Artificial Intelligence

COMP3702/COMP7702
Course Coordinator
Janet Wiles

Lecturers
Shoaib Sehgal
Ruth Schulz

Tutor
Suren Rathnayake
Aims and Objectives

¾ View: Methods and techniques within the field of

artificial intelligence and solve theoretical and practical
problems.

¾ Scope: The course provides an understanding of AI and

describes many of the most important algorithms and
techniques which are theoretcally pratically important.

¾ Purpose: The course helps the student to:

‰ gain an appreciation for the scientific context of
artificial intelligence
‰ understand and develop computing algorithms, and to
analyse their properties
‰ find the right techniques for solving specific
problems, and to implement these techniques
Aims and Objectives (cont)

¾ In general terms, it is expected that the student gains an

understanding of the theories, methods and practices which
form the basis of Artificial Intelligence.
¾ The course aims to introduce the basic concepts and
methods used in the field of artificial intelligence and
provide students with skills in the use of applying these
techniques.
¾ Specifically the course aims to give students an overview of
the following topics in artificial intelligence:
‰ Problem solving and optimisation (search algorithms)
‰ Reasoning with uncertain knowledge (probability
theory)
‰ Machine Learning (classification, etc)
‰ Probabilistic approaches for information retrieval
Aims and Objectives (cont)
After the course, you should

¾ be familiar with the historical context of artificial

intelligence
¾ know several definitions of artificial intelligence
¾ be familiar with an agent-based intelligent system design
¾ understand several problem solving and optimisation
techniques based on search (both uninformed and
informed)
¾ be able to implement and apply search techniques
¾ understand general principles of machine learning (both
supervised and unsupervised)
Aims and Objectives (cont)

In addition, you should

¾ know several machine learning techniques

(including decision tree learning and neural
networks)
¾ be able to implement, apply and systematically
evaluate machine learning techniques
¾ understand probability theory and how it can be
used for representing and reasoning with uncertain
knowledge
¾ be able to apply basic probability theory to
machine learning problems
¾ understand some probabilistic approaches to text
mining and information retrieval
Resources

¾ Highly recommended reading

‰ Russell S. and Norvig P., Artificial Intelligence:
A modern approach, 2nd edition, 2003.
‰ Used extensively (see reading list)
‰ Available for purchase from the University
Bookshop
‰ Several copies are also available at the library
Resources (cont)

¾ Handouts
‰ At www.itee.uq.edu.au/~comp3702 you will find
‰ Slides used in lectures
‰ Tutorials
‰ Assignments
‰ Readings and handouts
‰ Links and resources
Times and Venues

¾ Sign-up for Tutorials via SI-net

 Monday’s Lecture Reading Tutorial
Week Number Lecture Assessment
Date Number Tues 2-4pm
Russell and Session
Norvig, 2003 TBA

Part 1 Shoaib Sehgal

1 Chapters
Introduction to artificial 1, 2 and
1 21 July intelligence, an agent-based 26: pp. No tutorial
2 perspective (6pp, 2pp) 947-949,
958-960.
3 The definition
Solving problems by Assignment 1
2 28 July Chapter 3 of artificial
4 searching (6pp, 2pp) available
intelligence
5 Chapter 4 Problem
Informed search and (except Representatio
3 4 August
6 exploration (6pp, 2pp) 4.4 and n (Comments,
4.5) hints)
7 Informed
Adversarial search, game Search
4 11 August Chapter 6
8 playing (6pp, 2pp) (Comments,
hints)
9 Adversarial
Applications of AI
search
5 18 August i- Bioinformatics(Shoaib)
10 (Comments,
ii- TBA - External Speaker
hints)
Mid-semester exam; Mid-semester
Assignment 1
6 25 August 11 Discussion of assignment 1, exam
preparation
related theory (6pp, 2pp) (optional)
Part 2 Ruth Schulz

13
Assignment 1
Chapter 13
Probabilistic reasoning deadline
7 1 Sep. + 7.1-
14 (6pp, 2pp, example) (Monday 1st
7.2
Sep., 5pm)

15 Principles of machine
Chapters
learning, decision
18 Probabilistic Assignment 2
8 8 Sep. trees (6pp, 2pp,
16 and reasoning (Comments) available
examples with
19
decision tree learning)
17 Chapter 20
Statistical machine Machine learning
9 15 Sep. (20.1-
18 learning (6pp, 2pp) basics (Comments)
20.2)
19 Neural networks (6pp, 2pp Chapter 20
Current best learning and
(corrected 31/10/07), (20.5-
10 22 Sep. decision trees
20 more material on NNs, onwar
(Comments)
NetTalk audio (8MB)) ds)

Mid-semester break (one week)

29 Sep.

21 Learning algorithms for

Chapter 20 Decision Trees and Naïve
6 neural networks (6pp,
(20.5- Bayes Classification
11 Oct 2pp, (corrected
22 onwar (Comments, decision tree
ober 30/10/07) more on NN
ds) solution)
learning)
Robotics and discussion of
Neural networks
13 assignment 2 (6pp,
Chapter (Comments)
12 Oct 23 2pp,(corrected
23 Assignment 2
ober 30/10/07) more on
preparation
applications of NN)
 Revision Period
27 October

Exam Week 1 3 November

Final Exam

Exam Week 2 10 November

Assessment

¾ Assignments (30%)
‰ 2 assignments
‰ Assignment 1 (10%): Search – game playing
‰ Assignment 2 (20%): Machine learning – pattern recognition
¾ Tutorials (10%)
‰ Active participation mark (1% per tutorial)
¾ Final Examination (60%)
‰ During final examination period

‰ Covers lecture material

‰ Details
‰ 2 hours
‰ Closed-book
‰ Primarily short answer / short essay
Important Assessment Information

¾ All assessment is due at 5pm of the due date

¾ Assignments need to be submitted on-line at

http://submit.itee.uq.edu.au

¾ Late submission not accepted except for medical or

strong personal reasons (documentation required)
¾ The programming language will be Java
¾ Tutorials for C/C++ programmers are available at
the course website.
¾ Why Java ?
Why Java http://shootout.alioth.debian.org/

Intel AMD
 Assignments
¾ Two assignments – Two problems/applications which
require intelligence (artificial or natural)

‰ Problem-solving/optimisation
‰ Approach: clever search algorithms

‰ optimising outcomes on basis of a well-defined ‘current state’

‰ exposes computational complexity issues

‰ Pattern recognition
‰ Approach: learning-by-example/machine learning

‰ exposes difficulties of ‘representation by rules’

‰ illustrates the use of probabilistic methods and neural networks

Tutorials

¾ Aims
‰ to provide examples of potential examination
questions
‰ to enable and encourage peer-tutoring
‰ to provide an opportunity for questions
‰ to explore the theoretical concepts
‰ to apply the theoretical concepts
Tutorials
¾ You’ll receive 1% for participating in a tutorial
¾ Students will work in teams during tutorials
¾ Each week one person in the team will submit one tutorial
solution
¾ Each student will submit approximately three tutorial solutions
during the semester
¾ Participation marks will be allocated for active participation
‰ Working through problems

‰ Participating in discussions

‰ Answering questions

‰ Submitting tutorial solutions

¾ Feedback
‰ Tutors will provide some feedback on submitted work

‰ Work will be returned the week after submission

Lectures

¾ Aims
‰ to outline theories, methods and applications of
the field of AI
‰ to explain difficult concepts from the
recommended text and other sources
‰ to illustrate concepts in AI with diagrams and
examples
‰ to provide a forum for general discussion and
questions about the subject matter
Some tips…

¾ Don’t be shy, participate in lectures, ask

questions.
¾ Buy the book, read chapters as noted. It is
well-written, up-to-date, and an excellent
reference for later.
¾ AI is actually quite fun and useful, but you
need to work hard. Assignments and tutorials
will help you to work.
 Examination

¾ Closed book, non-programmable calculator allowed

¾ Knowledge questions - theory
‰ may have been in tutorials
‰ explicit in the recommended text book
‰ assessed on correctness of answer

¾ Knowledge questions - practical

‰ similar to those in tutorials
‰ method described in the recommended text and lectures
‰ assessed on correctness of method application

¾ Discussion questions
‰ may have been addressed in lectures
‰ not necessarily explicit in the readings
‰ assessed on insight / justifications
Artificial intelligence:
An introduction

¾ after which you should

‰ know a couple of definitions if intelligence and AI
‰ know how the Turing Test works
‰ know the Physical Symbol System Hypothesis
‰ understand the point made in Searle’s Chinese Room
Argument
‰ have some familiarity with the history of AI
‰ understand the concept of agents in the context of AI
‰ be familiar with some intelligent agent designs
¾ and feel prepared to discuss actual ways of making
machines clever and making them learn by
themselves…
What is Intelligence?

For each of the following, give three reasons

why:
(a) A dog is more intelligent than a worm.
(b) A human is more intelligent than a dog.
(c) An organisation is more intelligent than an
individual.
What is Intelligence?
(adapted from J. McCarthy – Stanford University)

¾ What is intelligence?

¾ One Answer
‰ Intelligence is the computational part of the ability to
achieve goals in the world. Varying kinds and degrees
of intelligence occur in people, many animals and
some machines.

¾ Do you agree with this definition?

What is AI?
(adapted from J. McCarthy – Stanford University)

¾ What is artificial intelligence?

¾ One possible answer

‰ 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
‰ AI does not have to confine itself to methods that are
biologically observable.
What is AI?
(adapted from J. McCarthy – Stanford University)

¾ Is there a solid definition of intelligence that doesn't

depend on relating it to human intelligence?

¾ Answer.
‰ Not yet. The problem is that we cannot yet
characterise in general what kinds of computational
procedures we want to call intelligent. We understand
some of the mechanisms of intelligence and not
others.
 Definitions of Artificial Intelligence
Systems that think like humans
“The exciting new effort to make
computers think … machines 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]
Systems that act like humans
“The act 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]
Systems that think rationally
“The study of mental faculties through the
use of computational models” [Charniak &
McDermott: 1985]
“The study of the computations that make it
possible to perceive, reason, and act”
[Winston: 1992]
Systems that act rationally
“Computational Intelligence is the study of
the design of intelligent agents” [Poole et al:
1998]
“AI … is concerned with intelligent
behaviour in artefacts” [Nilsson: 1998]
Science of AI
 The Turing Test
¾ Turing proposed a definition of Artificial
Intelligence independent of the inner
mechanisms of the machine

¾ http://www.abelard.org/turpap/turpap.htm

¾ Operational test of intelligent behaviour

‰ Given
‰ an interrogator, and

‰ (A) a computer and (B) a person in a

different room

The interrogator, through asking questions,

must determine whether respondent A or
respondent B is the computer. If the
computer “fools” the interrogator then we
can call it intelligent.
 The Turing Test http://chatbotgame.com/
Chatbot: Hello there human.
¾ What is You: are you bot
Chatbot: Of course. I'm a collection of software rules.
needed? You: how are you
Chatbot: Life is tolerable. How are things in your world?
You: excellent
‰ Knowledge Chatbot: Can we change the subject please? Do you like
swimming?
‰ Reasoning You: nop
Chatbot: "nop" ... What does that mean? It's not in my
‰ Language dictionary.
‰ Learning You: nop = no
Chatbot: That's too bad.
You: why
Chatbot: Because of an intense metaphysical connection in
the properties of submolecular particles
You: what?Chatbot: Do you understand me?
You: no
Chatbot: Well I don't understand humans.. Lol
Attributes of the Turing Test

¾ Positive
‰ “standard” test
‰ bypasses the true nature of intelligence
‰ removes bias
¾ Negative
‰ focus on symbolic tasks
‰ compares machine with ‘human’ intelligence
‰ far too restrictive
‰ difficult to do in practice
Let’s play: “Spot the Intelligence!”

¾ Order the following from most to least

intelligent
‰ Cockroach
‰ Computer
‰ Elevator
‰ Hammer
‰ Human
‰ Model-T ford motor car
‰ Sunflower
Answers?

¾ Simple yes / no answers

cannot be made from My Ranking
‰ Human
current definitions. ‰ Cockroach
¾ Your answer reflects ‰ Sunflower
‰ Computer
your beliefs, and is not
‰ Elevator
necessarily correct or ‰ Model-T Ford
incorrect. Motor Car
‰ Hammer
Examples
 Average
Co
ck

0
1
2
3
4
5
6
7
8

ro
a
Co c h
m
pu
te
El r
ev
at
or
Ha
m
m
er
Hu
m
an
Average Ranking

M
od
e
S u l- T
nf
lo
Average Rank (from 2000)

we
r
Results of Intelligence Ranking

Most Popular Response

8
7
6
5
Rank

4
3
2
1
0

-T

er
or
h

r
an

we

te
ac

el

m
at
m

pu
lo

ev

od
ro

m
Hu

m
nf
ck

Ha
El

M
Su

Co
Co
Physical Symbol System Hypothesis
(Simon and Newell, 1976)

A physical symbol system has the necessary and sufficient means for intelligent
action.

¾ A system:
‰ Consists of a set of entities, called symbols,

‰ Symbols can occur as components of another type of

entity called an expression (or symbol structure)
‰ A symbol structure is composed of a number of instances
(or tokens) of symbols related in some physical way (such
as one token being next to another).
‰ Contains a collection of these symbol structures

‰ Contains a collection of processes that operate on expressions

to produce other expressions: processes of creation,
modification, reproduction and destruction

¾ Set of finite symbols that can be composed to form a potential

infinite set of expressions
Physical Symbol System Hypothesis
(Simon and Newell, 1976)

“A physical symbol system has the necessary and sufficient means for
intelligent action.”

¾ Intelligent human thought: the symbols are encoded

in our brains. The expressions are thoughts. The
processes are the mental operations of thinking.

¾ A running artificial intelligence program: The

symbols are data. The expressions are more data. The
processes are programs that manipulate the data.
 Searle’s Chinese Room Thought
Experiment
¾ You are:
‰ Monolingual English speaker locked in a room

¾ You are given

‰ a large batch of Chinese writing
‰ a second batch of Chinese script as an output
‰ a set of rules in English for correlating the second batch with the
first batch
¾ The rules correlate one set of formal symbols with another set of
formal symbols
¾ Your responses are indistinguishable from those of Chinese
speakers

Just by looking at your answers, nobody can tell you "don't

speak a word of Chinese."
Searle’s Chinese Room

¾ The Point: No matter

‰ how intelligent a computer seems to behaves and

‰ no matter what programming makes it behave that
way
‰ since the symbols it processes are meaningless
(lack semantics) to it
‰ it’s not really intelligent
Searle’s Chinese Room

¾ Demonstrates that a system can be merely following

rules, but not understanding anything at all.

¾ The system he describes will pass the Turing Test

¾ Behaviour cannot determine extent of understanding

¾ Is following rules sufficient for intelligence?

Strong AI vs Weak AI

¾ Strong AI
‰ duplication of intelligence
‰ aims to understand intelligence

¾ Weak AI
‰ simulation of intelligence
‰ aims to make computers more ‘useful’
History of AI - Early 20th Century

¾ 1943
‰ McCulloch and Pitts lay the foundations for neural
networks.
‰ The term “cybernetics” is coined
¾ 1950
‰ Turing publishes “Computing Machinery and Intelligence”.
‰ Introduction of the Turing Test as a test for intelligent
behaviour.
‰ Claude Shannon publishes a detailed analysis of chess
playing as search.
Modern History

¾ 1956
‰ John McCarthy coined the term “artificial intelligence” as
the topic of the Dartmouth Conference.
‰ Demonstration of AI program - the Logic Theorist (LT) -
written by Newell, Shaw and Simon (CMU).
¾ 1952-62
‰ Samuel (IBM) wrote the first game-playing program that
learns.
¾ 1962
‰ First industrial robot company, Unimation, is founded.
Modern History

¾ 1965
‰ Weizenbaum (MIT) built ELIZA, an interactive program
that carries on a dialogue in English on any topic. It
became a popular toy at AI centres.
‰ See http://www-ai.ijs.si/eliza-cgi-bin/eliza_script
‰ Or http://www.manifestation.com/neurotoys/eliza.php3
¾ 1967
‰ Dendral program (Feigenbaum, Lederberg, Buchanan and
Sutherland) is the first successful knowledge-based
program for scientific reasoning.
¾ 1968
‰ Minsky and Papert publish “Perceptrons”, demonstrating
the limits of simple neural nets.
Modern History

¾ 1969
‰ SRI robot, Shakey, demonstrates combining
locomotion, perception and problem solving.
¾ 1974
‰ The first expert system - MYCIN (Stanford) -
demonstrates the power of rule-based systems for
knowledge representation and inference in the
domain of medical diagnosis and therapy.
 Modern History

¾ 1969-1979
‰ Knowledge-based (expert) systems
¾ 1980-1988
‰ Expert systems industry booms
¾ 1988-1993
‰ Expert systems industry busts – “AI winter”
1966 US Military report on machine translation: “The spirit is willing but the
flesh is weak” was re-translated (English-Russian-English) into
“the vodka is good but the meat is rotten”. Prospects seemed poor and
funding was cut.

2008 version (Babel fish): “Spirit is willingly ready but flesh it is weak”
Modern History
¾ Mid-1980’s-present
‰ The return of Neural Networks

¾ 1985
‰ Brooks (MIT) develops the concept of behaviour based robots

¾ 1988-present
‰ Resurgence of probability: general increase in technical depth
‰ Computational Intelligence / soft computing (Evolutionary
computing, Swarm Intelligence, Fuzzy systems, NNs)

¾ 1995-present
‰ Agents, agents everywhere …
Agents

¾ Generally computers are obedient, literal,

unimaginative servants
‰ acceptable for most applications
¾ However, increasingly we require systems
that can decide for themselves
¾ Such computer systems are known as agents
Definition of the term ‘Agent’

An agent is a computer system that is situated

in some environment, and that is capable of
autonomous action in this environment in
order to meet its design objectives.
Wooldridge and Jennings (1995)

Autonomy: act without the intervention of humans or

other systems, control both over their own internal
state and over their behaviour
I, Robot
 Agents and Environments
Sensors
Percepts

Environment ? Agent

Actions
The agent takes Effectors
sensory input from the
An agent consists of
environment and Rationality depends on
produces as outputs •an architecture, and •performance, degree of success,
actions that affect it. •a program. •perceptual history,
Agents include humans, •knowledge of environment, and
•actions available for deployment
robots, softbots,
thermostats, etc Let’s look at these aspects in terms
of different agent designs…
 Agent Types

¾ Simple reflex agents

¾ Model-based reflex agents
¾ Goal-based agents
¾ Utility-based agents

¾ Example: Automated taxi driver

Performance
Agent Type Environment Actuators Sensors
Measure
Roads, other Steer, Cameras,
Safe, fast, legal,
traffic, accelerate, speedometer,
Taxi Driver comfortable trip,
pedestrians, brake, talk to GPS, sonar,
maximise profits
customers passenger microphone
Simple reflex agents

Agent Sensors

What the world

Environment
is like now

What action I
Condition-action rules
should do now

Actuators
Model-based reflex agents

Agent Sensors

State
What the world is

Environment
How the world evolves like now

What my actions do

What action I
Condition-action rules
Should do now

Actuators
Goal-based agents

Agent Sensors

State
What the world is

Environment
How the world evolves like now

What my actions do What it will be like

if I do action A

What action I
Goals
should do now

Actuators
Utility-based agents

Agent Sensors

State What the world is

like now

Environment
How the world evolves
What it will be like
What my actions do if I do action A

How happy I will be

Utility in such a state

What action I
should do now

Actuators
Questions ?

Office Hours
Tuesday 10-12 AM
Room 308, Axon building ITEE

Acknowledgements:
http://domus.usherbrooke.ca
www.ironbot.com
for some of the images used in making of these slides.