Transcript Lecture - Computer Science

Introduction to AI
&
AI Principles (Semester 1)
WEEK 2 – Tuesday part A
(2008/09)
John Barnden
Professor of Artificial Intelligence
School of Computer Science
University of Birmingham, UK
Stupidity and Intelligence
So, we’re not intelligent creatures?
Could an ant/rat/dog/monkey be stupid in those ways?
Did those mistakes arise out of good reasons, actually?
The necessity of stupidity. Stupidity in the context of
intelligence.
Emotion and thought. (Huge but hot topic.)
We can intelligently reflect on our stupidity!
As well as being stupid about our intelligence!!
INTRODUCTION
to AI, at last!
What is the Field of
“Artificial Intelligence”
????
Not an easy question, for many reasons, incl.:


What is “intelligence”???? [often asked]
What does “artificial” mean? [not so often asked, but implicit in
discussions of whether AI can merely simulate intelligence or can
achieve it]
What is “AI” ???? contd
 Basically, AI (the field) is the study of
how to create artificial entities that
have features that we associate with intelligence,
and cognition more generally,
in humans and other living things,
such as reasoning, planning, communicating in language, solving
problems, seeing what’s around, moving around in the world,
creating artworks, playing games, learning, emoting, being
conscious, having society, etc.
Cautions 1
The “artificial objects” are currently computers, robots,
computer programs, etc., as we know them today, but
could include radically different types of artefact in future
…
And given the possibility of synthetic biology, we might
eventually grow our artefacts … and how different would
that be from a woman just giving birth?
Cautions 2
And who is it that “associates” those features with
intelligence? …
In fact AI covers a lot of things most people would not
have thought of as “intelligent”
– e.g. seeing that there’s a pencil on a desk. (Hence my
inclusion of “cognition in general”.)
Cautions 3
Many AI researchers don’t themselves produce or directly
study programs, computers, robots, etc., but instead
create:
underlying computational principles or frameworks,
mathematical theory, etc.
Why Do We Do AI?
An “Engineering” Aim
A “Psychological” Aim
A “General/Philosophical” Aim
“Engineering” Aim
To engineer, or provide computational principles and
methods for engineering,
useful artefacts that are arguably intelligent (in the broad
sense above), without necessarily having any mechanistic
similarity to human or animal minds/brains.
The usefulness may be in an industrial domain or an
everyday, practical domain, but may also be in other
domains such as art or mathematical theorem proving.
“Psychological” Aim
To devise computational principles, computationallydetailed theories, or running computational systems that
provide a basis for possible testable accounts of cognition
in human or animal minds/brains.
In short, contributing computationally to questions such
as, how does the human mind work?
But why not leave that to Psychologists? Answer: they
don’t know as much about computation. Also, thinking
“outside the biological box” is liberating.
“General/Philosophical” Aim
To devise computational principles, computationallydetailed theories, or running computational systems that
serve as, or suggest, possible accounts of cognition in
general, whether it be in human-made artefacts, in
naturally-occurring organisms, or in beings yet to be
discovered, or that
illuminate philosophical issues such as the nature of
mind, thought, intelligence, consciousness, perception,
language, representation, learning, creativity, rationality,
society, etc.
Mixing of Aims
 The three aims are often inextricably combined in a given
piece of AI work.

An individual researcher may subscribe to more than one of the
aims.

Developments in pursuit of any one of the aims could happen to
inspire advances towards one of the others.

Endeavours that have any one of the aims can deliberately look
for inspiration from research that has one of the other aims.
Why Am I (Puzzled Student) Doing AI??
i.e., “Eh? I?”
 It is a particularly fascinating, fun, liberating, inspiring,
challenging, breathtaking, … in short, sexy aspect of CS.
 AI technology (software, hardware) is creeping more and more into
practical applications.
 It’s a relatively people-orientated side of CS, and interacts/overlaps
with many other disciplines such as Psychology.
 It’s a good basis for interesting undergrad/master’s/doctoral
projects.
 It involves many general CS issues, and there are no clear
boundaries anyway.
Some Actual or Emerging Applications
Learning, planning & communication in computer games.
Diagnosis, in many areas including medical.
Intelligent conversational agents (incl. chatbots, helpful
avatars on company sites, ICAs fronting travel services).
Emotive ICAs.
Military apps, incl. battle planning, target identification.
Aircraft/spacecraft/planetcraft control & action planning.
[See Callan book]
Satnavs.
Some Example Applications, contd. 1
Stock market prediction.
Fraud-detection: credit cards, phone usage.
Data mining for marketing purposes.
Text summarization and information-extraction.
Artistic creativity (music, paintings, …).
Building design.
Intelligent transport systems, utility networks, etc.
Some Example Applications, contd. 2
Intelligent personalized web-search agents.
Policing and national security.
Machine translation (of language).
Semantic web.
“Digital Economy”
Major new national research-funding priority,
led by the EPSRC (Engineering and Physical Sciences Research
Council).
Includes concern for issues like how large computerbased systems can interact with people in a good way,
avoid social exclusion, etc.
Naturally brings in many AI and other CS concerns.
To be led by the needs of industry, community, society.