(TES3111 / TIC3151)

Artificial Intelligence

Lecture 1
Introduction
1

Reference Books
Textbook (Main):
1.

S. Russell and P. Norvig Artificial Intelligence: A Modern

Approach Prentice Hall
2. Etc.

LISP:
1. Paul Graham, The ANSI Common Lisp , Prentice Hall , 1995.
2. Peter Seibel, Practical Common Lisp: Online at
http://www.gigamonkeys.com/book/
3. David S. Touretzky, Common Lisp: A Gentle Introduction to
Symbolic Computation, Benjamin/Cummings, 1990.
4. Winston & Horn, LISP , 3rd edition, Addison-Wesley, Reading, MA,
1989.
5. P. Blackburn, J. Bos and K. Striegnitz, Learn Prolog Now! Free
version of the book is available.
2

Introduction

The field of artificial intelligence (AI) attempts not just

to understand how we think

but also to build intelligent entities

replicate how human

brain should work!

AI currently encompasses a huge variety of subfields

general-purpose areas - learning and perception

specific tasks - playing chess, proving mathematical
theorems, writing poetry, and diagnosing diseases.

AI is relevant to any sphere of human intellectual activity.

It is truly a universal field

Introduction
Example: Mundane tasks
Humans do such tasks easily
Natural language communicating with others in English or
other languages;

Robotics acting & moving in the world;

Vision making sense of what we see;
Planning putting a sequence of actions together to achieve
a goal.

Introduction
Example: Mundane tasks
A human expert required for:
Medical diagnostics
Equipment repair
Computer configuration
Financial planning
Mundane tasks in general are much harder for a machine.

What tasks require AI?

AI is the science and engineering of making intelligent machines
which can perform tasks that require intelligence when performed
by humans
AI is the study of how to make computers do things which, at
the moment, people do better.

What tasks require AI?

game playing
speech recognition
understanding natural language
computer vision

expert systems
6

What is AI?
Systems that think like humans

Systems that think rationally

The exciting new effort to make

computers thinks machine with
minds, in the full and literal sense.
(Haugeland 1985)
[The automation of] activities that we
associate with human thinking, activities
such as decision-making, problem
solving, learning (Bellman, 1978)

The study of mental faculties through the use

of computational models. (Charniak et al.
1985)

Systems that act like humans

Systems that act rationally

The art of creating machines that

perform functions that require
intelligence when performed by people
(Kurzweil, 1990)
The study of how to make computers do
things at which, at the moment, people
are better. (Rich and Knight, 1991)

Computational Intelligence is the study of

design of intelligent agents. (Poole et al.,
1998)

The study of the computation that make it

possible to perceive, reason, and act.
(Winston, 1992)

AI .. Is concerned with intelligent behavior in

artifacts. (Nilsson, 1998)
7

What is AI?
Systems that think like humans

Systems that think rationally

Thinking Humanly

Thinking Rationally

(The Cognitive Modeling)

(The Laws of Thought)

Systems that act like humans

Systems that act rationally

Acting Humanly

Acting Rationally

(The Turing Test)

(The Rational Agent)

Acting Humanly: The Turing Test

"If it looks, walks, and quacks like a duck, then it is a duck"
Alan Turing's 1950 article Computing Machinery and Intelligence
discussed conditions for considering a machine to be intelligent
Can machines think? Can machines behave intelligently?

Acting Humanly: The Turing Test

Turing test (1950):
Guess which of B and C
- is the machine?
- is the person?

C
Some Written questions
Written responses
Machine is intelligent if the interrogator cannot distinguish C from B

10

Acting Humanly: The Turing Test

Computer needs to possess capability of:
natural language processing to communicate (English)
knowledge representation to store what it knows/hears
automated reasoning to use the stored information to
answer questions and draw new conclusions
machine learning to adapt to new circumstances and detect
patterns

11

Acting Humanly: The Turing Test

Problem with Turing Test: it does not permits the interrogator to
observe the physical natures of B and C

Total Turing Test - to test subjects perceptual abilities, requires

computer to have:
computer vision to perceive objects
robotics to manipulate objects and move about

12

Thinking Humanly: Cognitive Modeling

If a program thinks like a human, we need to determine how
human thinks
Two ways:
1.

introspective catch our own thoughts as they go by

2.

psychological experiments once we have sufficiently

precise theory of the mind, can express theory in a computer
program

Cognitive Science brings together computer models from AI and

experimental techniques from psychology to try construct precise
and testable theories of the workings of human mind.
13

Thinking rationally: Laws of thought"

Capture "correct" reasoning processes
A loose definition of rational thinking: Irrefutable reasoning
process
How do we do this?
Develop a formal model of reasoning (formal logic) that
"always" leads to the "right" answer
Implement this model
How do we know when we've got it right?
When we can prove that the results of the programmed
reasoning are correct
Soundness and completeness of first-order logic
14

Thinking rationally: Laws of thought"

Aristotle:
Logical forms that rational thinking possesses.

Example: Socrates is a man, all men are mortal, therefore

Socrates is mortal.
Several Greek schools developed various forms of logic: notation
and rules of derivation for thoughts

15

Thinking rationally: Laws of thought"

Problems:
Not all intelligent behavior is mediated by logical deliberation
(not all facts are certain!)
Logical systems tend to do wrong thing in the presence of
uncertainty
Example: the flight might be delayed

16

Acting Rationally: The Rational Agent

Rational behavior: doing the right thing
An agent is an entity that perceives and acts

Maximize goal,
given available
information

This course is about designing rational agents

agents, build to in one way or another, act rational

Practically: For any given class of environments and tasks, we

seek the agent (or class of agents) with the best performance in a
given environment at a particular time.

17

Acting Rationally: The Rational Agent

Caveat: computational limitations make perfect rationality
unachievable (even if we had perfect it)

Some agent functions might be computationally

intractable
Goal: design best program for given machine resources

Rational agents do the right thing

Take actions that are optimal for achieving goals
Computational limits prohibit complete rationality
Thus, attempt to be as rational as possible given resource
constraints
Textbook: focuses on
acting rationally as
the definition of AI
18

Foundations of AI
Different fields have contributed to AI in the form of ideas,
viewpoints and techniques.
Philosophy: Logic, reasoning, mind as a physical system, foundations of
learning, language and rationality.
Mathematics: Formal representation and proof algorithms, computation,
(un)decidability, (in)tractability, probability.
Economics: formal theory of rational decisions, game theory.
Neuroscience: physical substrate for mental activities (study of the brain).
Psychology: How do humans and animals think and act? (adaptation,
phenomena of perception and motor control).
Computer Engineering: How can be built an efficient computer?
Control theory and Cybernetics: How can artifacts operate under their
own control? ( homeostatic systems, stability, optimal agent design).
Linguistics: How does language relate to thought? (knowledge
representation, grammar).
19

Foundations of AI
Philosophy
o
o
o
o

Can formal rules be used to draw valid conclusions?

How does the mental mind arise from a physical brain?
Where does knowledge come from?
How does knowledge lead to action?

Mathematics
o
o
o

What are the formal rules to draw valid conclusions?

What can be computed?
How do we reason with uncertain information?

Logic
Computation
Probability

Economics
o
o
o

How should we make decisions so as to maximise payoff?

How should we do this when others may not go along?
How should we do this when the payoff may be far in the future?
The mathematical treatment of preferred outcome
20

Foundations of AI

Neuroscience
o How do brain process information?

Psychology
o How do humans and animals think and act?

Computer engineering
o How can we build an efficient computer?

Control theory and Cybernetics

o How can artifacts operate under their own control?

Linguistics
o How does language relate to thought?

21

AI History
Modern founders of AI
McCulloch & Pitts (neural nets)
Russell and Whitehead (a formal analysis of propositional
logic)
Alan Turing (Turing Test)
Norbert Wiener (cybernetics)
John von Neumann (von Neumann computer, game theory)
Claude Shannon (information theory)
Newell & Simon (the Logic Theorist theorem proving)
John McCarthy (LISP, Artificial Intelligence term)
Marvin Minsky (first NN computer, microworlds-limited
domain problems)
and many more
22

A (Short) AI of History

23

AI Achievements and AI Today

AI techniques are used in many common applications:

Intelligent user interfaces

Spell/grammar checkers
Medical diagnosis systems
Regulating/controlling hardware devices and processes (e.g, in
automobiles)
Voice/image recognition (more generally, pattern recognition)
Scheduling systems (airlines, hotels, manufacturing)
Program verification/compiler and programming language
design
Web search engines/Web spiders
Web personalization
Recommender systems (collaborative/content filtering)
Credit card verification in e-commerce /fraud detection
Data mining and knowledge discovery in databases
Computer games
24

AI Achievements and AI Today

Many technologies widely used today were the direct or
indirect results of research in AI:
Robotics
Speech recognition
Character recognition
Computer games (e.g., Deep blue chess game)
Symbolic mathematical systems (e.g., Mathematica, Maple,
etc.)
Chip design and human interfaces
Web search

etc.
25

AI: Branches
Symbolic Computation (Logical AI & Search)
Using programs like LISP and PROLOG.
Emphasis in
this course

Natural language processing

Pattern recognition - is the computers vision not unlike the way

your eyes view patterns and recognize objects. It is a computers
way of figuring out what an object is by use of a pattern
Planning - to generate a strategy for achieving some goal
Learning from experience

26

AI State of the art

Have the following been achieved by AI?
World-class chess playing

Playing table tennis

Cross-country driving
Solving mathematical problems
Discover and prove mathematical theories
Engage in a meaningful conversation
Understand spoken language
Observe and understand human emotions
Express emotions

27

Primary AI languages
Prolog
First Prolog program: France, 1970
Based on theorem proving research
Major development at University of Edinburgh, 1975-79
Adopted by the Japanese Fifth Generation Computing Project

Logic programming language:

Programs composed of facts and rules
Executes by applying first-order predicate calculus/unification to
programs
Interactive interpreter, compiler
Tell the computer what is true and what needs to be done, rather
than how to do it.
28

Primary AI languages
Prolog Example
likes(deb,horses).

likes(deb,dogs).
?- likes(deb,horses).
yes

likes(deb,X).
X=horses
X=dogs
likes(deb,Y) :- horse(Y).
horse(robin).
?- likes(deb,robin).
yes
29

Primary AI languages
LISP
Proposed by McCarthy, late 1950s; contemporary of COBOL,
FORTRAN

Functional programming language based on lambda

calculus/recursive function theory
Intended as a language for symbolic rather than numeric
computation
Interactive interpreter, compiler
Uses atoms, lists, functions.

30

Primary AI languages
LISP Example
(defun hypotenuse (x y)
(sqrt (+ (square x) (square y))))
> (hypotenuse 4 3)
5

31

Weak and Strong AI Claims

Contrast between dualism (brain + mind) versus
materialism /mechanistic (Brains cause minds)
Weak AI: Machines can be made to act as if they were
intelligent (interested in getting programs to do particular tasks)
Turing Test (Turing, 1950)
Strong AI: Machines that act intelligently have real, conscious
minds (wants to build a model of the mind (or at least)
intelligence)
Consciousness:
Self-awareness
Subjective experiences
32

The Chinese Room Experiment

Searles Chinese Room
Thought experiment, 1980.
Refined consciousness objection.
There is a room, with a man who only speaks English.

Man has book, with instructions: given some scribble in Chinese,

output this scribble.
A man fluent in Chinese sends messages (in Chinese) into room,
and gets responses (also in Chinese).
He cant distinguish between man in room and fluent Chinese
speaker.
Does this mean the room knows Chinese??
33

The Chinese Room Experiment

34

The Chinese Room Experiment

Computers used syntactic rules to manipulate

Chinese words but have no understanding of
the meaning (semantics)

35

This experiment pass the Turing

test but cannot be considered
as to have
conscious (intelligence!)
- Turing test only tests for
weak AI, not strong

AI Programming Languages
Lisp
1. Allegro CL for Windows, Linux, and FreeBSD: a free demo
version from Franz Inc.
2. CLISP: a portable ANSI Common Lisp System (download).
3. CMUCL Home Page: the new official home of CMU
Common Lisp. (This implementation can be run on the
EMCF machines using the script at /usr/local/bin/cmulisp.
4. Freelisp for Windows or Linux, free demo versions of
LispWorks, and Lucid Common Lisp.

36

AI Programming Languages
PROLOG
1. Ivan Bratko, Prolog Programming for Artificial Intelligence,
third edition, Addison Wesley.
2. Leon Sterling and Ehud Shapiro, The art of Prolog:
Advanced programming techniques, MIT Press 1986
3. SWI Prolog Users Manual available at
http://www.swi-prolog.org/dl-doc.html
SWI Prolog web site http://www.swi-prolog.org/

37

The End

38