1.Introduction.ppt

ADVANCE ARTIFICIAL INTELLIGENCE
CS 5003
Dr. Bharat Singh
Department of Computer Science and Engineering
Indian Institute of Information Technology, Ranchi

COURSE OBJECTIVE
 To learn the difference between optimal reasoning Vs human
like reasoning.
 To understand the notion of state space representation along with
time and space complexities.
 To learn the methods of solving problems using Artificial
Intelligence.
 To introduce the concepts of machine learning.
 Able to work in uncertain environments using probabilistic
reasoning techniques.
2

SYLLABUS OF ADV. AI COURSE
 Unit-1
 Introduction
Introduction to AI, History, Overview, AI techniques, Turing test, Intelligent
agents, Performance measure, rationality, structure of agents, problem
solving agents,
 Problem Spaces & Search Defining problem as a Space, uninformed
Search strategies, informed heuristic search and exploration, greedy best
search, A* search, Memory bounded heuristic search, Heuristic function,
inventing admissible heuristic search, local search algorithm, Hill climbing,
simulated annealing, genetic algorithm, Online Search.
 Unit-2
 Constraint satisfaction problems, Backtracking search, variable and value
ordering, constraint propagation, intelligent backtracking, Local search for
CSP,
 Adversarial search, Games, the Minimax algorithm, alpha-beta pruning,
Imperfect real time decision, games that include an element of chance.
3

SYLLABUS OF ADV. AI COURSE (CONT..)
 Unit-3, Knowledge Based Agent
Logic, Propositional logic, inference, equivalence, validity and satisfiability,
resolution, forward and backward chaining, DPLL algorithm, Local search
Algorithm, FOL, Model for FOL, symbols and interpretation, terms, atomic
sentences, complex sentences, Quantifiers, inference in FOL, Unification and
lifting, forward chaining and backward chaining,
 Unit-4, Planning
 Planning, Language of planning problems, planning with state space search,
forward and backward state space search, Heuristics for state space search,
partial under planning, planning graphs, planning with propositional logic.
 Unit-5
 Uncertainty, Handling uncertain knowledge, rational decision, Probability, axioms of
probability, inference using full joint distributions, independence, Baye’s rule and
conditional independence, Bayesian Networks, Semantics of Bayesian networks,
Exact and approximate inference in Bayesian Network.
4

BOOK PREFERRED
1. Saroj Kaushik, “Artificial Intelligence”, Cengage Learning India Pvt Ltd
2. N.J. Nilsson, “Principles of Artificial Intelligence”, Narosa Publishing
House.
3. E. Rich and Knight, “Artificial Intelligence”, McGraw Hill International.
4. Stuart Russell, Peter Norvig, “Artificial Intelligence – A Modern Approach”,
Pearson Education / Prentice Hall of India.
5. Saroj Kaushik, “Logic and Prolog Programming”, New Age International
Pvt Ltd
5

WHAT IS ARTIFICIAL INTELLIGENCE?
 Someone intelligence is their ability to understand and
learn things.
 Intelligence is the ability to think and understand
instead of doing things by instinct or automatically.
 Thinking is the activity of using your brain to consider a
problem or to create an idea.
 The ability to learn and understand, to solve problems
and to make decision.

WHAT IS ARTIFICIAL INTELLIGENCE?
 Turing defined the intelligent behavior of a computer as
the ability to achieve human-level performance in
cognitive tasks.

SOME DEFINITIONS (I)
The exciting new effort to make
computers think …
machines with minds,
in the full literal sense.
Haugeland, 1985
(excited but not really useful)

SOME DEFINITIONS (II)
The study of mental faculties through the use
of computational models.
Charniak and McDermott, 1985
A field of study that seeks to explain and
emulate intelligent behavior in terms of
computational processes.
Schalkoff, 1990
(Applied psychology & philosophy?)

SOME DEFINITIONS (III)
The study of how to make
computers do things at which, at
the moment, people are better.
Rich & Knight, 1991
(I can almost understand this one).

AI DEFINITIONS
 The study of how to make programs/computers do
things that people do better
 The study of how to make computers solve problems
which require knowledge and intelligence
 The exciting new effort to make computers think …
machines with minds
 The automation of activities that we associate with
human thinking (e.g., decision-making, learning…)
 The art of creating machines that perform functions
that require intelligence when performed by people
 The study of mental faculties through the use of
computational models
 A field of study that seeks to explain and emulate
intelligent behavior in terms of computational
processes
 The branch of computer science that is concerned
with the automation of intelligent behavior
Thinking
machines or
machine
intelligence
Studying
cognitive
faculties
Problem
Solving and
CS

SOME DEFINITIONS (IV)
In Japan AI Experience in 2017,
"AI is a computer system able to perform tasks that
ordinarily require human intelligence... Many of
these artificial intelligence systems are powered by
machine learning, some of them are powered by
deep learning and some of them are powered by
very boring things like rules." DataRobot CEO
Jeremy Achin

DIMENSIONS IN AI DEFINITIONS
 Build intelligent artifacts vs. understanding human
behavior.
 Does it matter how I built it as long as it does the
job well?
 Should the system behave like a human or behave
intelligently?
The Turing Test

ACTING HUMANLY (TURING TEST)
14

TURING TEST..RESULT
15
If the interrogator can not reliably
distinguish the human from the
computer.
Then the computer does posses
artificial intelligence

WHAT IS AI?
Thought process and reasoning
 System that act like human system that think rationaly
Turing test Law of thought/Logic
Human performance Ideal performance measure
 System that think like human system that act rationally
 Cognitive Science Rational Agents
Behavior

WHAT IS AI?
Work started in Word War-II, and name coined in 1956 by John McCarthy.
Humanly
 Thinking Humanly- Automation of activities that we associate with
human thinking, decision making. (by Haugeland, 1978)
 Acting Humanly- The study of how to make computers do things at
which, at the moment, people are better. (by Rich and Knight, 1991)
1/14/2023
17

WHAT IS AI? (CONT..)
Rationally
 Thinking Rationally- The study of computations that make it
possible to perceive, reason, and act. (by Poole et al. 1998)
 Acting Rationally- AI concerned with intelligent behaviour.
(by Nilsson, 1998)
1/14/2023
18

ACTING HUMANLY (TURING TEST)
(CONT..)
 Coined in 1950.
 A computer passes the test if a human interrogation after
posing some written question, can not tell whether the
written response come from person or a computer.
1/14/2023
19

ACTING HUMANLY (TURING TEST) CONT..
 To pass the test, study require in 6 disciplines.
o NLP
o Knowledge Representation
o Automated Reasoning (To use stored information and draw
new conclusion.)
o Machine Learning ( to adapt new circumstances and to detect
and extrapolate patterns.)
o Computer Vision (To perceive object and environment)
o AI Robotics (To manipulated Object and to move)
1/14/2023
20

THINKING HUMANLY
Cognitive Modelling Approach
 If we say, program should think like Human.
We need to determining how human thing.
 1961- General problem solve by Herbert Simon and Allen Newell
(Trace of human subjects solving the same problem).
1/14/2023
21

THINKING HUMANLY (CONT..)
Cognitive Modelling Approach
 Interdisciplinary fields of cognitive science brings together
computer mode of AI and Experimental technique from
psychology to construct precise and testable theories of
human mind.
1/14/2023
22

THINKING RATIONALLY
Greek Philosopher (Aristole)
 He codify right thinking.
 He said correct conclusion can achieve when given correct
premises.
IF Socrates is a man and All men are mortal.
Then Socrates is mortal.
1/14/2023
23

THINKING RATIONALLY (CONT..)
But Rationally means-
 More than correct inferences only it is one that acts so as to achieve
the best outcome or when there is uncertainty.
Correct inferences + some situations when there are no provable
things to do
1/14/2023
24

SO WHAT IS AI?
 AI as a field of study
 Computer Science
 Cognitive Science
 Psychology
 Philosophy
 Linguistics
 Neuroscience
 AI is part science, part engineering
 AI often must study other domains in order to implement systems
 e.g., medicine and medical practices for a medical diagnostic system,
engineering and chemistry to monitor a chemical processing plant
 AI is a belief that the brain is a form of biological computer and
that the mind is computational
 AI has had a concrete impact on society but unlike other areas of
CS, the impact is often
 felt only tangentially (that is, people are not aware that system X has AI)
 felt years after the initial investment in the technology

WHAT IS INTELLIGENCE?
 Is there a “holistic” definition for intelligence?
 Here are some definitions:
 the ability to comprehend; to understand and profit from experience
 a general mental capability that involves the ability to reason, plan,
solve problems, think abstractly, comprehend ideas and language, and
learn
 is effectively perceiving, interpreting and responding to the
environment
 None of these tells us what intelligence is, so instead, maybe
we can enumerate a list of elements that an intelligence must be
able to perform:
 perceive, reason and infer, solve problems, learn and adapt, apply
common sense, apply analogy, recall, apply intuition, reach emotional
states, achieve self-awareness
 Which of these are necessary for intelligence? Which are
sufficient?
 Artificial Intelligence – should we define this in terms of human
intelligence?
 does AI have to really be intelligent?
 what is the difference between being intelligent and demonstrating
intelligent behavior?

WHAT DOES AI REALLY DO?
 Knowledge Representation
 (how does a program represent its domain of discourse?)
 Automated reasoning.
 Planning
 (get the robot to find the bananas in the other room).
 Machine Learning
 (adapt to new circumstances).
 Natural language understanding.
 Machine vision, speech recognition, finding data on the
web, robotics, and much more.

BRAIN VS. COMPUTER
 In AI, we compare the brain (or the mind) and the
computer
 Our hope: the brain is a form of computer
 Our goal: we can create computer intelligence through
programming just as people become intelligent by learning
But we see that the computer
is not like the brain
The computer performs tasks
without understanding what
its doing
Does the brain understand
what its doing when it solves
problems?

A BRIEF HISTORY OF AI
 The dream of making a computer imitate us began
many years ago…..
 The Dartmouth conference, Summer ‘56.
 Early enthusiasm 52-59:
 Puzzle solving with the General Problem Solver, Geometry
theorem prover, Checkers player, Lisp.
 Reality strikes:
 Programs don’t scale up.
 The problem is not as easy as we thought:
 The spirit is willing but the flesh is weak -->

HISTORY OF AI
 1943-1955
Recognition of AI was done by Warren McCulloch and Walter Rits. Its
comprises 3 fields—
• Knowledge of basic physiology and function of neurons.
• Propositional Logic by Russel et al..
• Turing’s Theory of Computations.
Two undergraduate students of Harvard Marvin Minisky and Dean Edmonds
built first Neural Network.
1/14/2023
30

HISTORY OF AI(CONT...)
 1947
Alan Turing influential lecture in 1947 introduced Turing Test, Machine
Learning and Genetic Algorithms and Reinforcement learning.
“He said instead of trying to produce a programme to simulate
the adult mind, why not rather try to produce one which
simulate the child mind.”
1/14/2023
31

 1956
First program to perform automated reasoning to prove theorems
was build by Newell and Simon.
 1952-1969
John McCarthy develop high level LISP which was to become
AI programming language for next 30 years.
1/14/2023
32

 1958
John McCarthy develop Advice Taker a hypothetical program that can be
seen as the first complete AI system.
It uses knowledge to search for solution to the program.
 For Example: It showed how simple program to generate a plan to drive to
the airport using Knowledge Representation and Reasoning.
1/14/2023
33

 1966-1973
Herbert Simon quoted that there is now a world machine can
think, learn and create.
He predicted that within 10 years a computer would be Chess
Champion. But the prediction came true actually after 40 years.
1/14/2023
34

LACK OF KNOWLEDGE REPRESENTATION
 Because early program knew nothing of their subject matter, they succeeded only by
means of syntactic manipulations.
 US National Research Council after the launch of Sputnik in
1957 wanted to speed up the project of translation the
Russian scientific papers to English.
 But, the projects cancelled because it depends only on the world replacement from an
electronic dictionary.
1/14/2023
35

LACK OF INTRACTABILITY
 Many problems that AI was attempting to solve tried different
combinations of steps until the solution was found.
 It worked initially because it contains very few actions and short
solution sequence.
 Before the theory of Computational Complexity was
developed, it widely thought that scaling up to larger problems
was simply matter of faster Hardware and larger memories
1/14/2023
36

KNOWLEDGE BASED SYSTEMS
 1969-1979
General Purpose Search mechanism, considered weak methods that do not scale up to
large or difficult problem instances.
 1969-1979
Alternatively, use domain specific knowledge that allows larger reasoning steps.
 1987 till today
Various AI techniques with Machine Learning methods used for large available datasets.
 Now, in current era, the availability of large datasets become common.
1/14/2023
37

MORE HISTORY
 Knowledge-based systems (expert systems) 1969-
1979:
 Ed Feigenbaum (Stanford): Knowledge is power! (as
opposed to weak methods)
 Dendral (inferring molecular structure from a mass spectrometer).
 MYCIN: diagnosis of blood infections
 AI becomes an industry:
 R1: configuring computers for DEC.
 Robotic vision applications

RECENT EVENTS: 1987-PRESENT
 AI turns more scientific, relies on more
mathematically sophisticated tools:
 Hidden Markov models (for speech recognition)
 Belief networks.
 Focus turns to building useful artifacts as opposed
to solving the grand AI problem.
 The victory of the neats over the scruffies?

A RICH HISTORY
 Philosophy
 Mathematics
 Biology
 Economics
 Neuroscience
 Psychology
 Control Theory
 Linguistic
 Computer Science
 John McCarthy- coined the term- 1950’s

RECENT AI SUCCESSES
 Deep Blue beats Kasparov (AI?)
 Theorem provers proved an unknown theorem.
 Expert systems: medical, diagnosis, design
 Speech recognition applications (in limited domains).
 Robots controlling quality in factories.
 Intelligent agents on board Deep Space 1.

SO WHAT DOES AI DO?
 Most AI research has fallen into one of two categories
 Select a specific problem to solve
 study the problem (perhaps how humans solve it)
 come up with the proper representation for any knowledge needed to
solve the problem
 acquire and codify that knowledge
 build a problem solving system
 Select a category of problem or cognitive activity (e.g., learning,
natural language understanding)
 theorize a way to solve the given problem
 build systems based on the model behind your theory as experiments
 modify as needed
 Both approaches require
 one or more representational forms for the knowledge
 some way to select proper knowledge, that is, search

WHY AI ? AND ITS IMPORTANCE
44

INTELLIGENT BEHAVIOUR
 Perception
 Reasoning
 Learning
 Understanding Language
 Solving Problem

APPLICATIONS
Computer vision
Image recognition
Robotics
NLP
Speech Processing
Machine Translation
Autonomous agents—Mangalyan, Mars Rover
Internet agents

PRACTICAL IMPACT OF AI
 AI components are embedded in numerous devices
e.g. copy machines.
 AI systems are in everyday use
 Detecting credit card fraud
 Configuring products
 Aiding complex planning tasks
 Advising physicians

TYPES OF AI
 Strong AI
 Weak AI
 Applied AI
 Cognitive AI

TYPES OF AI
 Strong AI aims to build machine that can truly
reason and solve problems which is self aware and
whose overall intellectual ability is
indistinguishable from that of a human being.
 Also called General AI.
 Human like
 Non-Human like

APPROACHES TO AI
 Weak AI deals with the creation of some form of
computer-based artificial intelligence that cannot
truly reason and solve problems, but can act as if it
were intelligent.
 Also called Narrow AI.
 Weak AI holds that suitably programmed machines
can simulate human cognition.

APPROACHES TO AI
 Applied AI- aims to produce commercial viable
smart system-such as for example a security
system that is able to recognize the faces of people
who are permitted to enter a particular building
 Applied AI has already enjoyed considerable
success.
 Example-Recognize people

APPROACHES TO AI
 Cognitive AI- Computer are used to test theories
about how the human mind works-for example,
theories about how we recognize faces and other
objects, or about how we solve abstract problem.

TYPES OF ARTIFICIAL INTELLIGENCE
 1. Reactive Machines
 2. Limited Memory
 3. Theory of Mind
 4. Self-Awareness

 1. Reactive Machines
 Reactive machines perceive present external information and plan actions
accordingly.
 The machines perform specialized duties and only understand the task at
hand.
 The most basic types of AI systems are purely reactive, and
have the ability neither to form memories nor to use past
experiences to inform current decisions.
 The machines’ behavior is consistent, given a repeated situation.
 In the 1990s, IBM developed a reactive machine named Deep Blue to
play competitive chess, predicting chess moves by identifying each
piece’s board placement.
 They can’t interactively participate in the world, the way we
imagine AI systems one day might. Instead, these machines
will behave exactly the same way every time they encounter
the same situation.

2. Limited Memory
 Machines can look into the past.
 It’s characterized by the ability to absorb learning data and
improve over time based on its experience similar to the way
the human brain’s neurons connect. This is the AI that is
widely used and being perfected today.
 Limited memory machines can harness recent observations to
make informed decisions.
 The machines consider observational data in reference to
their pre-programmed conceptual framework.
 The observational data is retained for a limited period and
then forgotten.
 Deep learning algorithms and the deep learning revolution of
2012 made limited memory AI possible.
 Self-Driving Car

 3. Theory of Mind
 Theory of mind– to understanding that people,
creatures and objects in the world can have thoughts
and emotions that affect their own behavior.
 Theory of mind machines can form thoughts and make
decisions in reference to emotional context; thus, they can
participate in social interaction.
 The machines are still in the development stage; however,
many exhibit aspects of human-like capability.
 For example
 Consider voice assistant applications that can comprehend basic
speech prompts and commands but cannot hold a conversation.
 Robots Kismet (introduced in 2000) and Sophia (2016)

 4. Self-Awareness
 To build systems that can form representations about themselves.
Ultimately, we AI researchers will have to not only understand
consciousness, but build machines that have it.
 Self-awareness machines demonstrate intelligent behavior through
 ideation, the formation of ideas or concepts.
 the formation of desires, and
 understanding their internal states.
 When self-aware AI is achieved we would have AI that has human-
level consciousness and equals human intelligence with the same
needs, desires and emotions.
 In 1950, Alan Turing developed the Turing Test to identify machines
that could behave indistinguishably from a human being.

TYPES OF LEARNING IN ARTIFICIAL INTELLIGENCE
 Narrow Intelligence
 General Intelligence
 Super Intelligence

 Narrow Intelligence
 Additionally known as Weak AI.
 Stage of Artificial Intelligence between machines that could
perform a narrowly defined set of specific tasks.
 The machine doesn’t possess any believing capability, but it
simply plays a set of functions that are pre-defined.
 Cases of Weak AI include Siri, Alexa, Self-driving automobiles,
Alpha-Go, Sophia the Ninja and so on. Just about all of AI-
based methods built till this date fall below the category of
Weak AI.

 General Intelligence
 Additionally known as Strong AI.
 Stage at the development of Synthetic Intelligence where
machines will probably possess the ability to believe and make
conclusions like people human beings.
 There are now no Present cases of Strong AI, yet, It is
considered that we will soon be able to generate machines
which can be as smart as human beings.
 Strong AI Is Regarded as a hazard to human existence
with lots of Scientists, including Stephen Hawking who
stated that:
 “The development of complete Artificial Intelligence could
stop the end of the Human Era…

 Super Intelligence
 Artificial Super Intelligence is the stage of Artificial Intelligence
as soon as the capability of computer systems may transcend
human beings.
 This is currently a hypothetical position as depicted in pictures
and science fiction novels, in which makers have obtained on
the planet.

WHY AI ? AND ITS IMPORTANCE (CONT..)
 AI start-up's in World
https://www.edureka.co/blog/artificial-intelligence-
applications/
 AI start-up's in India
https://www.mygreatlearning.com/blog/top-ai-startups-in-
india/
62

WHY AI ? AND ITS IMPORTANCE (CONT..
 World Wide Competition in AI field
https://www.kaggle.com/competitions
https://www.rsna.org/en/education/ai-resources-and-
training/ai-image-challenge
63

1.Introduction.ppt

More Related Content

What's hot (20)

Similar to 1.Introduction.ppt (20)

Recently uploaded (20)

1.Introduction.ppt