Sie sind auf Seite 1von 27

Artificial Intelligence



Topics we will cover
 Introduction to AI
 Objectives/Goals of AI
 What is AI?

 Approaches to AI – making computer:

 Think like a human ( Thinking humanly)
 Act like a human (Acting humanly)

 Think rationally (Thinking rationally)

 Act rationally (Acting rationally)

 The Foundations of AI
 Bits of History and the State of the Art 2
AI is the simulation of human intelligence done by machines programmed by us. The
machines need to learn how to reason and do some self-correction as needed along the
Artificial Intelligence is accomplished by studying how human brain thinks, learns,
decide, and work while trying to solve a problem, and then using the outcomes of this
study as a basis of developing intelligent software and systems. 3
What is AI?
 Artificial Intelligence (AI):
 AI is one of the newest sciences.
 AI is the study of intelligent behavior and

 AI is also concerned with the implementation of

a computer program which exhibits intelligent
 Intelligence
 Intelligence is the capability of observing, learning,
remembering & reasoning.
 AI attempts to develop intelligent agents.

 It was formally initiated in 1956

Approaches to AI
Approaches of AI fall into four categories:
 Making computer

Thinking humanly Thinking rationally

Acting humanly Acting rationally

 A human-centered approach must be an empirical science,

involving hypothesis and experimental confirmation.

 A rationalist approach involves a combination of mathematics

and engineering. 6
Acting humanly: Turing Test
 Making a computer act like a human being.

 Turing (1950) "Computing machinery and intelligence":

 "Can machines behave intelligently?"

 Operational test for intelligent behavior: the Imitation Game.

 Turing Test: Operational test for intelligent behavior.

 Suggested major components of AI: knowledge, reasoning,

language understanding, learning. 7
To pass the Turing Test computers would need to possess the
following capabilities

 Natural Language Processing - to enable it to communicate

successfully in English (or some other human language);

 Knowledge Representation - to store information provided

before or during the interrogation;

 Automated Reasoning - to use the stored information to

answer questions and to draw new conclusions;

 Machine learning - to adapt to new circumstances and to

detect and extrapolate patterns. 8
Thinking humanly: cognitive modeling
 Making a computer think like a human being.

 Requires scientific theories of internal activities of the brain.

 There are two ways to do this:

 Through introspection –

 The examination of one's own conscious thoughts and

feelings and,

 Through psychological experiments.

 The examination of one's soul in a spiritual context


 How to validate theories? Requires:

1) Predicting and testing behavior of human subjects.


2) Direct identification from neurological data.


 Both approaches (roughly, Cognitive Science and Cognitive

Neuroscience) are now distinct from AI.

Thinking rationally: laws of thought

 Making a computer think rationally

 Use the laws of logic to determine computer’s reasoning.

 A system is rational if it thinks the right thing through correct


 Aristotle: provided the correct arguments/ thought structures that

always gave correct conclusions given correct premises.

 Abebe is a man; all men are mortal; therefore Abebe is mortal

 11
These Laws of thought governed the operation of the mind and
initiated the field of Logic.

 Problem:

 Not all intelligent behavior is determined by logical

deliberation (e.g. taking your hand out of a fire)

 It is not easy to take informal knowledge.

 Also formal logic is unable to handle uncertain knowledge.

There is a big difference between being able to solve a problem

"in principle" and doing so in practice.
Acting rationally: rational agent
 Making a computer act rationally

An agent is just something that perceives and acts.

 Rational behavior: doing “the right thing”.

 “The right thing”: which is expected to maximize goal

achievement, given the available information.

 Doesn't necessarily involve thinking – e.g., blinking reflex – but

thinking should be in the service of rational action.

 Doing the right thing so as to achieve one’s goal, given one’s

beliefs. 13
 AI is the study and construction of rational

agents (an agent that perceives and acts).

 This is the focus of this course …

Rational agents

 An agent is an entity that perceives and acts.

 This course is about designing rational agents.

 Abstractly, an agent is a function from percept

histories to actions:
[f: P*  A]
A percept is a piece of information perceived by the

The Foundations of AI
 Philosophy (428 B . C .-present)

 Can formal rules be used to draw valid conclusions?

 How does the mental mind arise from a physical brain?

− Dualism: there is a part of the mind (or soul or spirit) that is outside
of nature, exempt from physical laws
− Materialism: holds that all the world (including the brain and mind)
operate according to physical law.

 Where does knowledge come from?

− Empiricism
− Induction
 How does knowledge lead to action?
The Foundations of AI

 Mathematics (800 - present)

 What are the formal rules to draw valid conclusions?
 What can be computed?

 How do we reason with uncertain information?

 Algorithm, Complexity, computability, Intractability,

Incompleteness theorem
 Intractable - the time required to solve instances of the
class grows at least exponentially with the size of the


 Economics (1776 - present)

 How should we make decisions so as to maximize payoff?

 How should we do this when others may not go along?

 How should we do this when the payoff may be far in the future?

 Utility: The mathematical treatment of "preferred outcomes

 Decision theory: combines probability theory with utility theory,

provides a formal and complete framework for decisions

The Foundations of AI

Neuroscience (1861-present)
 How do brains process information?

 Psychology (1879-present)
 How do humans and animals think and act?

 Computer engineering (1940-present)

 How can we build an efficient computer?

 Control theory and Cybernetics (1948-present)

 How can artifacts operate under their own control?

 Linguistics (1957-present)
 How does language relate to thought?
The History of Artificial Intelligence

 (1943 - 1955) - The gestation of artificial intelligence.

- 1943: McCulloch & Pitts: Boolean circuit model of brain.

- 1949: Hebbian Learning: demonstrated a simple updating rule

for modifying the connection strengths between neurons.

- 1950: Turing's "Computing Machinery and Intelligence".


 (1956) - The birth of artificial intelligence.

 The term "Artificial Intelligence" was first adopted.

 (1950’s) - A Time Of Optimism.

 (1958) - LISP invented.

- LISP stands for List Processing.

- The most common AI programming language.
 (1960’s) - AI Discovers Computational Complexity.

The History of Artificial Intelligence
 (1966 - 1973) - A dose of reality.

 AI discovers computational complexity.

 Neural network research almost disappears.

 (1969 - 1979) - Early development of knowledge-based systems

(expert systems). The key to power?

 (1980 - present) - AI becomes an industry.

 (1986 - present) - Neural networks return to popularity.

 (1987 - present) - AI becomes a science.

 (1995 - present) - The emergence of intelligent agents.
The State of the Art
 What can AI do today?

 Autonomous planning and scheduling.

 Game playing.
 Autonomous control.
 Diagnosis: Medical diagnosis programs.
 Logistics Planning.
 Robotics.
 Language understanding and problem solving.
Future of A.I
• The future is really UNKNOWN.
• But we can imagine two different kind of future of A.I.
They are :
1) Positive
2) Negative

Positive imagination of Future

Negative imagination of Future

Any Question?