Introduction of Artificial Intelligence related to BIT course.pdf

On this presentation slide we discussed about Artificial intelligence. Here is the content available In artificial intelligence.
What is Intelligence/Artificial Intelligence (AI)
• History of AI
• AI Perspectives (Defining AI)
• Turing Test
• Foundations of AI,
• Scope of Symbolic AI
• Applications of AI

1. Artificial Intelligence
BIT 4
th
Semester
Instructor:
Shiv Raj Pant
Prepared by Shiv Raj Pant1

2. About the course/Class plan
•The course introduces basic concepts behind AI technology
•Basic knowledge of Algorithms, Data structures, Logic theory, Algebra, calculus
is required
Books:
•
Text Book -Stuart Russel and Peter Norvig, Artificial Intelligence A
Modern Approach, Pearson
Accessory materials:

Lecture slides

Notes compiled by other instructors
Prepared by Shiv Raj Pant
2

3. Course Content:
Unit 1: Introduction to AI (3 Hrs.)
Unit 2: Agents and environments (5 Hrs.)
Unit 3: Problem solving by searching (10 Hrs.)
Unit 4: Knowledge representation (14 Hrs.)
Unit 5: Neural networks (2 Hrs.)
Unit 6: Machine learning (5 Hrs.)
Unit 7: Expert systems (3 Hrs.)
Unit 8: Natural language processing (3 hrs)
Prepared by Shiv Raj Pant3

4. Unit 1
Introduction to AI
•
What is Intelligence/Artificial Intelligence (AI)
•
History of AI
•
AI Perspectives (Defining AI)
•
Turing Test
•
Foundations of AI,
•
Scope of Symbolic AI
•
Applications of AI
Introduction
Prepared by Shiv Raj Pant4

5. What is Intelligence ?
Intelligence is:
–the ability to reason
–the ability to understand
–the ability to create
–the ability to Learn from experience
–the ability to plan and execute complex tasks
What is Artificial Intelligence?
"Giving machines ability to perform tasks normally associated with human
intelligence."
According to Barr and Feigenbaum:
“Artificial Intelligence is the part of computer science concerned with designing
intelligence computer systems, that is, systems that exhibit the characteristics we
associate with intelligence in human behavior.”
Prepared by : Shiv Raj Pant
5

6. AI Perspectives (AI approaches) ?
•
Since the inception of AI, researchers have tried to define AI with different perspectives.
•
The question here is what we want to achieve with AI ?
•
Different researchers have proposed different approaches to develop AI.
•
Accordingly, we have many different definitions of AI depending on the approach to deal
with AI.
•
There are 4 major perspectives in defining AI:
1.
Acting humanly approach
2.
Thinking humanly approach
3.
Thinking rationally
4.
Acting rationally
Prepared by : Shiv Raj Pant
6

7. 1. Acting humanly approach
•This perspective views the field of AI as a
science of creating systems that Act 1like
humans
“ AI is The art of creating machines that
perform functions that require intelligence
when performed by people.” (Kurzweil, 1990)
“AI is The study of how to make computer do
things at which, at the moment, people are
better.” (Rich and Knight, 1991)
2. Thinking humanly approach
•This approach tries to mimic the working of
human brain in machines.
“The exciting new effort to make computers
think…..machine with minds, in the full and
literal sense.” (Haugeland, 1985)
“[The automaton of] activities that we
associate with human thinking, activities such
as decision-making, problem solving,
learning…..” (Bellman, 1978)
Prepared by : Shiv Raj Pant
7
3. Thinking rationally
•
This approach tries to mimic the rational part of
the human mind in machines.
“The study of mental faculties through the use of
computational models.” (Charniakand McDermott,
1985)
“The study of the computations that make it possible
to perceive, reason, and act.” (Winston, 1992)
4. Acting rationally
“Computational Intelligence is the study of the
design of intelligent agents.” (Poole et al., 1998)
“AI… is concerned with intelligent behavior in
artifacts.” (Nilsson, 1998)

8. History of AI
•
The concept and theory of AI was first envisioned by Alan Turing. For his contribution,
Turing is called “Father of AI”
•
The first work in the field of AI was done by W. McCulloch and W. Pitts (1943).
•
They proposed a model of artificial neuron in which each neuron is characterized as “on” or
“off” occurring in response to stimulation by neighboring neurons.
•
They used 3 sources :
-
Neurology: Basic physiology and function of neurons in brain.
-
Propositional logic theory
-
Turing’s theory of computation
•
Then various researchers worked in the field of AI.
•
The name “Artificial Intelligence” was given in 1956.
•
Later, with the development of powerful computers, several AI languages were developed like
Lisp, prolog etc.
Prepared by : Shiv Raj Pant
8

9. Important timelines in History of AI
The beginning of concept of AI
•
1943: Warren Mc Cullochand Walter Pitts: a
model of artificial booleanneurons to
perform computations.
•
1951: Marvin Minsky and Dann Edmonds
constructed the first neural network computer
•
1950: Alan Turing’s “Computing Machinery and
Intelligence” -First complete vision of AI.
Great expectations (1952-1969):
•
A lot of theoretical concepts from
researchers but Very Little progress towards
implementation.
•
Theory mostly focused on imitation of human
problem-solving
•
Arthur Samuel (1952-) investigated game
playing (checkers ) with great success.
•
John McCarthy(1958-) : Inventor of Lisp
(second-oldest high-level language) Logic
oriented, separation between knowledge and
reasoning
Prepared by : Shiv Raj Pant
9
Collapse in AI research (1966 -1973) :
•
Progress was slower than expected.
•
Unrealistic predictions.
•
Some systems lacked scalability.
•
Fundamental limitations on techniques and
representations.
•
Lack of powerful machines.
AI revival through knowledge-based systems
(1969-1970):
•
General-purpose vs. domain specific
•
E.g. the DENDRAL project -First successful
knowledge intensive system.
•
Expert systems : MYCIN to diagnose blood
infections (Feigenbaumet al.)
•
Introduction of uncertainty in reasoning.
•
Increase in knowledge representation
research.
AI becomes industry and science (1980-present)
•
The first commercial expert system R1 in
1982
•
Speech recognition, neural networks,
robotics and many more…

10. Foundations of AI:
•
The field of AI requires theories, concepts,
and technologies from various other areas.
Some major foundational areas of AI are-
Philosophy:
o
Logic, reasoning, mind as a physical
system, foundations of learning, language
and rationality.

Where does knowledge come from?

How does knowledge lead to action?

How does mental mind arise from physical
brain?

Can formal rules be used to draw valid
conclusions?
Mathematics:
o
Formal representation and proof algorithms,
computation, undecidability, intractability,
probability.

What are the formal rules to draw the valid
conclusions?

What can be computed?

How do we reason with uncertain information?
Prepared by : Shiv Raj Pant
10
Psychology:
o
Adaptation, phenomena of perception and
motor control.

How humans and animals think and act?Economics:
o
Formal theory of rational decisions, game
theory, operation research.

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 future?Linguistics:
•
Knowledge representation, grammar

How does language relate to thought?
Neuroscience:
o
Physical substrate for mental activities

How do brains process information?
Control theory:
o
stability, optimal agent design

How can artifacts operate under their own
control?

11. More on AI approaches ...
1. Acting Humanly: The Turing Test Approach
•
The pioneer of AI theory –Alan Turing-
was the first to view AI as an approach to
build intelligent machines that act like
humans.
How to test whether a machine is intelligent
enough to act like humans ?
i. One way is to list out the actions of
humans! and test if a machine can do them
i.e. what humans do ?
Humans think.
Humans eat.
Humans swim.
Humans run ...
...
This list is exhaustive!
Prepared by : Shiv Raj Pant
11
ii. Alan Turing in his paper “computing
machinery and intelligence” suggested that
instead of enumerating the actions of
humans, we should see whether a machine can
pass a behavioral intelligence test
•
The Turing test, proposed by Alan Turing
(1950) was designed to convince the people
that whether a particular machine can think
or not.
•
He suggested a test based on
indistinguishability from undeniably
intelligent entities-human beings.
•
The test involves an interrogator who
interacts with one human and one machine.
Within a given time the interrogator has to
find out which of the two the human is, and
which one the machine.

12. Prepared by : Shiv Raj Pant
12
•
The computer passes the test if a human
interrogator after posing some written
questions, cannot tell whether the written
response come from human or not.
•
Turing proposed that to pass a Turing test,
a computer must have following capabilities:
o
Natural Language Processing: Must be able to
communicate successfully in English
o
Knowledge representation: To store what it
knows and hears.
o
Automated reasoning: Answer the Questions based
on the stored information.
o
Machine learning: Must be able to adapt in new
circumstances.
•
Turing test avoids the physical interaction
with human interrogator. Physical simulation
of human beings is not necessary for
intelligence.
Total Turing test
•
The total Turing test includes video signals
and manipulation capability so that the
interrogator can test the subject’s
perceptual abilities and object manipulation
ability.
•
To pass the total Turing test computer must
have following additional capabilities:
o
Computer Vision: To perceive objects
o
Robotics: To manipulate objects and move

13. How many questions to ask ?
What if the human interrogator distinguishes
half of the answers and unable to distinguish
other answers ?
•
To solve above problems, Turing suggested to
ask the questions for 5 minutes.
•
The machine(or program) passes the test if
it fools the interrogator 30% of the time.
What happened to the Turing test approach of
AI ?
•
In 1950, Turing predicted that, by the year
2000, a computer with a storage of 10
9
units
could be programmed well enough to pass the
Turing test.
•
This is 2022! Did it happen ?
Prepared by : Shiv Raj Pant
13
•
There were some programs which fooled people
for 5 minutes.
Examples:

Eliza program (see Wikipedia)

Internet chatbotMGONZ

Program ALICE fooled a judge in 2001
•
Because of the little applicability to
real-world problems, the researchers paid
little attention to Turing test approach.

14. 2. Thinking Humanly: Cognitive modeling approach
•
If we are going to say that a given program thinks like a human, we must have some way of
determining how humans think.
•
We need to get inside the actual workings of human minds.
•
There are two ways to understand how humans think:
Introspection: catch our own thoughts while they go by
psychological experiments. Observe other peoples behaviour
•
Once we have precise theory of mind, it is possible to express the theory as a computer
program.
•
The computer which runs such a program is said to be thinking like humans.
•
The field of cognitive science brings together computer models from AI and experimental
techniques from psychology to try to construct precise and testable theories of the workings
of the human mind.
Prepared by : Shiv Raj Pant
14

15. 3. Thinking rationally: The laws of thought
approach
•
The acting/thinking humanly approaches have
a demerit –irrationality!
•
Human mind have two parts:
Rational part: making right reasoning from
what is known.
Irrational part: making wrong inference
from given knowledge. For example: Human
may tell a lie.
•
Aristotle was one of the first who attempt
to codify the right thinking that is
irrefutable reasoning process.
•
He gave Syllogisms that always yielded
correct conclusion when correct premises are
given.
Prepared by : Shiv Raj Pant
15
For example:
Ram is a man.
All men are mortal.
⇒Ram is mortal
•
These laws of thought were supposed to
govern the operation of mind
•
This study initiated the field of logic.
•
The logicistapproach in AI hopes to create
intelligent systems using logic programming.

16. 4. Acting Rationally: The rational Agent
approach:
•
Agent is something that acts.
•
Computer agent is expected to have following
attributes:
o
Autonomous control
o
Perceiving their environment
o
Persisting over a prolonged period of time
o
Adapting to change
o
And capable of taking on another’s goal
What is Rational behavior?
-doing the right thing.
The right thing: that which is expected to
maximize goal achievement, given the available
information.
Rational Agent is one that acts so as to achieve
the best outcome or, when there is uncertainty,
the best expected outcome.
Prepared by : Shiv Raj Pant
16
•
In the “laws of thought” approach to AI, the
emphasis was given to correct inferences.
•
Making correct inferences is sometimes part of
being a rational agent, because one way to act
rationally is to reason logically to the
conclusion and act on that conclusion.
•
However, there are also some ways of acting
rationally that cannot be said to involve
inference.
•
For Example, recoiling from a hot stove is a
reflex action that usually more successful than
a slower action taken after careful
deliberation (thinking).
Advantages:
•
It is more general than laws of thought
approach, because correct inference is just one
of several mechanisms for achieving
rationality.
•
It is more amenable to scientific development
than are approaches based on human behavior or
human thought because the standard of
rationality is clearly defined and completely
general.

17. Rationality vs Omniscience
•
We need to be careful to distinguish between rationality and omniscience.
Omniscience
•
Omniscience means perfection.
•
A an omniscient agent knows the actual outcome of its actions and can act accordingly. An
omniscient agent is always successful.
•
But omniscience is impossible in reality.
Rationality
•
Rationality is not the same as perfection.
•
Rationality means “doing the right thing according to the given knowledge.”
•
It may not always be successful.
•
Rationality tries to maximize expected performance but Omniscience maximizes the actual
performance.
Example: Consider a person crossing the street.
Prepared by : Shiv Raj Pant
17

18. Applications of AI
•
Autonomous planning and scheduling:
o
NASA’s “Remote Agent Program” can control the scheduling and operation for a spacecraft.
•
Games:
o
IBM’s “Deep Blue” was the first computer program which defeated world champion in a chess
match.
•
Autonomous control:
o
The ALVINN computer vision system was trained to steer a car to keep it following a lane.
•
Expert systems:
o
Medical diagnosis programs based on probabilistic analysis have been able to perform at
the level of an expert physician.
•Robotics
oMicrosurgery, bomb disposal, manufacturing etc.
Prepared by : Shiv Raj Pant
18

19. End of Unit 1
Thank you!
Prepared by Shiv Raj Pant19