:59] Scope Cleaver: A lot of people said ...
[13:59] Ze Novikov: it imitates
[13:59] Paula Dix: why?
[13:59] Ze Novikov: that is not thinking
[13:59] Paula Dix: but doenst a child brain imitates its parents??
[13:59] Scope Cleaver: when a computer can beat the world champion at chess it will have some thinking...
[14:00] Paula Dix: its build the same way, and trained by parents
[14:00] Ze Novikov: he did it by random numbers
[14:00] Scope Cleaver: Ze you can't hold that kind of argument unless you come clean on what you *mean* by thinking.
[14:00] Ze Novikov: look at the program used
[14:00] herman Bergson: That is the point Scope...it looks like thinking..
[14:00] Samuel Okelly: how can we program intuitiveness? modern computers cant even produce 1 single random number yet
[14:00] Gemma Cleanslate: until it gets to about 13 paula
[14:00] Scope Cleaver: But thats begging the question herman
[14:00] Paula Dix: lol Gemma
[14:00] Scope Cleaver: For all I know you *look* like you are thinking too
[14:01] herman Bergson: In a way ..yes...again the definition issue here
[14:01] Scope Cleaver: But we aren't having the solipsistic discussion are we?
[14:01] Gemma Cleanslate: time to go
[14:01] Gemma Cleanslate: lol
[14:01] Jangle McElroy: Bye gemma
[14:01] herman Bergson: Well.....Be well Gemma...I'll get ready for next Thursday :-)
[14:02] Jangle McElroy: If to err is human, than perhaps computers are already intelligent?
[14:02] Paula Dix: lol
[14:02] herman Bergson: May I thank you all for this inspiring discussion....
[14:02] Gemma Cleanslate: certainly was lo
[14:02] Samuel Okelly: i should go too .. tc every1 :)
[14:02] Gemma Cleanslate: see you thursday
[14:02] Jangle McElroy: See you all soon
[14:02] Scope Cleaver: See you all!
[14:02] Paula Dix: bye!
[14:02] Ze Novikov: Must be off.... ty herman bb all
[14:02] Gemma Cleanslate: we hope
[14:02] Jangle McElroy: Thanks HErman
[14:02] Gemma Cleanslate: Bye
ELIZA: Hello.
ELIZA: How do you do. Please state your problem.
Me : I must find an answer to an important question
ELIZA: You say you must find an answer to an important question ?
Me : yes
ELIZA: You seem to be quite positive.
Me : Can computers think?
ELIZA: Do computers worry you ?
Me : Absolutely
ELIZA: I'm not sure I understand you fully.
Me : The idea that computers can think scares me
ELIZA: Why do you mention computers ?
Me : Because a thinking computer scares me
ELIZA: What do you think machines have to do with your problem ?
Me : Don't you listen?
ELIZA: We were discussing you -- not me.
Me : I am afraid of thinking computers
ELIZA: Don't you think computers can help people ?
Me : Maybe
ELIZA: You don't seem quite certain.
Me : no
ELIZA: Are you saying no just to be negative?
Me : I say no because I am not certain
ELIZA: Is it because you are not certain that you came to me ?
Me : yes
ELIZA: You are sure.
Me : I am
ELIZA: How long have you been ?
Me : all the time
ELIZA: Please go on.
Me : Are you a computer?
ELIZA: What about machines worrys you ?
Me : machines have no soul
ELIZA: You are being a bit negative.
Me : it is true
ELIZA: What does that suggest to you ?
Me : It suggests that machines are not human
ELIZA: Did you think they might not be not human ?
Me : They are not human
ELIZA: Would you like it if they were not not human ?
Me : Oh yes
ELIZA: I see.
See: http://www.chayden.net/eliza/Eliza.html
She understood and took the burden from my shoulders. I could go on with my preparations and reading.
I am a great fan of Weizenbaum. He died in 2008 in Berlin. He was the creator of the computerprogram ELIZA, which acts as an empathic psychologist. Conversations with her can be really hilarious.
Some people really took ELIZA serious and openen their heart for her, discussing intimate personal problems. But it all began with Alan Turing, who wrote in 1950 a paper named "Computing Machinery and Intelligence".
He begins his article thus: "I propose to consider the question "Can machines think?" This should begin with definitions of the meaning of the terms 'machine' and 'think'. The definitions might be framed so as to reflect so far as possible the normal use of the words, but this attitude is dangerous."
Why is this attitude dangerous? Simple, there is no proper definition for "thinking" availabe. But then it goes wrong, when he says "Instead of attempting such a definition I shall replace the question by another, which is closely related to it and is expressed in relatively unambiguous words."
Turing comes with the appealing question "Can machines think?" A question that drives me into the arms of ELIZA, but there could a hidden agenda too ;-) and then suggest to replace it with another question. What in the world may mean CLOSELY related?
Anyway, his new question is something like "Are there imaginable digital computers which would do well in the [Turing test]"? The picture here (also found at http://en.wikipedia.org/wiki/Turing_test#cite_note-2) shows you the set up of the test.
So this is closely related "thinking"? The fact that a computer can show the same conversational behavior as a human being? The Turing test is based on the assumption that human beings can judge a machine's intelligence by comparing its behaviour with human behaviour.
There it is....intelligence. There neither is a proper definition for intelligence. What kind of intelligence? Didn't we learn about emotional intelligence by Goodman's book, or social intelligence?
I think, that Turing was trapped in his historical context. Just imagine, though gigantic machines, they had the first electric machines which were able to perform complex calculations, just like a human can.
And then we have our imagination.......if this is possible, there definitely is more to come. When I am not mistaken, the computer Turing describes in his test, wasn't even technically possible in 1950, only theoretically.
The Turing Test has become a classic in the history of artificial intelligence and Weizenbaum's ELIZA would be a candidate to pass the test. But philosophically it doesn't bring much if it is regarding the question "Can my computer think?"
But we live in a time period in which people more and more tend to believe that computers can think. Did it never occur to you? You call the office and say "You sent me a letter telling me that I haven't paid my bill."
"Yes we did, sir" " But I have paid my bill, weeks ago." "I am sorry Sir, but the computer says that you have not paid your bill, so we sent you that reminder." "But your computer is wrong!" "I am sorry Sir, it is says it here on my screen: you have not yet paid."
It can be even more bizarre. Here in the Netherlands we have a TV commercial like this; Young couple in bar. She just met him. Is he the perfect partner? Send an SMS message with your name and his name to number XXXX and the computer gives you the answer in no time.
Girl does so. Score 3% !!! As a conclusion she throws the content of her glass in the face of the young man.
Or this one: Send your name and your wifes name to number XXXX and the computer gives you in no time the perfect name for your baby.
Something is wrong here with man and computer. So it wont be overkill to spend another lecture on the philosophy of Artificial Intelligence: AI
John McCarthy, who coined the term in 1956, defines it as "the science and engineering of making intelligent machines." But is intelligence the product of the interaction of all kinds of materials, put together in a factory, or is it product of the fact that we are dealing LIVING matter?
The Discussion
[13:23] herman Bergson: If you have any questions or remarks.....
[13:23] Scope Cleaver: Floor's open?
[13:23] Gemma Cleanslate: already?
[13:23] herman Bergson: go ahead :-)
[13:24] Laila Schuman: i find it necessary to get that definition of intelligence... just as we need a definition of education... depends upon whom is speaking
[13:24] Scope Cleaver: Well I think the Turing test is still very relevant today and better than anything we have so far
[13:24] Scope Cleaver: That definition of intelligence I think is a bit tricky
[13:24] Ludwig John: sometimes I suppose that my writing programm is intelligent: I write some letters and it completes the whole word
[13:24] Laila Schuman: yes
[13:24] herman Bergson: Well in the world of AI it has not so much meaning
[13:24] Laila Schuman: agreed
[13:24] Scope Cleaver: If I introduce you to someone and ask you, do you find that person intelligent, you don't go into definitions of intelligence
[13:24] Laila Schuman: no but i have opinions
[13:25] Scope Cleaver: intelligence is about human performance
[13:25] Paula Dix: ive read a part of a Dennet book where he talk about turing test, very interesting
[13:25] Scope Cleaver: Machine intelligence is about how well it emulates that
[13:25] Laila Schuman: and a lot depends upon whether they or i are significantly more or less intelligent than the other
[13:25] herman Bergson: yes if you define intelligence in a behavioristic or functional way
[13:25] Scope Cleaver: Thats the generally received view yes
[13:26] herman Bergson: Yes...so we call emulation intelligent behavior
[13:26] Scope Cleaver: Note that it's one thing to have an intelligent thinking machine
[13:26] Scope Cleaver: that doesnt mean it's conscious or that even has to be
[13:26] herman Bergson: But the reat question was Can machines think
[13:26] Laila Schuman: finding the result of a math problem or giving a reply based on grammar...does not mean one doing it can apply it
[13:26] Scope Cleaver: Yes they can
[13:27] Gemma Cleanslate: and will they ever really be able to
[13:27] Scope Cleaver: Unless we're willing to debate whether anything *can* think, machines certainly can.
[13:27] Paula Dix: When ive seen this would be the theme for today, i put this question to the people here at home, and Fat said its unfair to judge computers in the state they are, since they are so limited
[13:27] herman Bergson: If so ...it is only a matter of definition
[13:27] Paula Dix: He say they are intelligent, but something in the level of a lizzard :)))
[13:27] Scope Cleaver: Yes you can debate semantics of course
[13:27] Scope Cleaver: But like I said it's interesting that people dont' do that when you ask them to judge if another person is intelilgent.
[13:28] Jangle McElroy: I guess it might be easier to suggest machines can calculate based on rules, where as man can create without rules.
[13:28] Scope Cleaver: they seem to have a sensible sense of what intelligent means when asked about someone else but then discuss and debate semantics when it's a machine.
[13:28] herman Bergson: Another person you say....not another computer,,,that is the difference
[13:28] Gemma Cleanslate: that makes sense to me jangle
[13:28] Paula Dix: Scope but we see all the time people passing judgments because someone was born here or there and so on, its the same thing
[13:28] Scope Cleaver: But thats the beauty of the Turing test
[13:28] Scope Cleaver: it doesn't tell you at the onset which is which
[13:29] Scope Cleaver: You only interact with the other entity via teletext
[13:29] herman Bergson: Yes indeed....
[13:29] Scope Cleaver: But whether you pass judgment or not is completely immaterial to whether an entity is or isn't intelligent.
[13:29] herman Bergson: But Turing had to change his question...and had to drop the fundamental concept of thinking
[13:30] Scope Cleaver: What matters is, is there a test we can devise that would realiably tell us if an entity is intelligent
[13:30] Paula Dix: oh, you dont know what we brazilians think about portugueses!!!! :))))
[13:30] herman Bergson: Yes Scope..from a pragmatic point of view..if it works it works...whether it is a computer or a man doing the job
[13:30] Ze Novikov: lol
[13:30] Scope Cleaver: Right
[13:31] herman Bergson: But the question Can computers think suggest much more...
[13:31] Scope Cleaver: It's a very complex question, but I think the interesting part reside in our formulating the question the right way, somehow the turing test helps us do that
[13:31] herman Bergson: it is close to the idea..Have computers a mind?
[13:31] Scope Cleaver: it's has flaws of course but it's overall a good test
[13:31] Scope Cleaver: hehe yes indeed but the word "mind" opens a big can of worms semantically speaking
[13:31] Scope Cleaver: For instance you'd get a different definition for everyone in this room
[13:31] herman Bergson: A test of what...? That human behavior can be emulated by machines?
[13:31] herman Bergson: We have factories ful of robots
[13:32] Scope Cleaver: A test for intelligence
[13:32] Scope Cleaver: Which is plain performance of cognitive functions
[13:32] herman Bergson: No....for it only is about emulating huma conversatin
[13:32] Scope Cleaver: memory, reasoning, inference, and so on
[13:32] Scope Cleaver: the human conversation part is just a tool
[13:32] herman Bergson: And I would define intelligence as the ability to have a converstaion
[13:32] Paula Dix: I would say the moment computers have capacity similar to a human brain the question will be solved
[13:33] Scope Cleaver: I belive it' s more than that but the point is, a conversation can help you determine if an entity is intelligent
[13:33] Scope Cleaver: we'd have to establish necessary and sufficient conditions
[13:33] herman Bergson: memory, reasoning, inference yes....but imagination, intentionality, emotions..no
[13:33] Scope Cleaver: Thats a scientic question herman
[13:33] herman Bergson: also part of us as intelligent beings
[13:34] herman Bergson: It is an observation
[13:34] Scope Cleaver: For instance it isn't clear how *we* aren't machines that in itself is very debetable
[13:34] Paula Dix: but we are machines
[13:34] Jangle McElroy: Intelligence tests often rely on observation and calculation. Making a calculator with a large memory capable of simulating intelligence.
[13:34] Scope Cleaver: Like I said it's debatable
[13:34] Scope Cleaver: The general received view is that we are.
[13:35] Paula Dix: the divisions, the labels, are all artificial
[13:35] Scope Cleaver: it depends what your materialistic position is
[13:35] herman Bergson: The Turing Test is just a very narrow view on what thinking entails,I would say
[13:35] Scope Cleaver: It's just a tool
[13:35] Paula Dix: i bet tons of humans wouldnt pass turing test
[13:35] herman Bergson: but completely outdated...
[13:35] Scope Cleaver: You see it has to have something ont he table that *both* humans and machine could do
[13:36] Scope Cleaver: communicating through text is one
[13:36] Laila Schuman: is an education the ability to pass tests... or is there something within the educated person that has much more going on... whether they pass tests or not
[13:36] Scope Cleaver: there is a clear reason why he choose to have text via teletext instead of voice for instance or acting
[13:36] Ze Novikov: in the end it is about sequencial series of data or non sequential...
[13:36] Paula Dix: dennett mentions a situation where people couldnt pick the machine on the test
[13:37] Scope Cleaver: Yes paula when that happens
[13:37] Scope Cleaver: when you results are 50/50 you can safely say you have true AI
[13:37] Laila Schuman: can a computer tell a joke that is not fed to it
[13:37] Scope Cleaver: It could why not?
[13:37] Gemma Cleanslate: to me it demonstrates that a machine can be intelligent but does it really have intelligence if it does not have feelings too
[13:37] Paula Dix: they made a program to simulate paranoia, and psychiatrists had to say who was the machine between real paranoics
[13:37] herman Bergson: It could?
[13:37] Scope Cleaver: Can a bunch of gray matter cells tell a joke thats not fed to it? haha
[13:37] Paula Dix: they couldnt, because they couldnt go really deep on the questions for fear of causing damage to the persons
[13:38] Paula Dix: so they all got it wrong
[13:38] herman Bergson: Yes it can Scope...you can think of a new joke..a play of words and so on
[13:38] Scope Cleaver: in that case yes, I think a machine could do that as well ㋡
[13:38] Paula Dix: lol somewhere the jokes where invented
[13:38] Scope Cleaver: Exactly my point
[13:39] Scope Cleaver: So there is cognitive work being done to come up with a joke
[13:39] herman Bergson: Yes..jokes are invented..but not rule based, but by deviating from rules
[13:39] Ludwig John: can a computer write a novel or a poem?
[13:39] Scope Cleaver: Something an AI could do
[13:39] Jangle McElroy: Tell a joke? Or create a joke? Creation is perhaps closer to free will and intelligence than calculation. It's a bit like the tests that show computers can write music - it often sounds fomulaic rather than freeform and flowing.
[13:39] Scope Cleaver: Ludwig yes
[13:39] Scope Cleaver: and in fact, theoretically it doesn't even need to be intelligent to do it
[13:39] Scope Cleaver: it could do it by fluke hehe ㋡
[13:39] herman Bergson: Yes Jangle....that is the case indeed
[13:40] Paula Dix: but its again the same point, computers "brains" are still too small
[13:40] Paula Dix: they do music or anything with this small capacity
[13:40] Scope Cleaver: Someone mentioned Dennett earlier I think he has that monkey typing a play by fluke as an example
[13:40] Paula Dix: surely not like us, but that will change with time
[13:40] Paula Dix: No that was Dawkins :))
[13:41] Scope Cleaver: We're a bunch of decades away but I don't think the question of when is that interesting
[13:41] herman Bergson: We'll definitely dig into those issues the next lecture..:-)
[13:41] herman Bergson: Well the bet is set to 2029 now
[13:41] Paula Dix: i wonder, someone here think computers cant thing?
[13:41] Qwark Allen: yes
[13:41] Scope Cleaver: i think thats way early hehe
[13:41] Qwark Allen: close to that hermman
[13:41] Gemma Cleanslate: i do not think they can tghink
[13:41] Daruma Boa: i think they think^
[13:42] Qwark Allen: they will
[13:42] Daruma Boa: my mac does^^
[13:42] Qwark Allen: in time
[13:42] Scope Cleaver: I wont' pass *my* turing test by 2029 not to be presumptuous lol
[13:42] herman Bergson: No...in my opinion computer cant think...
[13:42] Paula Dix: my vote is for yes too :))
[13:42] Gemma Cleanslate: i think they only think what we tell them to think
[13:42] Gemma Cleanslate: so far anyway
[13:42] Jangle McElroy: :)
[13:42] herman Bergson: the perfor rulebased actions
[13:42] Daruma Boa: no, these days there r programms, which learn
[13:42] Paula Dix: but herman our brains also dont work like that?
[13:42] Ze Novikov: every evening I fight with my computer .." Now time to turn you off and go to bed....."
[13:43] Scope Cleaver: What is of interest is that *once* they are as intelligent as we are... they won't stay that miserable for long
[13:43] herman Bergson: no....the idea that the brain is a computer is a nice metaphor but a flaw
[13:43] Laila Schuman: lol
[13:43] Paula Dix: lol imagine when computers are bigger and start to have personalities?? :)))
[13:43] Gemma Cleanslate: or much smaller and do that
[13:43] Daruma Boa: right scope^^
[13:43] Paula Dix: why herman??
[13:43] Scope Cleaver: I think the substrate question is tangential
[13:43] herman Bergson: No. Paula....the size of a machine doesnt count....
[13:43] Scope Cleaver: what intelligence is implement *in* doesn't really matter
[13:43] herman Bergson: To begin with it is Life that counts..
[13:44] herman Bergson: Our mental states are caused by the fact that we are alive
[13:44] Scope Cleaver: herman do you belive there is a quality to life such as an "elan vital"?
[13:44] Jangle McElroy: If computers are ever capable of sentient independent thought, I wonder if they will be like the Orangutan? The human tribes people of Borneo say Orangutans only don't speak because someone would give them a broom and order them to sweep up if they could communicate.
[13:44] Scope Cleaver: A something above the atoms that makes our being?
[13:44] herman Bergson: When you put all chemicals together that consitute life....it still hasnt happen to create living material
[13:45] herman Bergson: No Scope....
[13:45] herman Bergson: It must be those atoms, I think...
[13:45] Scope Cleaver: So what is the magic property?
[13:45] Scope Cleaver: and why machines can't have it?
[13:45] Paula Dix: we are going toward biological computers... will them be alive?
[13:45] herman Bergson: If you knew that we could create living matter in the laboratory
[13:46] Scope Cleaver: I belive thats flawed but thats my opinion.
[13:46] herman Bergson: We may be going to biochemical computers..that's the difference
[13:46] Paula Dix: i think like scope, no difference if its carbon or metal
[13:47] Paula Dix: or silica :))))
[13:47] herman Bergson: Ok.....good material for the next lecture...:-)
[13:47] Scope Cleaver: You'll often find difference in the metaphysics of being is rooted in deep beliefs...
[13:47] Paula Dix: lol yes, thats fun
[13:47] Scope Cleaver: About other stuff ㋡
[13:48] Scope Cleaver: We hope Eliza didn't cause you too many turmoils herman
[13:48] herman Bergson: Well she is a great help :-)
[13:48] Scope Cleaver: Could you believe how many people she helped.
[13:48] Scope Cleaver: Maybe more than any single *human* therapist hehe
[13:48] Gemma Cleanslate: she is a loop of nothing lol
[13:49] herman Bergson: Some conversations...especially with an improved version can have a positive behavioral effect, I think
[13:49] Gemma Cleanslate: yes lol
[13:49] Gemma Cleanslate: is fun
[13:49] Scope Cleaver: Thanks for the stimulating lecture
[13:49] herman Bergson: You should visit Elbot....
[13:49] Daruma Boa: here?
[13:49] Paula Dix: whats elbot?
[13:49] herman Bergson: He is fun to talk with
[13:49] Paula Dix: oh, another program?
[13:49] herman Bergson: Just google on ELBOT
[13:49] Paula Dix: ok
[13:50] herman Bergson: He is another chatbot
[13:50] Daruma Boa: ok will do
[13:50] herman Bergson: Won the Loebner prize
[13:50] Scope Cleaver: *nods*
[13:50] Paula Dix: dennet mention another funnny experiment, a program made to think it was some american politician
[13:50] Scope Cleaver: We'll start to worry when it's more fun to talk to a bot than our spouses
[13:50] Paula Dix: (I think it was american...)
[13:50] Samuel Okelly: i found elbot like having a conversation with a child who wont take its eyes of the TV to talk to you ;-)
[13:50] Scope Cleaver: lol Samuel
[13:51] herman Bergson: Yes Scope.....then we have s lost scope indeed
[13:51] Scope Cleaver: Ai is very childlike which is why it isn't fully Ai yet
[13:51] Paula Dix: exact
[13:51] Scope Cleaver: But hey it beat Kasparov at chess, first step to everything
[13:51] herman Bergson: I dont like that argument.....
[13:51] Paula Dix: talking TV, loved the peanuts cartoon!
[13:51] Daruma Boa: why herman?
[13:51] herman Bergson: the argument of poverty...
[13:51] Scope Cleaver: I meant to say it's in developmental stage
[13:52] Scope Cleaver: Learning is a very big component to AI
[13:52] herman Bergson: Well ..it is the same as the belief that eventually Science will solve all problems
[13:52] Daruma Boa: why is it pover?
[13:52] Scope Cleaver: the reasearchers are still baning their head in implementing better learning algorithms in AI
[13:52] Paula Dix: yes, its a belief
[13:52] Paula Dix: should i add "but a right one"? :)))))
[13:53] herman Bergson: It is poverty as it states..oh I am a poor guy now, but wait..I will get rich :-)
[13:53] herman Bergson: When we evaluate AI we have to evaluate it as it IS.
[13:53] herman Bergson: Not on what it promisses to become
[13:53] Scope Cleaver: But thats the beauty of science it can build upon it's own knowledge and discovery
[13:53] Scope Cleaver: yes I agree on that point
[13:54] Paula Dix: yes, its true, but we think it can do this or that and this is a belief
[13:54] Scope Cleaver: But whats important about the "what ir promises to become" part is that we can use that to start now, to think about the ethical problems we will face.
[13:54] Samuel Okelly: science often argues from what may be
[13:54] herman Bergson: Yes...that is what we humans can....
[13:54] herman Bergson: Have hope and expectation...how to program that in a computer?
[13:55] Daruma Boa: mh, i must leave. #
[13:55] Paula Dix: agree, the belief is necessary to orient the way to work
[13:55] Gemma Cleanslate: exactloy
[13:55] Gemma Cleanslate: at least now anyway
[13:55] Paula Dix: bye daruma
[13:55] Gemma Cleanslate: Bye
[13:55] herman Bergson: Bye Daruma
[13:55] Daruma Boa: will see u thursday i hope.
[13:55] Jangle McElroy: Bye Daruma
[13:55] Rodney Handrick: Bye Gemma
[13:55] Ze Novikov: sequence vs. non-sequence at the end of the day still a machine...
[13:55] herman Bergson: I'll be here...deo volente ^_^
[13:56] Rodney Handrick: Bye Daruma
[13:56] Scope Cleaver: What do you mean Ze?
[13:56] Gemma Cleanslate: Bye
[13:56] Gemma Cleanslate: oh i was saying to daruma lollol
[13:56] Scope Cleaver: I don't get the sequence vs. non-sequence, do you mean in the comp-sci way?
[13:56] Ze Novikov: yep
[13:56] Scope Cleaver: i.e. machines can operate in parallel
[13:56] Scope Cleaver: and massively parallel
[13:56] Qwark Allen: i know
[13:57] Ze Novikov: does not change anything
[13:57] Scope Cleaver: So whats your point?
[13:57] herman Bergson: Fuzzy logic, stochatic procedures..you can program it all
[13:57] Ze Novikov: sequence
[13:57] Ze Novikov: non sequence
[13:57] Ze Novikov: that is not thinking
[13:57] Scope Cleaver: Well how do you suppose the brain works?
[13:57] Ze Novikov: no more than you
[13:57] herman Bergson: Ze. computers don't think..that still is my thesis
[13:57] Ze Novikov: i am explaining the machine
[13:58] Samuel Okelly: has the notion of free-will been abandoned?
[13:58] Jangle McElroy: Perhaps read Idoru by William Gibson if you enjoy storues about computers that are more human than human. It is enjoyable.
[13:58] Scope Cleaver: I agree with herman, maybe they do on a very primitive level
[13:58] Paula Dix: Ze, couldnt that be the first steps toward real thinking?
[13:58] herman Bergson: That is such an issue Samuel
[13:58] Gemma Cleanslate: oh gosh
[13:58] herman Bergson: Not even on a primitive level....
[13:58] Ze Novikov: nope look at the way they work
[13:59] herman Bergson: the working of a computer can not be called thinking
[13:59] Scope Cleaver: I beg to differ
[13:59] Ze Novikov: i agree herman
[13:59] herman Bergson: Yes we should Scope ^_^
[13:59] Gemma Cleanslate: me too herman
[13:59] Ze Novikov: it is not thinking
[13:59] Paula Dix: when they have a power of processing equivalent to a human brain it will be thinking
[13:59] Scope Cleaver: Everytime we come up with a definition of thinking and a computer emulates it. we struggle to find something else ㋡
[13:59] Ze Novikov: nope it does not
[13:59] Scope Cleaver: A lot of people said ...
[13:59] Ze Novikov: it imitates
[13:59] Paula Dix: why?
[13:59] Ze Novikov: that is not thinking
[13:59] Paula Dix: but doenst a child brain imitates its parents??
[13:59] Scope Cleaver: when a computer can beat the world champion at chess it will have some thinking...
[14:00] Paula Dix: its build the same way, and trained by parents
[14:00] Ze Novikov: he did it by random numbers
[14:00] Scope Cleaver: Ze you can't hold that kind of argument unless you come clean on what you *mean* by thinking.
[14:00] Ze Novikov: look at the program used
[14:00] herman Bergson: That is the point Scope...it looks like thinking..
[14:00] Samuel Okelly: how can we program intuitiveness? modern computers cant even produce 1 single random number yet
[14:00] Gemma Cleanslate: until it gets to about 13 paula
[14:00] Scope Cleaver: But thats begging the question herman
[14:00] Paula Dix: lol Gemma
[14:00] Scope Cleaver: For all I know you *look* like you are thinking too
[14:01] herman Bergson: In a way ..yes...again the definition issue here
[14:01] Scope Cleaver: But we aren't having the solipsistic discussion are we?
[14:01] Gemma Cleanslate: time to go
[14:01] Gemma Cleanslate: lol
[14:01] Jangle McElroy: Bye gemma
[14:01] herman Bergson: Well.....Be well Gemma...I'll get ready for next Thursday :-)
[14:02] Jangle McElroy: If to err is human, than perhaps computers are already intelligent?
[14:02] Paula Dix: lol
[14:02] herman Bergson: May I thank you all for this inspiring discussion....
[14:02] Gemma Cleanslate: certainly was lo
[14:02] Samuel Okelly: i should go too .. tc every1 :)
[14:02] Gemma Cleanslate: see you thursday
[14:02] Jangle McElroy: See you all soon
[14:02] Scope Cleaver: See you all!
[14:02] Paula Dix: bye!
[14:02] Ze Novikov: Must be off.... ty herman bb all
[14:02] Gemma Cleanslate: we hope
[14:02] Jangle McElroy: Thanks HErman
[14:02] Gemma Cleanslate: Bye
Subscribe to:
Posts (Atom)
