of 26/26
Lecture 2 CSE 140 - Intro to Cognit ive Science 1 The Turing Test: Simulating Intelligence

Lecture 2CSE 140 - Intro to Cognitive Science1 The Turing Test: Simulating Intelligence

  • View
    213

  • Download
    0

Embed Size (px)

Text of Lecture 2CSE 140 - Intro to Cognitive Science1 The Turing Test: Simulating Intelligence

  • The Turing Test: Simulating Intelligence

    CSE 140 - Intro to Cognitive Science

  • Alan Turing (1912-1954)The Turing machine and the mathematization of the notion computable functionThe halting problemColossus & breaking the Enigma codeACE: Englands first large scale general computing machineDeveloped an influential mathematical model of embryological development

    Hodges, Andrew (1983) Alan Turing: The Enigma. Simon and Schuster, New York

    CSE 140 - Intro to Cognitive Science

  • Computing Machinery & Intelligence (1950)

    I PROPOSE to consider the question, 'Can machines think?' This should begin with definitions of the meaning of the terms 'machine and 'think'.

    CSE 140 - Intro to Cognitive Science

  • Think: The Imitation GameIt is played with three people, a man (A), a woman (B), and an interrogator (C). The interrogator stays in a room apart from the other two. The object of the game for the interrogator is to determine which of the other two is the man and which is the woman.The interrogator is allowed to put questions to A and B.It is A's object in the game to try and cause C to make the wrong identification. In order that tones of voice may not help the interrogator the answers should be typewritten.

    CSE 140 - Intro to Cognitive Science

  • The Imitation Game IIWe now ask the question, 'What will happen when a machine takes the part of A in this game?' Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, 'Can machines think?'

    CSE 140 - Intro to Cognitive Science

  • Machine: An Electronic ComputerThe question which we put in 1 will not be quite definite until we have specified what we mean by the word 'machine the present interest in 'thinking machines' has been aroused by a particular kind of machine, usually called an 'electronic computer' or 'digital computer'. Following this suggestion we only permit digital computers to take part in our game.

    CSE 140 - Intro to Cognitive Science

  • Summary of Turings TestCan an appropriately programmed digital computer consistently deceive a critical observer given that:

    The observer is free to ask the computer any question;The machine is free to lie;The observer distinguishes the machine from a human at chance.

    CSE 140 - Intro to Cognitive Science

  • The Turing Test: A Sufficient TestMay not machines carry out some-thing which ought to be described as thinking but which is very different from what a man does? This objection is a very strong one, but at least we can say that if, nevertheless, a machine can be constructed to play the imitation game satisfactorily, we need not be troubled by this objection.

    CSE 140 - Intro to Cognitive Science

  • A Key Implicit Claim

    A perfect simulation of intelligence would be indistinguishable from the real thing so that we would have no reason to say that the simulation is not intelligent.

    CSE 140 - Intro to Cognitive Science

  • What Is a Computing Machine?Motivation: A human computer doing arithmetic, e.G. Adding two large numbers.

    The human computer:Follows fixed rules, stored in a book altered when he is put on to a new job.Has an unlimited supply of paper.

    CSE 140 - Intro to Cognitive Science

  • A Computing Machine Consists ofAn executive unit, which carries out a fixed set of simple rules.A store, which is used as a notepad.To write down the results of its calculations.To remember which rules to use in which order.A control, which makes sure that the instructions are carried out correctly and in the right order.

    CSE 140 - Intro to Cognitive Science

  • Turings Question:Given that this model can simulate ANY digital computer (!!),Is human intelligence inside the set of computable functions; That is, the set of functions that can be computed by an algorithm

    CSE 140 - Intro to Cognitive Science

  • Will Computers Pass the Turing Test?Turings belief: in about 50 years (last year!!) It will be possible to program computers With a storage capacity of 109 bits (~100 megabytes)Guessing error rates of 30%After 5 minutes questioning

    CSE 140 - Intro to Cognitive Science

  • Potential Objections Raised by Turing

    CSE 140 - Intro to Cognitive Science

  • The Theological ObjectionThinking is a function of mans immortal soul. God has given an immortal soul to every man and woman, but not to any other animal or to machines. Hence no animal or machine can think.

    Should we not believe that He has the freedom to confer a soul on an elephant if he sees fit?We are in either case [constructing machines or procreation] instruments of His will providing mansions for the souls that He creates.Such [theological] arguments have been found unsatisfactory in the past.

    CSE 140 - Intro to Cognitive Science

  • The Heads in the Sand ObjectionThe consequences of machines thinking would be too dreadful. Let us hope and believe that they cannot do so.

    Alas.

    CSE 140 - Intro to Cognitive Science

  • The Mathematical ObjectionGdels theorem states that there are certain things that [a digital] machine cannot do. If it is rigged up to give answers to questions as in the imitation game, there will be some questions to which it will either give a wrong answer, or fail to give an answer at all

    It has only been stated, without any sort of proof, that no such limitations apply to the human intellect.

    CSE 140 - Intro to Cognitive Science

  • The Argument from ConsciousnessNot 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 No mechanism could feel pleasure at its successes, grief when its valves fuse, be warmed by flattery, be made miserable by its mistakes, be charmed by sex, be angry or depressed when it cannot get what it wants By this argument, the only which by which one could be sure that a machine thinks is to be the machine and to feel oneself thinking.

    CSE 140 - Intro to Cognitive Science

  • Arguments from Various Disabilities you will never be able to make one to do X. where X can be: make mistakes, enjoy strawberries and cream, do something novel, fall in love, make someone fall in love with it, tell right from wrong

    For each X, we are faced with an analytical problem. Is it really true that X lies outside our power to create algorithms for simulating X? As it stands, this objection simply asserts that something is impossible without offering any proof.

    CSE 140 - Intro to Cognitive Science

  • Lady Lovelaces ObjectionThe Analytical Engine has no pretensions of originate anything. It can do whatever we know how to order it to perform. (1842)One reading: machines will not surprise us. But Machines take me by surprise with great frequency.This is due to the [false] assumption that as soon as a fact is presented to the mind all consequences of that fact spring into the mind simultaneously with it.Otherwise, Good Point!!! Turing presents a framework for machine learning that is still with us today.

    CSE 140 - Intro to Cognitive Science

  • Argument from Continuity in the Nervous SystemThe nervous system is not a discrete-state machine [which the computing machine surely is]. Because it has continuous states, a discrete state machine cannot simulate a nervous system.This presupposes that the nervous system is capable of making infinitely fine distinctions which is false. We can digitize all sorts of continuous phenomena (witness CDs and DVDs); why cant we similarly digitize the signals of the nervous system?

    CSE 140 - Intro to Cognitive Science

  • Argument From Informality of BehaviorIt is not possible to produce a set of rules purporting to describe what a man should do in every conceivable set of circumstancesTo attempt to provide rules of conduct to cover every eventualityappears to be impossible.The argument seems to be If each person had a definite set of rules of conduct by which his or her life is regulated, then people would be no better than machines. There are no such rules, so people are not machines. No! Compare: If it is snowing, then I will be cold. It isnt snowing, so Im not cold.

    CSE 140 - Intro to Cognitive Science

  • Argument from ESP (!!)Unfortunately the statistical evidence, at least for telepathy, is overwhelming. A telepath might guess, for example, what card someone is holding, more often than a machine.To put the competitors in a telepathy-free room would satisfy all requirements.

    CSE 140 - Intro to Cognitive Science

  • ELIZAWritten by Weizenbaum in 1966Simulated a non-directive psychotherapistSample Dialogue:

    Men are all alike. IN WHAT WAY? They're always bugging us about something or other. CAN YOU THINK OF A SPECIFIC EXAMPLE? Well, my boyfriend made me come here. YOUR BOYFRIEND MADE YOU COME HERE He says I'm depressed much of the time. I AM SORRY TO HEAR YOU ARE DEPRESSED It's true. I am unhappy

    CSE 140 - Intro to Cognitive Science

  • ELIZA DO YOU THINK COMING HERE WILL HELP YOU NOT TO BE UNHAPPY I need some help, that much seems certain. WHAT WOULD IT MEAN TO YOU IF YOU GOT SOME HELP Perhaps I could learn to get along with my mother. TELL ME MORE ABOUT YOUR FAMILY My mother takes care of me. WHO ELSE IN YOUR FAMILY TAKES CARE OF YOU My father. YOUR FATHER You are like my father in some ways. WHAT RESEMBLANCE DO YOU SEE

    CSE 140 - Intro to Cognitive Science

  • Recent ObjectionsCompiled by David J. Chalmers, Department of Philosophy, University of Arizona, Tucson AZ 85721. E-mail: [email protected] Block, N. 1981. A look-up table could pass the Turing test, and surely isn't intelligent. The TT errs in testing behavior and not mechanisms. Moor, J. H. 1976. The basis of the Turing test is not an operational definition of thinking, but rather an inference to the best explanation Searle, J. R. 1980. Implementing a program is not sufficient for mentality, as someone could e.g. implement a "Chinese-speaking" program without understanding Chinese. So strong AI is false, and no program is sufficient for consciousness.Maudlin, T. 1989 Computational state is not sufficient for consciousness, as it can be instantiated by a mostly inert object.

    CSE 140 - Intro to Cognitive Science