CSC311 - Artificial Intelligence

CHAPTER 1: Introduction

Lecturer: Prof Louise Leenen

e-mail: csc311@[Link]

Slides for this course created by L. Leenen and the following sources:
• Provided by Pearson Publishers
• Penn State University
• EECS at Berkeley University ([Link]
• Duke University ([Link]
• Princeton University ([Link]
• Bryn Maer College, D. Kumar ([Link]

1
 Basic information
Prescribed textbook: “Artificial Intelligence: A Modern
Approach” by Russel & Norvik (R&N)
(The newer edition is also fine)

2
 Take note….
Contents
• Prescribed book - read it

• Do a lot of exercises!

• Make an appointment if you need help.

• No extension on any assessment.

3
 What will we cover in this course?
Contents
• Introduction to AI

• Intelligent Agents

• Solve problems by searching for solutions

• Basic Search – Uninformed & Informed
• Local Search
• Adversarial Search (Game Playing)

• Introduction to logic-based AI

4
 Read Chapter 1 of R&N first…
Contents
• What is AI?

• What is not AI?

• Foundational Disciplines

• History of AI

• Applications

5
 What is artificial intelligence?
• John McCarthy coined the term in 1956 and
described AI as “the science and engineering of
making intelligent machines”

• Luger & Stubblefield, 1993 “...the automation of

intelligent behaviour”

• Kurzweil, 1990 “...creating machines that perform

functions that require intelligence when performed by
people”

6
 What is artificial intelligence?
• The engineering of artifacts that act intelligently
(rational behaviour in artifacts)

• The science of artifacts that reason intelligently

(rational reasoning by computational models)

• An intelligent agent can improve its performance by

learning from observations

• An intelligent agent can derive (search for)

conclusions that are not explicitly represented

7
 What is intelligence?
• How to recognise it: Linguistic, musical, logical,
spatial, kinesthetic, etc.

• How to replicate it: Reasoning, learning, problem

solving, perception, etc.

8
What is intelligence?

9
How to test for “intelligence”?

“Chinese room” argument

[Searle 1980]
image from [Link]

• A person who knows English but not Chinese sits in room.

• He receives notes in Chinese.
• He has a systematic English rule book for how to write new Chinese
characters based on input Chinese characters, and returns his notes.
– Person=CPU, rule book=AI program, really also need lots of paper
(storage)
– He has no understanding of what they mean
– ..but from the outside, the room gives perfectly reasonable
answers in Chinese!
• Searle’s argument: the room has no intelligence in it! 10
 The Turing Test
• (Human) judge communicates with a human and a machine
over text-only channel.
• Both the human and the machine try to act like a human
• Judge tries to tell which is which.
• Current programs struggling with this.
• Supplementary reading: Read Alan Turing’s 1950 paper (
on iKamva).

image from [Link] 11

 Turing Test on unsuspecting judges
• It is possible to (temporarily?) fool humans who do not
realize they may be talking to a bot.

• The well-known ELIZA program [Weizenbaum 66] simply

rephrases a user’s statements and questions.

• Recently developed chatbots are much more advanced.

12
 But is the Turing Test the right goal?
• “Aeronautical engineering texts do not define the goal of
their field as making ‘machines that fly so exactly like
pigeons that they can fool even other pigeons.’” [Russell
and Norvig]

13
 Different approaches to AI…
Systems that think Systems that think
like humans rationally
Systems that act Systems that act
like humans rationally

• We will follow an “act rationally” approach:

• acting rationally/like a human presumably requires (some
sort of) thinking rationally/like a human,
• humans much more rational anyway in complex domains.

14
• General-purpose AI like the robots of science fiction is
incredibly hard
– Human brain appears to have lots of special and general
functions, integrated in some amazing way that we really
do not understand at all (yet).

• Special-purpose AI is more doable (nontrivial)

– E.g., chess/poker playing programs, logistics planning,
automated translation, voice recognition, web search,
data mining, medical diagnosis, keeping a car on the
road, … … … …

15
 Another definition2)

The field of Artificial Intelligence (AI) involves the

design and implementation of systems
that exhibit capabilities of the human mind, such
as reasoning, knowledge, perception,
planning, learning, and communication.

16
 What is not AI?

• Specialised systems that can solve specific pre-

specified problems
• Artifacts that are unable to reason, adapt or learn
• Look-up systems that retrieve explicit data
• Anything that lives
• Collections of living things, such as a swarm of bees
• Collective intelligent abstract constructs such as
organisations

17
 Lessons from AI research
• Clearly-defined tasks that we think require intelligence &
education from humans tend to be doable for AI
techniques
– Playing chess, drawing logical inferences from clearly-stated facts,
performing probability calculations in well-defined environments,
…
– Although, scalability can be a significant issue.
• Complex, messy, ambiguous tasks that come natural to
humans (in some cases other animals) are much harder for
AI
– Recognizing your grandmother in a crowd, drawing the right
conclusion from an ungrammatical or ambiguous sentence,
driving around the city, …
18
 Lessons from AI research

• Humans are better at coming up with reasonably good

solutions in complex environments.

• Humans are better at adapting/self-evaluation/creativity

(“My usual strategy for chess is getting me into trouble against this
person… Why? What else can I do?”)

19
 Foundational disciplines

• Philosophy, psychology, linguistics, information

science

• Mathematics, statistics, biology, neuroscience

• Engineering, computer science

20
 Applications
• Robotics
• Decision theory
• Gaming
• Knowledge engineering
• Natural language processing
• Neural networks
• Constraint satisfaction
• Data science
• Baysian networks
• Semantic Technologies
21
22
 Early history of AI
• 50s & 60s: Early successes! AI can draw logical conclusions,
prove some theorems, create simple plans… Some initial
work on neural networks…
• Led to overhyping: researchers promised spectacular
progress, but started running into difficulties:
– Ambiguity: highly funded translation programs (Russian to
English) were good at syntactic manipulation but bad at
disambiguation
• “The spirit is willing but the flesh is weak” becomes “The
vodka is good but the meat is rotten”.
– Scalability/complexity: early examples were very small, programs
could not scale to bigger instances.
– Limitations of representations used.
23
 History of AI…
• 70s & 80s: Creation of expert systems (systems specialized
for one particular task based on experts’ knowledge), wide
industry adoption.

• Again, overpromising…

• … led to AI winter(s)
– Funding cutbacks, bad reputation.

• But now we may now be in the AI spring.

24
 Modern AI
• More rigorous, scientific, formal/mathematical.
• Fewer grandiose promises.
• Divided into many subareas interested in particular
aspects.
• Some senior AI researchers are calling for re-integration
of all these topics, return to more grandiose goals of AI.
• Artificial General Intelligence is receiving attention. AGI
looks for a universal algorithm for learning & acting in any
environment.
• Very large data sets may also change the focus of AI.

25
 This course

• Intelligence (for our purposes) is concerned with

rational action.

• Focus on general AI techniques that have been

useful in many applications.

• Will try to avoid application-specific techniques

although they are still interesting and worthwhile.

26