19-10-2024

UNIT-1

Artificial Intelligence

V SEM
Department of Computer Science (AIML)

Nawab Shah Alam Khan College of Engineering and Technology

Unit-1 Syllabus
 Introduction to Intelligence
 Foundation of AI
 History of AI
 Structure of Agents
 Problem Solving
 Formulating Problems
 Problem types, States and Operators
 State Space
 Search Strategies – Uninformed & Informed
 - BFS, DFS, Best First Search, A* , Heuristic functions,
Iterative deepening A*
 Adversarial search / Game Playing – Perfect decision
games and imperfect decision games, evaluation function,
alpha beta pruning

1
 19-10-2024

Intelligence & Cognition

Intelligence: is the ability to easily learn or understand

things and to deal with new or difficult situations.

Cognition: is the mental process of acquiring

knowledge and understanding through thought,
experience, and the senses.

Cognition

2
 19-10-2024

Cognition

Emotion & Cognition

3
 19-10-2024

Emotion & Cognition

Emotion & Cognition

4
 19-10-2024

Emotion – Positive effect

Emotion – Positive effect

10

5
 19-10-2024

Emotion – Negative effect

11

Language & Thought

Which came first?
Language or Thought?

The thought comes first, while language is an

expression.

Research
Babies are born with the ability to think about objects
before they learn language.

12

6
 19-10-2024

Intelligence & Cognition

13

Difference - Intelligence and Cognition

Intelligence Cognition
Intelligence is the ability to Cognition is the mental process
easily learn or understand of acquiring knowledge and
things and to deal with new or understanding through thought,
difficult situations. experience, and the senses.

Capacity for learning, Mechanisms of how you learn,

reasoning, understanding, and remember, and pay attention.
similar forms of mental
activity;
Intelligence is basically the Without cognition abilities, one
ability to perform well in cannot show his intelligence.
cognitive tasks.

14

7
 19-10-2024

Artificial Intelligence

What Is AI?
Universally accepted definition:
AI is the study of how to make computers do things
which, at the moment, people do better.

15

Artificial Intelligence – by John McCarthy

 The term Artificial Intelligence was coined by John
McCarthy, father of Artificial Intelligence, in 1956.

 According to John McCarthy, AI is “The science

and engineering of making intelligent machines,
especially intelligent computer programs”.

 He defines intelligence as “the computational part

of the ability to achieve goals in the world”.

16

8
 19-10-2024

Intelligence
 Intelligence is that property of mind which encompasses many
related mental abilities like:
 Learning
 Reasoning
 Inferencing
 Problem solving
 Perceiving
 Store Knowledge
 Comprehend ideas and communicate in Natural language
 Learn new concepts and ideas from environment
 Offer advice based on rules and situations

17

Understanding of AI

Understanding of AI requires understanding of related

terms such as :
Intelligence,
knowledge,
reasoning,
cognition
learning
and other computer related terms.

18

9
 19-10-2024

Artificial Intelligence
Various authors have defined AI differently. These definitions of AI are organized
into four categories:

Thinking Humanly Thinking Rationally

“The exciting new effort to make computers think . . “The study of mental faculties through the
. machines with minds, in the full and literal sense.” use of computational models.”

“The study of the computations that make

“[The automation of] activities that we associate
it possible to perceive, reason, and act.”
with human thinking, activities such as decision-
making, problem solving, learning . . .”
Acting Humanly Acting Rationally
“The art of creating machines that perform “Computational Intelligence is the study
functions that require intelligence when performed of the design of intelligent agents.”
by people.”

“AI . . . is concerned with intelligent

“The study of how to make computers do behaviour in artifacts.”
things at which, at the moment, people are better.”

19

Categories of AI
So the definitions of AI are organized into four categories:
Thinking Humanly Thinking Rationally
Acting Humanly Acting Rationally

The definitions on the :

i) top are concerned with thought processes and reasoning
ii) whereas on the bottom address behaviour.
The definitions on the :
i) left measure success in terms of fidelity (degree of exactness) to
human performance,
ii) whereas the ones on the right measure against an ideal
performance measure, called rationality.
What is rationality?
Doing the “right thing” is known as rationality.
When do you call a system is rational?
A system is rational if it does the “right thing,” given what it knows.

20

10
 19-10-2024

AI Goals

 There are two views of AI goals

 – Duplicating what brain does (cognitive science)
 – Duplicating what brain should do (logical thinking)

21

Categories of AI Systems

 To design intelligent systems, it is important to categorize these

systems. There are four categories of such systems:

1) Systems that think like humans

2) Systems that act like humans
3) Systems that think rationally
4) Systems that act rationally

22

11
 19-10-2024

Types of AI
 There are mainly two types of main categorization which are
based on capabilities and based on functionally of AI. Following
is flow diagram which explain the types of AI.


23

Types of AI
 Type 1- Based on capabilities:
i) Artificial Narrow Intelligence – dedicated to one task
ii) Artificial General Intelligence – performs like humans
iii)Artificial Super intelligence – more intelligent than humans

 Type 2- Based on functionalities:

i) Reactive machines: Doesn’t have memory to store past

experience. Focus on current scenario to react with best
possible action.
 Examples: IBM's Deep Blue system and Google’s AlphaGo
are reactive machines.
24

12
 19-10-2024

Types of AI
ii) Limited Memory : use past- and present-moment data to
decide on a course of action most likely to help achieve a
desired outcome.
Examples:
1. Generative AI tools like ChatGPT, Bard and DeepAI.
2. Virtual assistants and chatbots: Siri, Alexa, Google
Assistant, Cortana (Microsoft) and IBM Watson Assistant
3. Self-driving cars: Autonomous vehicles use Limited
Memory AI to understand the world around them in real-
time and make informed decisions on when to apply
speed, brake, make a turn, etc.

25

Types of AI
 iii) Theory of Mind: understand the thoughts and emotions
of other entities. Currently in development.

iv) Self-awareness: is a kind of functional AI class for

applications that would possess super AI capabilities.
It is still hypothetical and under development.
To date, Emotion AI is unable to understand and respond to
human feelings.

26

13
 19-10-2024

1. Systems that think like humans – Cognitive Approach

 One has to know the functioning of the brain and its
mechanism for possessing information.
 To say that a program thinks like a human, we
must know how humans think. We can learn about
human thought in three ways:
a) introspection—trying to catch our own thoughts
as they go by;
b) psychological experiments—observing a person
in action;
c) brain imaging—observing the brain in action.

 Neural network is a computing model for processing

information similar to the brain.
27

2. Systems that act like humans - The Turing Test

approach
 The overall behavior of the system should be human like.
 It could be achieved by observation.
 The Turing Test, proposed by Alan Turing (1950), was
designed to provide a satisfactory operational definition of
intelligence.
 Turing Test: A computer passes the test if a human
interrogator, after posing some written questions, cannot
tell whether the written responses come from a person or
from a computer.

28

14
 19-10-2024

The Turing Test

Turing proposed operational test for intelligent behavior in 1950.

Human

Human ?
Interrogator
AI system

29

2. Systems that act like humans - The Turing Test

approach
 Programming a computer to pass a rigorously applied test requires:

1. Natural language processing to communicate successfully in a

human language.
2. Knowledge representation to store what it knows or hears;
3. Automated reasoning to answer questions and to draw new
conclusions.
4. Machine learning to adapt to new circumstances and to detect and
deduce patterns.
 To pass the total Turing Test the computer will need 2 more:

 5) Computer vision: to perceive objects

 6) Robotics: to manipulate objects and move about.
 These six disciplines compose most of AI.

30

15
 19-10-2024

ELIZA

31

3. Systems that think rationally

 Such systems rely on logic rather than humans to measure
correctness.
 For thinking rationally or logically, logic formulas and
theories are used for synthesizing outcomes.
 For example, given:
Ahmad is a human. All humans are mortal.
Then one can conclude logically that “Ahmed is mortal”

Intelligent behavior is converted into logical

considerations.

32

16
 19-10-2024

4. Systems that act rationally

 Acting rationally is the final category of Intelligent
System
 Rational behavior means doing the right thing.
 Even if method is illogical, the observed behavior
must be rational (right) .

33

Characteristics required in AI Program

AI programs must have the capability and characteristics of
intelligence such as:
1) learning
2) reasoning,
3) inferencing,
4) perceiving
5) and comprehending information (to understand something
completely) .

34

17
 19-10-2024

Components of AI Program
 AI techniques must be independent of the problem domain.

 Any AI program should have the following components:

1) Knowledge base: AI program should be learning in nature
and update its knowledge accordingly.
 Knowledge base generally consists of facts and rules and has
the following characteristics:
o a) It is voluminous in nature and requires proper structuring
o b) It may be incomplete and imprecise
o c) It may be dynamic and keep on changing

 2) Control Strategy: determines which rule to be applied. To

know the rule some heuristics or thumb rules based on
problem domain are used.
35

Components of AI Program
 3) Inference mechanism: requires:

- searching through knowledge base

- and derive new knowledge using the existing knowledge
with the help of inference rules.

36

18
 19-10-2024

Foundations of AI
Foundation of AI is based on
1. Philosophy
2. Mathematics
3. Economics
4. Neuroscience
5. Control Theory
6. Linguistics
7. Computer Engineering
8. Psychology

37

Foundations – 1. Philosophy

 Can formal rules be used to draw valid conclusions?

 How does the mind arise from a physical brain?
 Where does knowledge come from?
 How does knowledge lead to action?

38

19
 19-10-2024

Foundations – 2. Mathematics
 Use of more formal logical methods
Boolean logic
Fuzzy logic
 Uncertainty
The basis for most modern approaches to handle
uncertainty in AI applications can be handled by
 Probability theory
 Modal and Temporal logics

39

Foundations – 3. Economics

 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?

40

20
 19-10-2024

Foundations – 4. Neuroscience
How do the brain work?
Early studies (1824) relied on injured and abnormal
people to understand what parts of brain work
More recent studies use accurate sensors to correlate
brain activity to human thought
By monitoring individual neurons, monkeys can now
control a computer mouse using thought alone
How close are we to have a mechanical brain?
Parallel computation, remapping, interconnections,….

41

Foundations – 5. Control Theory

Machines can modify their behavior in response
to the environment (sense/action loop)
Water-flow regulator, steam engine governor,
thermostat
The theory of stable feedback systems (1894)
Build systems that make transition from initial
state to goal state with minimum energy

In 1950, control theory could only describe

linear systems and AI largely rose as a
response to this shortcoming

42

21
 19-10-2024

Foundations – 6. Linguistics
 How does language relate to thought?
 Speech demonstrates so much of human intelligence
Analysis of human language reveals thought taking place
in ways not understood in other settings
Children can create sentences they have never heard
before
Language and thought are believed to be tightly
intertwined

43

Foundations – 7. Computer Engineering

 How can we build an efficient computer ?

8. Psychology
 How do humans and animals think and act ?

44

22
 19-10-2024

Sub-areas of AI
 AI is interdisciplinary field having diverse subfields.
Each one these fields in an area of research in AI
itself. Some of them are:
1. Neural networks
2. Computer vision
3. Robotics
4. Data mining
5. Game playing methodologies
6. Fuzzy, model and temporal logic
7. Expert problem solving
8. Web agents

45

Sub-areas of AI – Contd..
 1. Neural Networks
 Neural Networks are inspired by human brains
and copies the working process of human brains.
It is based on a collection of connected units or
nodes called artificial neurons or perceptrons.


 The Objective of this approach was to solve the

problems in the same way that a human brain does.

46

23
 19-10-2024

Sub-areas of AI – Contd..
Computer Vision
 Processing any image/video sources to extract
meaningful information and take action accordingly.

 AI has also developed such kind of robots which are

acquiring human activities
 Example: object recognition, image understanding ,
Playing Robots etc.

47

Sub-areas of AI – Contd..
Machine Learning
 The capability of Artificial Intelligence systems to learn by
extracting patterns from data is known as Machine Learning.
 It is a subset of Artificial Intelligence that is based on the idea
that machines can be given access to data along with the
ability to learn from it.

Speech Processing / Speech Recognition

 Speech Processing / Recognition is the ability of a computer
program to identify words and phrases in the spoken language
and convert them to machine readable format.
 The real life examples of Speech processing are Google
Assistant ,Amazon Alexa and Apple’s Siri Application etc.

48

24
 19-10-2024

Sub-areas of AI – Contd..
Robotics
 Robots are the artificial agents which behaves like humans
Build for the purpose of manipulating the objects by
o perceiving,
o picking,
o moving,
o modifying the physical properties of object,
 thereby freeing manpower from doing repetitive functions
without getting bored, distracted, or exhausted.

49

Applications

 Some of the applications are given below:

 Business: Financial strategies, give advice
 Engineering: check design, offer suggestions to create new
product
 Manufacturing: Assembly, inspection & maintenance
 Mining: used when conditions are dangerous
 Hospital: monitoring, diagnosing & prescribing
 Education: In teaching
 Household: Advice on cooking, shopping etc.
 Farming: prune trees & selectively harvest mixed crops.

50

25