Transcript COMP 3009 Introduction to AI

COMP 3009
Introduction to AI
Dr Eleni Mangina
[email protected]
http://www.cs.ucd.ie/staff/emangina
TEXTBOOK
“Artificial Intelligence”
Structures and Strategies for
Complex Problem Solving
George F. Lugger & William A.
Stubblefield
ARTIFICIAL INTELLIGENCE
“Artificial Intelligence (AI) may be defined as a
branch of computer science that is concerned with
the automation of intelligent behavior”
WHAT IS INTELLIGENCE?
- Is it a single characteristic? OR…
- Is it a collection of distinct and related abilities?
- Is it learned OR… is it a priori existence?
AI – Think about it!
INTUITION
LEARNING
CREATIVITY
KNOWLEDGE
REPRESENTATION
SELFAWARENESS
INFERENCE
FROM
OBSERVABLE
BEHAVIOR
EVIDENCE OF A
PARTICULAR
INTERNAL
MECHANISM?
QUESTIONS:
• Is it necessary to pattern an intelligent computer program after what is
known about human intelligence OR is a strict “engineering” approach
to the problem sufficient?
• Is it even possible to achieve intelligence on a computer OR does an
intelligent entity require the richness of sensation and experience that
might be found only in a biological existence?
DID YOU GET CONFUSED?
DO NOT WORRY!
AI offers a unique and
powerful tool for exploring
exactly these questions!!
“Artificial Intelligence is the science that investigates
human intelligence by trying to build it in
artificial”
http://library.thinkquest.org/2705/
Let’s talk about it!!
HISTORY
• Aristotle: “most fascinating aspect of nature is: change”,
philosophy of nature = study of things that change 
Philosophical basis for: Symbolic computing and data
abstraction
LOGIC = the “instrument”(organon)
Study of thought itself at the basis of all knowledge!
Propositions  true OR false
P: “all men are mortal”
Q: “Socrates is a man”
“Socrates is mortal”
Deductive form  modus ponens
HISTORY (cont.)
• Renaissance: Thinking about humanity and its
relations to the natural world.
Empiricism  understanding nature
Scientists and philosophers realized that thought
itself, the way that knowledge was represented and
manipulated in the human mind, was a difficult
but ESSENTIAL subject for scientific study!
HISTORY (cont.)
• Copernican: Ancient Earth centered model  Rotates
around the sun
Our ideas about the world were seen as fundamentally
distinct from its appearance!
• Galileo: His scientific observations further contradicted
the “obvious”truths about the natural world and whose
development of mathematics as a tool for describing that
world emphasized the distinction between our ideas about
the world and the world it self!
• Descartes: Cognitive introspection
“Cogito ergo sum” = I think, therefore I am!!
THE TURING TEST
1950: British Mathematician Alan Turing
MACHINE INTELLIGENCE
Question: Can a machine actually think?
interrogator
Turing test (cont.)
Measures the performance of an allegedly intelligent
machine against that of a human being.
Turing called the “imitation game” places the machine and
a human counterpart in rooms apart from a second human
being, referred to as the “interrogator”. The interrogator is
not able to see or speak directly to either of them, does not
know which entity is actually the machine and may
communicate with them solely by the use of a textual
device such as a terminal. The interrogator is asked to
distinguish the computer from the human being solely on
the basis of their answers and questions asked over their
device!
Turing test (cont.)
If the interrogator cannot distinguish the machine from
the human then, Turing argues that the machine may be
assumed to be intelligent!
The important features of this test are:
1. It gives us an objective notion of intelligence
2. It prevents us from being sidetracked by such confusing
and currently unanswerable questions as whether or not
the computer uses the appropriate internal process or
whether or not the machine is actually conscious of its
actions!
3. It eliminates any bias in favor of living organisms by
forcing the interrogator to focus solely on the content of
the answers to questions
Turing test (cont.)
Turing test has been used to evaluate many AI
programs BUT there have been some criticisms:
1. It does not test abilities requiring perceptual skill
or manual dexterity (important components of
human intelligence)
2. Constrains machine intelligence to fit a human
DO WE REALLY WISH A MACHINE TO DO
MATHEMATICS AS SLOWLY AND
INACCURATELY AS A HUMAN?
AI Application areas
•
•
•
•
•
•
•
•
•
Game playing
Automated reasoning and Theorem proving
Expert Systems
Natural Language Understanding and Semantic
Modeling
Modeling Human Performance
Planning and Robotics
Languages and Environments for AI
Machine Learning
Emergent Computation