Can machines think?
Alan Turing and AI
1
"If a computer could think, how could we tell?"
Alan Turing
(1912-54)
•
Cambridge Mathematician
•
Code-breaker during WW2
(Enigma)
•
convicted after admitting a
sexual relation to another man.
Suicide in ’54.
•
Turing Machine
•
Turing Test (AI)
The paper in our book is “Computing Machinery and Intelligence” from 1950
2
Language. Some of early computing was text processing, though of highly
mathematical sort. Allan T worked on decrypt encoded language.
Loebner prize is an award for whoever can produce a machine (software)
than can beat the turing test. (dr eliza ai chat in on the cd with our book)
Turing machines are extremely basic abstract symbol-manipulating devices which, despite their
simplicity, can be adapted to simulate the logic of any computer that could possibly be constructed. (as
far as it is what we call a computer. If a computer code can run on these machines, we say the
compute. thus the expression “does not compute” when something is wrong..
The chapter "Turing Conceives of the All-Purpose Computer" in The Universal Computer by Martin
Davis (Norton 2000) is a good description of Turing Machines. (compendium)
•
The imitation game (later known as the turing test)
•
•
as a tool to distinguish between human and machine, in a playful manner
where the machine must prove it’s “brains”. Originally a party game to
distinguish between man and woman.
3 people. One man (A), one woman (B) and an
interrogator (C) of either sex.
•
The interrogator is a a separate room, and his/hers task is to figure out the sex
of the two others.
He addresses the two as X and Y, and at the end of the game he must say “X is
B”, “X is A”, “Y is A”, “Y is B”.
•
The interrogator is allowed to ask questions of the sort: Will X please tell me
the length of his or her hair.
3
“I propose to consider the question, ‘can machines think?’”
need to Define machine, define think.
To make it harder for the interrogator the voice is hidden by text, or by teleprompter. Today we use simple
chat software (MSN messenger, (m)IRC ol)
Critique of the new problem
“are there imaginable digital computers
whitch would do well in the imitation
game?”
•
• “Let us fix our attention on one particular
computer C. Is it true that by modifying
this computer to have an adequate storage,
suitably increasing its speed of action, and
providing it with an appropriate
programme, C can be made to play
satisfactory the part of A in the imitation
game, the part of B being taken by a man?
4
To the definition of machine, we only allow digital computers or electronic computer. We must accept
that the digital computer is based on the same concepts and Babbages analytical engine (planned
1828-39 which was entirely mechanical), and (I’m going real fast through the text here..) they are
descrete sated machines (has only a limited numbers of states (resources) i thing cpu, hard drive space
and ram is included here)
We must accept that machines will preform much better in matematical logic, but struggle with
semantics, feelings and other humans behavours. A computer can multiply 999.999 by 12.34 easily,
but have no knowledge of being tired, happy or angry. I will beat the average Joe in chess every time,
but does not see the beauty of the craftsmanship on the pieces.
This makes the software used a bit tricky, and we build in traps (mistakes and assumptions) into the
code. (spellingmisstakes, mathmatical errors, and absolute knowledge (to day the computer is tired))
Contrary Views on the Main Question
(1) Theological Objection
(2) 'Heads in the Sand' Objection
(3) Mathematical Objection
(4) Argument From Consciousness
5
1 Theological Objection: This states that thinking is a function of man's immortal soul and therefore a machine can not
think. Turing replies by saying that he sees no reason why it would not be possible for God to grant a computer a soul if
He so wished.
2 'Heads in the Sand' Objection: "The consequences of machines thinking would be too dreadful. Let us hope and
believe that they cannot do so." This objection is a fallacious appeal to consequences, confusing what should not be
with what can or cannot be. Closely connected to the prior.
3 Mathematical Objections: This objection uses mathematical theorems, such as Gödel's incompleteness theorem, to
show that there are limits to what questions a computer system based on logic can answer. Turing suggests that
humans are too often wrong themselves and pleased at the fallibility of a machine. (This argument would be made again by
philosopher John Lucas in 1961 and physicist Roger Penrose in 1989.)[2] What do you think of Piccaso?
4 Argument From Consciousness: This argument, suggested by Professor Geoffrey Jefferson in his 1949 Lister
Oration states that "not until a machine can write a sonnet or compose a concerto because of thoughts and
emotions felt, and not by the chance fall of symbols, could we agree that machine equals brain."[3] Turing replies by
saying that we have no way of knowing that any individual other than ourselves experiences emotions, and that
therefore we should accept the test and adds "I do not wish to give the impression that I think there is no mystery about
consciousness ... [b]ut I do not think these mysteries necessarily need to be solved before we can answer the question
[of whether machines can think]."
basically that a machine must be human to be intelligent...
Contrary Views on the Main Question
(5) Arguments from various disabilities.
(6) Lady Lovelace's Objection
(7) Argument from continuity in the nervous
system
6
5 Arguments from various disabilities. These arguments all have the form "a computer will never do X". Turing offers a selection:
Be kind, resourceful, beautiful, friendly, have initiative, have a sense of humour, tell right from wrong, make mistakes, fall in love, enjoy
strawberries and cream, make some one fall in love with it, learn from experience, use words properly, be the subject of its own
thought, have as much diversity of behaviour as a man, do something really new.
Turing notes that "no support is usually offered for these statements," and that they depend on naive assumptions about how versatile
machines may be in the future, or are "disguised forms of the argument from consciousness." He chooses to answer a few of them:
1. Machines cannot make mistakes. He notes it's easy to program a machine to appear to make a mistake.
2. A machine cannot be the subject of its own thought (or can't be self-aware). A program which can report on its internal states and
processes, in the simple sense of a debugger program, can certainly be written. Turing asserts "a machine can undoubtably be its own
subject matter."
3. A machine cannot have much diversity of behaviour. He notes that, with enough storage capacity, a computer can behave in a
astronomical number of different ways.
6 Lady Lovelace's Objection of the most famous objections states that computers are incapable of originality. This is largely
because, according to Ada Lovelace, machines are incapable of independent learning. Turing contradicts this by arguing that Lady
Lovelace's assumption was affected by the context from which she wrote, and if exposed to more contemporary scientific knowledge,
it would become evident that the brain's storage is quite similar to that of a computer.
“a machine can never ‘take us by surprise’”.
Turing further replies that computers could still surprise humans, in particular where the consequences of different facts are not
immediately recognizable.
7 Argument from continuity in the nervous system - “the nervous system is certainly not a discrete state machine”. it something
quite different. (almost opposite) Small mistakes will make rather large difference to the end reaction. (wasp). It is hard (impossible?)
to mimic this kind of behaviour with a discrete state machine. Reply: add “salt” or randomness to some tasks, like answering the
question what is PI (array(3.14, 3.12, 3.12, 3.14, 3.1416)
Contrary Views on the Main Question
(8) Argument from the informality of
behaviour
(9) The argument from Extra-sensory
perception
•
telepathy, clairvoyance, precognition and psycho-kinesis
7
8 Argument from the informality of behaviour argument states that any system governed by laws will be
predictable and therefore not truly intelligent. Turing replies by stating that this is confusing laws of behaviour
with general rules of conduct, and that if on a broad enough scale machine behaviour would become
increasingly difficult to predict. He argues that, just because we can't immediately see what the laws are,
does not mean that no such laws exist. He writes "we certainly know of no circumstances under which we
could say, 'we have searched enough. There are no such laws.'"
9 The argument from Extra-sensory perception Turing seems to be in defence thise “unscientific”
banches of behaviour. But argues that we could place the contestants in rooms that shields out this kind of
behaviour. So for this test, we can ignore them..
Learning Machines
Back to Lady Lovelace’s objection: a computer can only
do what we tell it to do.
How do we teach the computer?
How do we teach the computer to be critical?
Can a computer conduct deductive reasoning?
8
The last 4 pages is about if and how we can teach computers, as we can teach each other, and specifically
“can we teach a computer so that it can resemble the “brains” of a child?
Can a machine be super-critical? (aware of its own knowledge and assumptions)
How do we teach the computer? T wonders about punishment/reward learning (after Miss Hellen Kellers
case, we know that eyes and legs is unnecessary to become bright)
critical - the teacher says “do your homework”, who can we separate that as an imperative, and not just
another piece of information? “the teacher is always correct?”
conduct deductive reasoning?
?
“We can only see a short distance ahead, but we can
see plenty there that needs to be done.”
9
Hello, my name is [...]
and I’m a human.
10
The loebner prize. a annual contest where software (computers) compete in the Turing test.
none has got the Grand Prize of $100,000 and a Gold Medal (pictured above) for the first computer whose responses were
indistinguishable from a human'. That is what dr loebner promised in 1990.
Every year new bots test their chatting skills, and if we have time I’d like to let out Professor
do a 2min chat with a bot, just to check out what this is al about.
If we don’t have time, the.. eliza dr bot is chipped with our book, the new media reader (on
the cd/dvd thingy)
.
From Plans and Situated actions (Lucy A. Suchman)
Alice -Artificial Intelligence Foundation
11