Chapter 1

Introduction to Artificial
Intelligence

Berhanu F.(MSc)
 Intelligence
• Intelligence is the capability of observing, learning,
remembering and reasoning
• AI attempts to develop intelligent agents

Characteristics of Intelligent systems

• Use vast amount of knowledge
• Learn from experience and adopt to changing
environment
• Interact with human using natural language and
speech
• Respond in real time
• Tolerate error and ambiguity in communication
 Artificial Intelligence
• The concern of AI is to enable computers behave
like human and emulate the reasoning power of
humans
• in order to do tasks that require human intelligence.
– Which task requires intelligence?
– Complex arithmetic operations
– For instance, Solving 220 * 350?
– Mundane tasks
– Example, Natural language understanding; face
recognition
– Expert tasks: which require specialists knowledge
– Example, Medical diagnosis; computer maintenance
 Artificial Intelligence
• AI is one of the newest sciences
– Work started after world war II
– The name itself coined in 1956
• Definitions according to 8 text books
Systems that think like humans Systems that think rationally

Systems that act like humans Systems that act rationally

• Taken as approaches
• All four approaches to AI have been followed
 Acting humanly: the Turing test
approach
• “Computing machinery and intelligence” (Turing,
1950)
• “Can machines think?” -> “Can machines behave
intelligently?”
• Instead of proposing a long and controversial list
of qualifications required for intelligence, he
suggested a test based on indistinguishability
from human beings
– operational test for intelligence - the imitation game,
now called the Turing Test
 Acting humanly: the Turing test
approach – cont.

• The computer (AI system) passes the test if a human

interrogator, after posing some written questions,
cannot tell whether the written responses come from a
person or not
 Acting humanly: the Turing test
approach – cont.
• Predicted that by 2000, a machine might have a 30% chance
of fooling a lay person for 5 minutes
• Suggested major components of AI:

– Knowledge base – to store what the system knows or hears

– Reasoning – to use the stored knowledge to answer questions, to
draw new conclusions
– language understanding – to communicate successfully in natural
(human) language
– Learning – to adapt to new circumstances, to detect and to extrapolate
patterns

• Problem: Turing test is not reproducible, constructive, or

amenable to mathematical analysis
 Thinking humanly: the cognitive
modeling approach
Requires scientific theories of internal activities of
the brain
 We need to get the inside the actual workings of human
minds.
Through introspection – trying to catch our own thoughts as they
go by (direct identification from neurological data – bottom-up)
Through psychological experiments – predicting and testing
from human subjects (top-down approach)
Once we have a sufficiently precise theory of the
mind, it becomes possible to express the theory as a
computer program
 Thinking rationally: the “laws of thought
approach”

Aristotle was the first to attempt to codify rational

thinking - «right thinking»
 syllogism provided patterns for argument structures that
always gave correct conclusions given correct premises
Syllogism is a kind of logical argument in which one proposition
(the conclusion) is inferrred from two or more others (the
premises) of a certain form
Socrates is a man
all men are mortal
therefore Socrates is mortal
These laws of thought were supposed to govern the
operation of the mind, and initiated the field of logic
 Thinking rationally: the “laws of thought
approach”

Several Greek schools developed various forms of

logic 
 Notations and rules of derivation for thoughts
By 1965, programs existed that could take a
description of a problem in logical notation and find
the solution to the problem, if one exists
The logicist approach within artificial intelligence
hopes to build on such programs to create intelligent
systems
 Thinking rationally: the “laws of thought
approach”
Problems:
 it is not easy to state informal knowledge in formal terms
required by logical notation
particularly when the knowledge is not completely certain
 there is a big difference between being able to solve a
problem “in principle” and doing so in practice
problems with just a few dozen facts can exhaust the
computational resources of any computer unless it has some
guidance as to which reasoning steps to try first
 Acting rationally: the rational agent approach

Acting rationally – doing the right thing so as to

achieve one's goal, given one's beliefs
The law of thought approach to AI concentrate on
correct inference – right thinking
However, doing the right thing doesn't necessarily
involve thinking
 e.g. Reflex action
 But right thinking should be in the service of right action
one way to act rationally is to reason logically to the conclusion
that a given action will achieve one's goals, and then to act on that
conclusion
 Acting rationally: the rational agent approach
In this approach, AI is viewed as the study and
construction of rational agents
Agent – an agent is something/an entity that
perceives and acts
Rational agents – agents that act rationally/right
This approach to AI has the following advantages:
 it is more general than the “laws of thought” approach
 it is more amenable to scientific development than
approaches based on human behavior or human thought
the standard of rationality is clearly defined and completely general
 AI Prehistory

Disciplines that contributed ideas, viewpoints and

techniques to AI
 Philosophy
• Logic – set of rules that can describe the formal, rational part
of the mind
• Mind as a physical system that manipulates knowledge
• Establishing the source of knowledge
• Empiricism theory - asserts that knowledge comes primarily
via sensory experience
• Mind as the connection between knowledge and action
 Vital for AI

• Intelligence requires action and reasoning

 AI Prehistory – cont.

 Mathematics
Formal representation and proof algorithms e.g. Boolean logic
Computation

Incompleteness theorem - Some functions can not be

computed or represented by an algorithm
Intractability – a problem is intractable if the time required
to solve instances of the problem grows exponentially with the
size of the instances
•Exponential growth means that even moderately large
instances can not be solved in any reasonable time
•Divide the overall problem of generating intelligent system
into tractable sub-problems
 AI Prehistory – cont.

Probability – helps to deal with uncertain measurements and

incomplete theories
Thomas Bayes : proposed a rule for updating probabilities in
the light of new evidence
Bayes’ rule and the resulting field called Bayesian analysis
form the basis of most modern approaches to uncertain
reasoning in AI systems
 Economics
Utility– the mathematical treatment of preferred outcomes
Decision theory – combines probability theory and utility theory

Provides a formal and complete framework for decisions made

under uncertainity
Game theory – for some games, a rational agent should adopt
policies that are (or least appear to be) randomized
 AI Prehistory – cont.

 Neuroscience
Isthe study of the nervous system, particularly the brain
plastic physical substrate for mental activity

 Psychology
phenomena of perception and motor control, experimental
techniques
 Computer engineering
building fast computers necessary for the development of AI
systems
 Control theory and Cybernetics
design
systems that maximize an objective function over time
homeostatic systems, stability simple optimal agent designs
 AI Prehistory – cont.

Linguistics
 knowledge representation, grammar
 History of AI

1943 - McCulloch & Pitts: Boolean circuit model of

brain
 Proposed a model of artificial neurons in which each neuron is
characterized as being “on” or “off”
aswitch to “on” occurring in response to stimulation by a
sufficient number of neighboring neurons
 Showed that any computable function could be computed by
some network of connected neurons, and that all the logical
connectives (and, or, not) could be implemented by simple net
structure
 Suggested that suitably defined networks could learn
 History of AI – cont.

1949 – Donald Hebb demonstrated Hebbian

learning, an influential model to this date
 a simple updating rule for modifying the connection strength
1950 – Turing’s “computer Machinery and
intelligence”
 Introduced the Turing test, machine learning, genetic
algorithms and reinforcement learning
1956 -The birth of AI
 Dartmouth workshop
Introduceall the major figures to each other
Name of the field, Artificial Intelligence suggested by McCarthy,
has been adopted
 History of AI – cont.

 Why it was necessary for AI to become a separate field?

AIembraced the idea of duplicating human faculties like
creativity, self-improvement, and language use
Methodology:

A branch of computer science

Attempt to build machines that will function autonomously in

complex, changing environment
Early enthusiasm, great expectation (1952-1969)
A dose of reality (1966-1973)
Knowledge-based systems: the key to power (1969-
1979)
 History of AI – cont.

AI becomes an industry (1980-present)

The return of neural network (1986-present)
AI becomes a science (1987 - present)
The emergence of intelligent agents (1995-present)
 State of the Art in AI
Autonomous planning and scheduling
A hundred million miles from Earth, NASA's Remote Agent program
became the first on-board autonomous planning program to control
the scheduling of operations for a spacecraft
Game playing
IBM's Deep Blue became the first computer program to defeat the world
champion in a chess match
 Autonomous control
 Diagnosis
 Logistic planning
 Robotics
 Language understanding and problem solving
 Assignment 1 (5%)

•Compare and contrast Strong AI vs.

Weak AI?
Assignment