Fundamentals of Artificial Intelligence

Introduction to AI

Shyamanta M Hazarika
Indian Institute of Technology Guwahati
s.m.hazarika@iitg.ac.in

http://www.iitg.ac.in/s.m.hazarika/

Quest for Artificial Intelligence

The quest for Artificial Intelligence began with
dreams as all quests do.
People have long imagined machines endowed with
human abilities automata that move and devices that reason.
Human-like automatons are described in many stories
and are pictured in sculptures, paintings, and
drawings.
© Shyamanta M Hazarika, ME, IIT Guwahati

1
Dreams and Dreamers
For suppose thatevery tool we had
could perform its task, either at our
bidding or itself perceiving the need,
and if like… self-moved they enter the
shuttles in a loom
assembly of gods;
could fly to and fro and a plucker play a
lyre of their own accord, then master
craftsmen would have no need of servants
nor masters of slaves.
Aristotle (384 - 322 BC) Aristotle (384 - 322 BC)
The Politics The Politics
© Shyamanta M Hazarika, ME, IIT Guwahati

Dreams and Dreamers

Leonardo Da Vinci sketched designs
for a humanoid robot in the form of a
medieval knight around the year 1495.

No one knows whether Leonardo Da Vinci or

his contemporaries tried to build his
design. Leonardo’s knight was supposed to be
sit up, move
able to its arms and head,
and open its jaw.
© Shyamanta M Hazarika, ME, IIT Guwahati

2
Dreams and Dreamers
In 1651, Thomas Hobbes published his book Leviathan about
the social contract and the ideal state.

Hobbes seems to say

In the introduction that it might be
possible to build artificial animal.

George Dyson refers to

For this reason, science historian
Hobbes as the patriarch of artificial intelligence.
© Shyamanta M Hazarika, ME, IIT Guwahati

Dreams and Dreamers

For seeing life is but a motion of
limbs, the beginning whereof is in some
why may we not
principal part within,
say that all automata (engines that
move themselves by springs and wheels as doth
a watch) have an artificial life? For
what is the heart, but a spring; and the nerves,
but so many strings; and the joints, but so many
Thomas Hobbes (1588-1679) wheels, giving motion to the whole body.
Leviathan Thomas Hobbes (1588-1679)
Leviathan
© Shyamanta M Hazarika, ME, IIT Guwahati

3
Dreams and Dreamers

constructed actual automata

Several people that
moved in startlingly lifelike ways.

The most sophisticated of these was the mechanical

duck which could quack, flap its wings, paddle,
drink water, and eat and “digest" grain.
© Shyamanta M Hazarika, ME, IIT Guwahati

Dreams and Dreamers

In 1738, French inventor and engineer

Jacques de Vaucanson
(1709-1782) displayed the duck his
masterpiece.

Frederic Vidoni's ANAS, inspired by Vaucanson's duck.

© Shyamanta M Hazarika, ME, IIT Guwahati

4
Dreams and Dreamers
1801 - French silk weaver and
inventor Joseph Marie
Jacquard invents an
automated loom that is
controlled by punch cards.
Within a decade it is being mass-
produced, in great use across
Europe.
© Shyamanta M Hazarika, ME, IIT Guwahati

Robots in Fiction
Frank Baum in 1900 invented
one of the literary world’s most
beloved robots in The
Wonderful Wizard of Oz.
It is Tin Woodsman, a
mechanical man in search
of a heart!
© Shyamanta M Hazarika, ME, IIT Guwahati

5
Robots in Fiction
Some 17 years after Frank Baum, JosephCapek wrote
the short story Opilec describing automatons.

Joseph Capek’s brother Karel

Capek introduced the
term robot in the play R.U.R (Rossum’s Universal
Robots).

© Shyamanta M Hazarika, ME, IIT Guwahati

Rossum’s Universal Robots

RUR centers around a mad-

scientist who tries to usurp the
powers of God for man has
acquired the technology
and intelligence to create
life.

© Shyamanta M Hazarika, ME, IIT Guwahati

6
Elmer and Elsie
In 1948, Dr. W. Grey
Walter was interested if robots
could model brain
functions.
He built two small robots; called
tortoises and named them
Elmer and Elsie; a marvel of
the day.
© Shyamanta M Hazarika, ME, IIT Guwahati

Elmer and Elsie

The most revolutionary
thing about Elmer and Elsie is that
they did not have any brains
or pre-programming.

Had basic analog circuits,

vacuum tubes, a touch and light
sensor, could recharge their
own batteries.
© Shyamanta M Hazarika, ME, IIT Guwahati

7
What is Artificial Intelligence?
Artificial Intelligence is the ability of machines to seemingly
think for themselves.Artificial Intelligence is
demonstrated when a task performed by a human
and thought of as requiring the ability to learn,
reason and solve problems can be done by a
machine.
© Shyamanta M Hazarika, ME, IIT Guwahati

What is Artificial Intelligence?

This is itself a deep philosophical question, and
attempts to systematically answer it fall within
the foundations of Artificial Intelligence as a
rich topic for analysis and debate.

Nonetheless, a provisional answer can be given:

© Shyamanta M Hazarika, ME, IIT Guwahati

8
What is Artificial Intelligence?

AI is the field devoted to building artifacts capable

of displaying, in controlled, well-understood environments,
and over sustained periods of time, behaviors that we
consider to be intelligent, or more generally,
behaviors that we take to be at the heart of what it
is to have a mind.

© Shyamanta M Hazarika, ME, IIT Guwahati

What is Artificial Intelligence?

This gives rise to further questions.
1. What exactly constitutes intelligent behavior?
2. What it is to have a mind?
3. How humans actually manage to behave intelligently?

© Shyamanta M Hazarika, ME, IIT Guwahati

9
What is Artificial Intelligence?
How humans actually manage to behave intelligently?

question is empirical. It is predominantly for

This

psychology and cognitive science to answer.

insight
However, this question is pertinent! This is because any
into human thought might help build machines
that work similarly.
© Shyamanta M Hazarika, ME, IIT Guwahati

What is Artificial Intelligence?

What it is to have a mind?

This question on what is the mark of the mental, is

philosophical. Thrust on Artificial Intelligence has
lent significant urgency to it. Careful philosophical
contemplation of this question has influenced the
course of Artificial Intelligence itself.
© Shyamanta M Hazarika, ME, IIT Guwahati

10
What is Artificial Intelligence?
What exactly constitutes intelligent behavior?

specifying precisely what is to

The first question of
count as intelligent behavior, has traditionally been met
by proposing particular behavioral tests whose
successful passing would signify the presence of
intelligence.

© Shyamanta M Hazarika, ME, IIT Guwahati

Dimensions of Artificial Intelligence

Thinking

Thinking Humanly Thinking Rationally

Rationally
Humanly

Acting Humanly Acting Rationally

Acting
© Shyamanta M Hazarika, ME, IIT Guwahati

11
Dimensions of Artificial Intelligence
1. Think like Human
– model human cognition
2. Think Rationally
– formalize the inference process.
3. Act Rationally
– doing the right thing
4. Act like Human
– exhibit human behaviour
© Shyamanta M Hazarika, ME, IIT Guwahati

Dimensions of Artificial Intelligence

1. Think like Human
– model human cognition

1960s "cognitive revolution": information-processing

psychology
Requires scientific theories of internal activities
of the brain

Example: Newell & Simon’s GPS (General Problem Solver)

© Shyamanta M Hazarika, ME, IIT Guwahati

12
Thinking Humanly
The General Problem Solver, developed in 1957 by
Alan Newell and Herbert Simon, embodied a grandiose
vision: a single computer program that could
solve any problem, given a suitable description of the problem.

caused quite a stir when it

The General Problem Solver
was introduced, and some people in AI felt it would sweep in
a grand new era of intelligent machines.
© Shyamanta M Hazarika, ME, IIT Guwahati

Dimensions of Artificial Intelligence

2. Think Rationally
– formalize the inference process.

Several Greek schools developed various forms of

logic: notation and rules of derivation for
thoughts; may or may not have proceeded to the idea of
mechanization

Direct line through mathematics and philosophy to modern AI

© Shyamanta M Hazarika, ME, IIT Guwahati

13
Thinking Rationally
Aristotle considered rationality to be an essential
characteristic of the human mind. Perhaps the
deepest contribution of Aristotle to artificial intelligence
was the idea of formalism ….. notion that
remains at the heart of the contemporary computational
theory of the mind and what is strong AI. The very idea of an intelligent
machine was often tantamount to a machine that can
perform logical inference.
© Shyamanta M Hazarika, ME, IIT Guwahati

Thinking Rationally
Aristotle was one of the firsts to attempt to codify
"thinking". His syllogisms provided patterns of
argument structure that always gave correct
conclusions, given correct premises.
Example:
All computers use energy.
Using energy always generates heat.
Therefore, All computers generate heat.
© Shyamanta M Hazarika, ME, IIT Guwahati

14
Thinking Rationally
Main obstacles to the logistic approach in building
programs to create intelligence
1. Not all intelligent behavior is mediated by logical
deliberation.

2. What is the purpose of thinking? What thoughts should I

have?

3. Informal knowledge is not precise; Difficult to model

uncertainty.

4. Theory and practice, hard to put together.

© Shyamanta M Hazarika, ME, IIT Guwahati

Dimensions of Artificial Intelligence

3. Act Rationally
– doing the right thing

Rational behavior: doing the right thing - one which is

expected to maximize goal achievement, given the available
information

Does not necessarily involve thinking

– e.g., bionic reflex.
© Shyamanta M Hazarika, ME, IIT Guwahati

15
Acting Rationally
It is more general than the logical approach.

Amenable to scientific development than approaches

based on human behavior or human thought

perfect rationality in complex environments is

Achieving
not possible because the computational demands are too
high.

© Shyamanta M Hazarika, ME, IIT Guwahati

Dimensions of Artificial Intelligence

4. Act like Human
– exhibit human behaviour

machines that perform functions

Creating
that require intelligence when the same
tasks are performed by people.

© Shyamanta M Hazarika, ME, IIT Guwahati

16
 Acting Humanly
A number of capabilities need to be incorporated
 Natural language processing
 Knowledge Representation
 Automated Reasoning
 Machine Learning
 Computer Vision
 Robotics
© Shyamanta M Hazarika, ME, IIT Guwahati

Machines with True Intelligence

In 1950 Alan Turing published a
landmark paper in which he speculated
about the possibility of creating
machines with true intelligence. He
noted that "intelligence" is difficult to define and
devised his famous Turing Test.
The Turing Test was the first serious
proposal in the philosophy of
artificial intelligence.
A. M. Turing (1950) Computing Machinery and Intelligence. Mind 49: 433-460.
© Shyamanta M Hazarika, ME, IIT Guwahati

17
Can machines think?
Alan Turing, laying
the ground for what later became known
as artificial intelligence, starts his landmark paper
Computing Machinery and Intelligence with the words:

I propose to consider the question, ‘Can machines think?’

.
A. M. Turing (1950) Computing Machinery and Intelligence. Mind 49: 433-460.

© Shyamanta M Hazarika, ME, IIT Guwahati

Imitation Game
Turing test or Imitation Game as it was called in the paper, was
put forth as a simple test that could be used to prove that machines

could think.
Human

Human Interrogator
AI System

© Shyamanta M Hazarika, ME, IIT Guwahati

18
Turing and Heuristic Search
code-breakers were pitted against
When Turing and fellow

the Nazi code machine Enigma. Turing broke the previously

inviolable indicator system of Enigma, and helped design
electromechanical machines to read thousands of German radio
intercepts.

employed heuristic searching, which is now a

These devices
central idea of AI. They found the right answer -often enough and fast
enough to be read in real time. Without such machines, German U-boats would
have decimated the North Atlantic convoys providing a lifeline to Britain.
© Shyamanta M Hazarika, ME, IIT Guwahati

Early Days of AI
In late 1955, Allen Newell and
Herbert Simon developed The
Logic Theorist, considered by
many to be the first AI program.

The program, representing

each problem as a tree
model, would attempt to solve it by
selecting the branch that would most
result in the correct
likely
conclusion.
© Shyamanta M Hazarika, ME, IIT Guwahati

19
Early Days of AI
In 1956 John McCarthy organized
a conference to draw the talent and
expertise in machine intelligence for a
month of brainstorming. He invited them
to Vermont for The Dartmouth
Summer Research Project on
Artificial Intelligence. From that
point on, the field would be known as
Artificial intelligence. The Dartmouth
conference served lay the
to
groundwork for the future of AI
research.
© Shyamanta M Hazarika, ME, IIT Guwahati

Term ‘Artificial Intelligence’

We propose that a 2 month, 10 man study ofartificial intelligence be
carried out during the summer of 1956 at Dartmouth College in
Hanover, New Hampshire. The study is to proceed on the basis of the
conjecture that every aspect of learning or any other feature of
intelligence can in principle be so precisely described that a
machine can be made to simulate it.
J. McCarthy, M. L. Minsky, N. Rochester, and C.E. Shannon. August 31, 1955.

John McCarthy is one of the "founding fathers" of artificial intelligence,

together with Marvin Minsky, Allen Newell, and Herbert A. Simon.
© Shyamanta M Hazarika, ME, IIT Guwahati

20
Weak vs. Strong AI

Weak AI aims at building machines that act

intelligently, without taking a position on whether or not the
machines actually are intelligent.

Strong AI is the field devoted to building persons!

Charniak and McDermott (1985) concede in their classic introduction
to AI that we are very far from achieving strong AI.

© Shyamanta M Hazarika, ME, IIT Guwahati

Weak vs. Strong AI

Theultimate goal of AI, which we are very far from
achieving, is to build a person, or, more humbly, an animal.
Charniak & McDermott 1985

Charniak and McDermott don’t

say that the ultimate goal
is to build something that appears to be a person. Their
brand of AI is so-called strong AI, an ambitious form of the field aptly
summed up by Haugeland.
© Shyamanta M Hazarika, ME, IIT Guwahati

21
Weak vs. Strong AI
The fundamental goal [of AI research] is not merely to
mimic intelligence or produce some clever fake. Not at all. AI
wants only the genuine article: machines with minds, in the
full and literal sense. This is not science fiction, but real science,
based on a theoretical conception as deep as it is
daring: namely, we are, at root, computers
ourselves.
Haugeland 1985
© Shyamanta M Hazarika, ME, IIT Guwahati

Weak vs. Strong AI

The goal of work in artificial intelligence is to build machines

that perform tasks normally requiring human
intelligence.
Nilsson, Nils J. (1971)
Problem-Solving Methods in Artificial Intelligence

© Shyamanta M Hazarika, ME, IIT Guwahati

22
What is involved?
Interaction with the real world i.e., perceive, understand, and
act. For example: a. speech recognition and b. image
understanding.
Reasoning and planning involving a. modeling the external world
b. planning and decision making and c. deal with unexpected
problems and uncertainties.

Learning and Adaptation through Internal models being always

updated such as a baby learning to categorize and recognize
animals
© Shyamanta M Hazarika, ME, IIT Guwahati

What is involved?
Philosophy Logic, Methods of reasoning, Mind as physical
system, Foundations of Learning / Language
Rationality.

Mathematics Formal representation and Proof Theory

Algorithms: computation – Decidability /
Tractability

Statistics/ Modeling uncertainty, Learning from data

Probability
Economics Utility, Decision Theory
© Shyamanta M Hazarika, ME, IIT Guwahati

23
What is involved?
Neuroscience Neurons as information processing units.

Psychology/ How do people behave, perceive, process

Cognitive Science cognitive information, represent knowledge
Computer Engg. Building fast computers

Control theory Design systems that maximize an objective

function over time
Linguistics Knowledge Representation, Grammars
© Shyamanta M Hazarika, ME, IIT Guwahati

History of AI
1943: early beginnings
 McCulloch & Pitts: Boolean circuit model of brain

1950s: initial promise

 Turing's "Computing Machinery and Intelligence“
 Samuel's checkers program
 First program of Machine Learning
 Early AI programs
 Newell & Simon's Logic Theorist

© Shyamanta M Hazarika, ME, IIT Guwahati

24
History of AI
1955-65: Great enthusiasm!
Dartmouth meeting
"Artificial Intelligence“ name adopted
Newell and Simon: GPS, general problem solver
Gelertner: Geometry Theorem Prover
McCarthy: invention of LISP

1966—73: Reality dawns

Realization that many AI problems are intractable
Limitations of existing neural network methods
identified
© Shyamanta M Hazarika, ME, IIT Guwahati

History of AI
1969—85: Adding domain knowledge
 Development of knowledge-based systems
 Success of rule-based expert systems,
 E.g., DENDRAL, MYCIN
 But were brittle and did not scale well in practice

1986 onwards : Rise of machine learning

 Neural networks return to popularity
 Major advances in machine learning
 Algorithms and Applications

© Shyamanta M Hazarika, ME, IIT Guwahati

25
History of AI

Beginning 1990: Role of uncertainty

Bayesian networks
 knowledge representation framework

Beginning 1995: AI as Science

Integration of learning, reasoning, knowledge
representation
AI methods used
 vision, language, data mining, etc

© Shyamanta M Hazarika, ME, IIT Guwahati

History of AI

© Shyamanta M Hazarika, ME, IIT Guwahati

26
Annotated Bibliography
Artificial Intelligence: A
Philosophical Introduction
Jack Copeland
Backwell, 1993

The book reviews the progress made in

AI since the inception of the field in
1956. Copeland goes on to analyze what
those working in AI must achieve before
they can claim to have built a thinking
machine!
© Shyamanta M Hazarika, ME, IIT Guwahati

Annotated Bibliography
The Quest for Artificial
Intelligence
Nils J. Nilsson
Cambridge, 2009

End-of-chapter notes with citations to

important materials will be of great
use to AI scholars and researchers.
This book traces the history of a field
that has captivated the imaginations
of scientists, philosophers, and
writers for centuries.
© Shyamanta M Hazarika, ME, IIT Guwahati

27
Annotated Bibliography
Machine Learning:
The New AI
Ethem Alpaydin
MIT Press, 2016

A concise overview of machine

learning — computer programs that
learn from data — which underlies
applications that include
recommendation systems, face
recognition, and driverless cars.
© Shyamanta M Hazarika, ME, IIT Guwahati

28