1 ai - introduction bertil ekdahl [email protected]

39
1 AI - Introduction Bertil Ekdahl [email protected]

Upload: joel-lane

Post on 17-Dec-2015

224 views

Category:

Documents


2 download

TRANSCRIPT

1

AI - Introduction

Bertil Ekdahl

[email protected]

2

It is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable.

What is AI?

John McCarthy, http://www-formal.stanford.edu/jmc/whatisai/

3

When did it start?

• Alan Turing – 1947 lecture on intelligent machines.– 1950 article: Computing Machinery and

Intelligence.

4

The Turing Test

(Borrowed from Ola Flygt.)

5Penrose, Roger, “The Emperor’s New Mind”

A Turing Machine

6

Two symbols where B means “blank” and 1 is the tally symbol.

{ }1,B

7

ENIAC(Electronic Numerical Integrator And Calculator)

8

EDVAC(Electronic Discrete Variable Automatic Computer)

9

Dartmouth Artificial Intelligence (AI) Conference

• We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire.

• The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.

• Dartmouth AI Project proposal: J. McCarthy et al.; Aug. 31, 1955.

(http://www-formal.stanford.edu/jmc/history/dartmouth.html)

10

Why such a conjecture?

… every aspect of learning […] be so

precisely described that a machine

can be made to simulate it.

Tabula rasa

11

But the human intellect, which is the lowest in the order of intellects and the most removed from the perfection of the Divine intellect, is in potency with regard to things intelligible, and is at first "like a clean tablet on which nothing is written", as the Philosopher [Aristotle] says. (Thomas Aquinas, Summa Theologica 1.79.2, 1265 - 1274)

12

Connectionism Rumelhart & McClelland: generic

associationist networks, subjected

to massive amounts of training, could explain all of cognition.

13

Rumelhart & McClelland was left with the answer: Why is people smarter than rats?Answer: [The] question does seem a bit puzzling. […] another aspect to the difference […] is that the human environment includes other people and the CULTURAL devices that they have developed to organize their thinking process.(1986)

A problem

14

Why such a conjecture?…make machines use language, form abstractions and concepts …

No difference between formal language (programming language) and natural language?

15

Everything is a question of representation?

16

Roman digits

III

XV

17

Binary digits

11

1111

18

2

367654983274

10

100100...1shift out

19

Time

2h 30min 15s + 3h 45min 37s =?

20

Fifty-fiveFemtiofemFünfundfünfzigFemoghalvtreds

21

What is a Knowledge Representation?

• A KR is a Surrogate

• A KR is a Set of Ontological Commitments

• A KR is a Fragmentary Theory Of Intelligent Reasoning

• A KR is a Medium for Efficient Computation

• A KR is a Medium of Human Expression

(R. Davis, H. Shrobe, and P. Szolovits. What is a Knowledge Representation? AI Magazine, 14(1):17-33, 1993).

22

Knowledge representation in AI

• Logic

• Semantic nets

• Conceptual graphs

• Neural nets

• Object oriented methods

23

Logic

Production system: IF … THEN

IF you have pain in your backAND you have eaten many PanocodeTHEN examine your bile.

24

Pain in your back?

Yes NoEatenPanocode?

25

G: Men are all alike.P: IN WHAT WAY?

G: They’re always bugging us about something or other.P: CAN YOU THINK OF A SPECIFIC EXAMPLE?

G: Well, my boyfriend made me come here.P: YOUR BOYFRIEND MADE YOU COME HERE?

G: He says I’m depressed much of the time.P: I AM SORRY TO HERE THAT YOU ARE DEPRESSED.

G: It’s true. I am unhappy.P: DO YOU THINK COMING HERE WILL HELP YOU NOT TO BE UNHAPPY?

G: I need some help, that much seems certain.P: WHAT WOULD IT MEAN TO YOU IF YOU GOT SOME HELP?

G: Perhaps I could learn to get along with my mother.P: TELL ME MORE ABOUT YOUR FAMILY.

26

(I remember X) ---> (Why do you remember X just now?)

Person: I remember my mother.ELIZA: Why do you remember your mother just now?

(My {family-member} Y me) ---> (Who else in your family Y you?)

Person: My father hates me.ELIZA: Who else in your family hates you?

Rules without left hand side.

ELIZA: Can you elaborate on that for me?, or

ELIZA: That is very interesting. Why do you say that?

27

Logic goes beyond IF … THEN-rules

The author of the Iliad wrote the Odyssey; therefore someone wrote both the Iliad and the Odyssey.

No one admires anyone who admires everyone who admires someone.

28

Semantic nets

29

Mammal BearAnimal

FishWater

Whale

Fur

Vertebra

Cat

isa

isa

lives in

lives in

isa

has

isa

has

has

30

Nixon

Pacifist

Quaker Republican

isa isa

isa isa

31

CG: John is going to Boston by bus.

Conceptual Graph

32

/* schema facts */ man::person. woman::person. person[hasFather=>man].person[hasMother=>woman].

/* facts */ abraham:man. sarah:woman. isaac:man[hasFather->abraham; hasMother->sarah].

/* rules consisting of a rule head and a rule body */ FORALL X,Y X[hasSon->>Y] <- Y:man[hasFather->X].

/* query */ FORALL X,Y <- X:woman[hasSon->>Y[hasHasFather->abraham]].

Object method

33

(John McCarthy, http://www-formal.stanford.edu/jmc/whatisai/)

Q. How far is AI from reaching human-level intelligence? When will it happen? A. A few people think that human-level intelligence can be achieved by writing large numbers of programs of the kind people are now writing and assembling vast knowledge bases of facts in the languages now used for expressing knowledge. However, most AI researchers believe that new fundamental ideas are required, and therefore it cannot be predicted when human level intelligence will be achieved.

34

Why have we made limited progress in AI? Because we haven't developed sophisticated models of thinking, we need better programming languages and architectures, and we haven't focused on common sense problems that every normal child can solve.

(Minsky, 2002)

35

What’s the problem?

Semantics

36

Fifty-fiveFemtiofemFünfundfünfzigFemoghalvtreds

Roman: LV

Binary: 110111

Octal: 101101

Decimal: 55

37

Susan saw the man in the park with a dog.

Susan saw the man in the park with a statue.

Susan saw the man in the park with a telescope

39

END