Ai notes

ARTIFICIAL INTELLIGENCE
Digital Notes By
BIGHNARAJ NAIK
Assistant Professor
Department of Master in Computer Application
VSSUT, Burla

Syllabus
5th
SEMESTER MCA
F.M.- 70
MCA-308 ARTIFICIAL INTELLIGENCE (3-1-0)Cr.-4
Module I ( 10 hrs. )
Introduction to Artificial Intelligence: The Foundations of Artificial Intelligence, The History of
Artificial Intelligence, and the State of the Art. Intelligent Agents: Introduction, How Agents
should Act, Structure of Intelligent Agents, Environments. Solving Problems by Searching:
problem-solving Agents, Formulating problems, Example problems, and searching for Solutions,
Search Strategies, Avoiding Repeated States, and Constraint Satisfaction Search. Informed
Search Methods: Best-First Search, Heuristic Functions, Memory Bounded Search, and Iterative
Improvement Algorithms.
Module II ( 10 hrs. )
Agents That Reason Logically; A Knowledge-Based Agent, The Wumpus World Environment,
Representation, Reasoning & Logic prepositional Logic : A very simple Logic, An agent for the
Wumpus World.
First-Order Logic; Syntax and Semantics, Extensions and National, Variations, using First Order
Logic, Logical Agents for the Wumpus World, A Simple Reflex Agent, Representing Charge in
the World, Deducing Hidden Properties of the World, Preferences Among Actions, Toward A
Goal-Based Agent.
Building a Knowledge Base; Properties of Good and Bad Knowledge Bases, Knowledge
Engineering. The Electronic Circuits Domain, General Outology, The Grocery Shopping World.
Inference in First-Order Logic : Inference Rules Involving Quantifiers, An Example Proof.
Generalized Modus Ponens, Forward and Backward, Chaining & Completeness, Resolution: A
complete Inference Procedure, Completeness of Resolution.
Module III (10 hrs. )
Planning A Simple Planning Agent Form Problem Solving to Planning. Planning in Situation
Calculus. Basic Representations for Planning. A Partial-Order planning Example, A partial Order
planning algorithm, Planning With partially Instantiated Operators, Knowledge Engineering for
Planning.
Making Simple Decision: Combining Beliefs and desires under uncertainty. The Basis of Utility
Theory, Utility Functions. Multi attribute utility Functions, Decision Networks. The Value of
Information. Decision – Theoretic Expert Systems.

Learning in Neural and Belief Networks’ How the Brain Works, Neural Networks, perceptions,
Multi-layered Feed Forward Networks Applications Back propagation algorithm Applications of
Neural Networks.
Module IV ( 10 hrs. )
Knowledge in Learning: Knowledge in Learning, Explanation-based Learning, Learning Using
Relevance Information, Inductive Logic Programming. Agents that Communicate:
Communication as action, Types of Communicating Agents, A Formal Grammar for A subset of
English Syntactic Analysis (Parsing), Definite Clause Grammar (DCG), Augmenting A
Grammar. Semantic Interpretation. Ambiguity and Disambiguation. A Communicating Agent.
Practical Natural Language processing Practical applications. Efficient Parsing Scaling up the
lexicon. Scaling up the Grammar Ambiguity. Discourse Understanding.
Reference Books:
1. Elaine Rich, Kevin Knight, & Shivashankar B Nair, Artificial Intelligence, McGraw Hill, 3rd
ed.,2009
2. Introduction to Artificial Intelligence & Expert Systems, Dan W Patterson, PHI.,2010

MODULE WISE DESCRIPTIONS OF ALL THE CONCEPTS
Module 1:
What is Artificial Intelligence?
Artificial Intelligence (AI) is a branch of Science which deals with helping machines finding
solutions to complex problems in a more human-like fashion. This generally involves borrowing
characteristics from human intelligence, and applying them as algorithms in a computer friendly
way. A more or less flexible or efficient approach can be taken depending on the requirements
established, which influences how artificial the intelligent behaviour appears. AI is generally
associated with Computer Science, but it has many important links with other fields such as
Maths, Psychology, Cognition, Biology and Philosophy, among many others. Our ability to
combine knowledge from all these fields will ultimately benefit our progress in the quest of
creating an intelligent artificial being.
AI currently encompasses a huge variety of subfields, from general-purpose areas such as
perception and logical reasoning, to specific tasks such as playing chess, proving mathematical
theorems, writing poetry, and diagnosing diseases. Often, scientists in other fields move
gradually into artificial intelligence, where they find the tools and vocabulary to systematize and
automate the intellectual tasks on which they have been working all their lives. Similarly,
workers in AI can choose to apply their methods to any area of human intellectual endeavour. In
this sense, it is truly a universal field.
HISTORY OF AI
The origin of artificial intelligence lies in the earliest days of machine computations. During the 1940s
and 1950s, AI begins to grow with the emergence of the modern computer. Among the first researchers to
attempt to build intelligent programs were Newell and Simon. Their first well known program, logic
theorist, was a program that proved statements using the accepted rules of logic and a problem solving
program of their own design. By the late fifties, programs existed that could do a passable job of
translating technical documents and it was seen as only a matter of extra databases and more computing
power to apply the techniques to less formal, more ambiguous texts. Most problem solving work revolved
around the work of Newell, Shaw and Simon, on the general problem solver (GPS). Unfortunately the

GPS did not fulfill its promise and did not because of some simple lack of computing capacity. In the
1970’s the most important concept of AI was developed known as Expert System which exhibits as a set
rules the knowledge of an expert. The application area of expert system is very large. The 1980’s saw the
development of neural networks as a method learning examples.
Prof. Peter Jackson (University of Edinburgh) classified the history of AI into three periods as:
1. Classical
2. Romantic
3. Modern
1. Classical Period:
It was started from 1950. In 1956, the concept of Artificial Intelligence came into existance. During this
period, the main research work carried out includes game plying, theorem proving and concept of state
space approach for solving a problem.
2. Romantic Period:
It was started from the mid 1960 and continues until the mid 1970. During this period people were
interested in making machine understand, that is usually mean the understanding of natural language.
During this period the knowledge representation technique “semantic net” was developed.
3. Modern Period:
It was started from 1970 and continues to the present day. This period was developed to solve more
complex problems. This period includes the research on both theories and practical aspects of Artificial
Intelligence. This period includes the birth of concepts like Expert system, Artificial Neurons, Pattern
Recognition etc. The research of the various advanced concepts of Pattern Recognition and Neural
Network are still going on.
COMPONENTS OF AI
There are three types of components in AI
1) Hardware Components of AI
a) Pattern Matching
b) Logic Representation
c) Symbolic Processing
d) Numeric Processing
e) Problem Solving
f) Heuristic Search
g) Natural Language processing
h) Knowledge Representation
i) Expert System
j) Neural Network

k) Learning
l) Planning
m) Semantic Network
2) Software Components
a) Machine Language
b) Assembly language
c) High level Language
d) LISP Language
e) Fourth generation Language
f) Object Oriented Language
g) Distributed Language
h) Natural Language
i) Particular Problem Solving Language
3) Architectural Components
a) Uniprocessor
b) Multiprocessor
c) Special Purpose Processor
d) Array Processor
e) Vector Processor
f) Parallel Processor
g) Distributed Processor
10Definition of Artificial intelligence
1. AI is the study of how to make computers do things which at the moment people do
better. This is ephemeral as it refers to the current state of computer science and it
excludes a major area ; problems that cannot be solved well either by computers or by
people at the moment.
2. AI is a field of study that encompasses computational techniques for performing tasks
that apparently require intelligence when performed by humans.
3. AI is the branch of computer science that is concerned with the automation of intelligent
behaviour. A I is based upon the principles of computer science namely data structures
used in knowledge representation, the algorithms needed to apply that knowledge and the
languages and programming techniques used in their implementation.

4. AI is the field of study that seeks to explain and emulate intelligent behaviour in terms of
computational processes.
5. AI is about generating representations and procedures that automatically or autonomously
solve problems heretofore solved by humans.
6. A I is the part of computer science concerned with designing intelligent computer
systems, that is, computer systems that exhibit the characteristics we associate with
intelligence in human behaviour such as understanding language, learning, reasoning and
solving problems.
7. A I is the study of mental faculties through the use of computational models.
8. A I is the study of the computations that make it possible to perceive, reason, and act.
9. A I is the exciting new effort to make computers think machines with minds, in the full
and literal sense.
10. AI is concerned with developing computer systems that can store knowledge and
effectively use the knowledge to help solve problems and accomplish tasks. This brief
statement sounds a lot like one of the commonly accepted goals in the education of
humans. We want students to learn (gain knowledge) and to learn to use this knowledge
to help solve problems and accomplish tasks.
WEAK AND STRONG AI
There are two conceptual thoughts about AI namely the Weak AI and Strong AI. The strong AI is very
much promising about the fact that the machine is almost capable of solve a complex problem like an
intelligent man. They claim that a computer is much more efficient to solve the problems than some of the
human experts. According to strong AI, the computer is not merely a tool in the study of mind, rather the
appropriately programmed computer is really a mind. Strong AI is the supposition that some forms of
artificial intelligence can truly reason and solve problems. The term strong AI was originally coined by
John Searle.
In contrast, the weak AI is not so enthusiastic about the outcomes of AI and it simply says that some
thinking like features can be added to computers to make them more useful tools. It says that computers
to make them more useful tools. It says that computers cannot be made intelligent equal to human being,
unless constructed significantly differently. They claim that computers may be similar to human experts
but not equal in any cases. Generally weak AI refers to the use of software to study or accomplish specific
problem solving that do not encompass the full range of human cognitive abilities. An example of weak
AI would be a chess program. Weak AI programs cannot be called “intelligent” because they cannot
really think.

TASK DOMAIN OF AI
Areas of Artificial Intelligence
- Perception
• Machine Vision: It is easy to interface a TV camera to a computer and get an
image into memory; the problem is understandingwhat the image represents.
Vision takes lots of computation; in humans, roughly 10% of all calories
consumed are burned in vision computation.
• Speech Understanding: Speech understanding is available now. Some systems
must be trained for the individual user and require pauses between words.
Understanding continuous speech with a larger vocabulary is harder.
• Touch(tactile or haptic) Sensation: Important for robot assembly tasks.
- Robotics Although industrial robots have been expensive, robot hardware can be cheap: Radio
Shack has sold a working robot arm and hand for $15. The limiting factor in application of
robotics is not the cost of the robot hardware itself. What is needed is perception and intelligence
to tell the robot what to do; ``blind'' robots are limited to very well-structured tasks (like spray
painting car bodies).
- Planning Planning attempts to order actions to achieve goals. Planning applications include
logistics, manufacturing scheduling, planning manufacturing steps to construct a desired product.
There are huge amounts of money to be saved through better planning.
- Expert Systems Expert Systems attempt to capture the knowledge of a human expert and make
it available through a computer program. There have been many successful and economically
valuable applications of expert systems. Expert systems provide the following benefits
• Reducing skill level needed to operate complex devices.
• Diagnostic advice for device repair.
• Interpretation of complex data.
• ``Cloning'' of scarce expertise.
• Capturing knowledge of expert who is about to retire.
• Combining knowledge of multiple experts.
- Theorem Proving Proving mathematical theorems might seem to be mainly of academic interest.
However, many practical problems can be cast in terms of theorems. A general theorem prover can
therefore be widely applicable.
Examples:

• Automatic construction of compiler code generators from a description of a CPU's instruction set.
• J Moore and colleagues proved correctness of the floating-point division algorithm on AMD CPU
chip.
- Symbolic Mathematics Symbolic mathematics refers to manipulation of formulas, rather than
arithmetic on numeric values.
• Algebra
• Differential and Integral Calculus
Symbolic manipulation is often used in conjunction with ordinary scientific computation as a
generator of programs used to actually do the calculations. Symbolic manipulation programs are an
important component of scientific and engineering workstations.
- Game Playing Games are good vehicles for research because they are well formalized, small, and
self-contained. They are therefore easily programmed. Games can be good models of competitive
situations, so principles discovered in game-playing programs may be applicable to practical
problems.
AI Technique
Intelligence requires knowledge but knowledge possesses less desirable properties such as
- It is voluminous
- it is difficult to characterise accurately
- it is constantly changing
- it differs from data by being organised in a way that corresponds to its application
An AI technique is a method that exploits knowledge that is represented so that
- The knowledge captures generalisations; situations that share properties, are grouped together,
rather than being allowed separate representation.
- It can be understood by people who must provide it; although for many programs the bulk of the
data may come automatically, such as from readings. In many AI domains people must supply the
knowledge to programs in a form the people understand and in a form that is acceptable to the
program.
- It can be easily modified to correct errors and reflect changes in real conditions.
- It can be widely used even if it is incomplete or inaccurate.
- It can be used to help overcome its own sheer bulk by helping to narrow the range of possibilities
that must be usually considered.

Problem Spaces and Search
Building a system to solve a problem requires the following steps
- Define the problem precisely including detailed specifications and what constitutes an acceptable
solution;
- Analyse the problem thoroughly for some features may have a dominant affect on the chosen
method of solution;
- Isolate and represent the background knowledge needed in the solution of the problem;
- Choose the best problem solving techniques in the solution.
Defining the Problem as state Search
To understand what exactly artificial intelligence is, we illustrate some common problems. Problems
dealt with in artificial intelligence generally use a common term called 'state'. A state represents a
status of the solution at a given step of the problem solving procedure. The solution of a problem,
thus, is a collection of the problem states. The problem solving procedure applies an operator to a
state to get the next state. Then it applies another operator to the resulting state to derive a new state.
The process of applying an operator to a state and its subsequent transition to the next state, thus, is
continued until the goal (desired) state is derived. Such a method of solving a problem is generally
referred to as state space approach For example, in order to solve the problem play a game, which is
restricted to two person table or board games, we require the rules of the game and the targets for
winning as well as a means of representing positions in the game. The opening position can be
defined as the initial state and a winning position as a goal state, there can be more than one. legal
moves allow for transfer from initial state to other states leading to the goal state. However the rules
are far too copious in most games especially chess where they exceed the number of particles in the
universe 10. Thus the rules cannot in general be supplied accurately and computer programs cannot
easily handle them. The storage also presents another problem but searching can be achieved by
hashing. The number of rules that are used must be minimised and the set can be produced by
expressing each rule in as general a form as possible. The representation of games in this way leads
to a state space representation and it is natural for well organised games with some structure. This
representation allows for the formal definition of a problem which necessitates the movement from a
set of initial positions to one of a set of target positions. It means that the solution involves using
known techniques and a systematic search. This is quite a common method in AI.
Formal description of a problem
- Define a state space that contains all possible configurations of the relevant objects, without
enumerating all the states in it. A state space represents a problem in terms of states and operators
that change states

- Define some of these states as possible initial states;
- Specify one or more as acceptable solutions, these are goal states;
- Specify a set of rules as the possible actions allowed. This involves thinking about the generality of
the rules, the assumptions made in the informal presentation and how much work can be anticipated
by inclusion in the rules.
The control strategy is again not fully discussed but the AI program needs a structure to facilitate the
search which is a characteristic of this type of program.
Example:
The water jug problem :There are two jugs called four and three ; four holds a maximum of four
gallons and three a maximum of three gallons. How can we get 2 gallons in the jug four. The state
space is a set of ordered pairs giving the number of gallons in the pair of jugs at any time ie (four,
three) where four = 0, 1, 2, 3, 4 and three = 0, 1, 2, 3. The start state is (0,0) and the goal state is
(2,n) where n is a don't care but is limited to three holding from 0 to 3 gallons. The major production
rules for solving this problem are shown below:
Initial condition goal comment
1 (four,three) if four < 4 (4,three) fill four from tap
2 (four,three) if three< 3 (four,3) fill three from tap
3 (four,three) If four > 0 (0,three) empty four into drain
4 (four,three) if three > 0 (four,0) empty three into drain
5 (four,three) if four+three<4 (four+three,0) empty three into four
6 (four,three) if four+three<3 (0,four+three) empty four into three
7 (0,three) If three>0 (three,0) empty three into four
8 (four,0) if four>0 (0,four) empty four into three
9 (0,2) (2,0) empty three into four
10 (2,0) (0,2) empty four into three
11 (four,three) if four<4 (4,three-diff) pour diff, 4-four, into four from three

12 (three,four) if three<3 (four-diff,3) pour diff, 3-three, into three from four and
a solution is given below Jug four, jug three rule applied
0 0
0 3 2
3 0 7
3 3 2
4 2 11
0 2 3
2 0 10
Control strategies.
A good control strategy should have the following requirement: The first requirement is that it causes
motion. In a game playing program the pieces move on the board and in the water jug problem water
is used to fill jugs. The second requirement is that it is systematic, this is a clear requirement for it
would not be sensible to fill a jug and empty it repeatedly nor in a game would it be advisable to
move a piece round and round the board in a cyclic way. We shall initially consider two systematic
approaches to searching.
Monotonic and Non monotonic Learning :
Monotonic learning is when an agent may not learn any knowledge that contradicts what it already
knows. For example, it may not replace a statement with its negation. Thus, the knowledge base may
only grow with new facts in a monotonic fashion. The advantages of monotonic learning are:
1.greatly simplified truth-maintenance
2.greater choice in learning strategies
Non-monotonic learning is when an agent may learn knowledge that contradicts what it already
knows. So it may replace old knowledge with new if it believes there is sufficient reason to do so.
The advantages of non-monotonic learning are:

1.increased applicability to real domains,
2.greater freedom in the order things are learned in
A related property is the consistency of the knowledge. If an architecture must maintain a consistent
knowledge base then any learning strategy it uses must be monotonic.
7- PROBLEM CHARACTERISTICS
A problem may have different aspects of representation and explanation. In order to choose the most
appropriate method for a particular problem, it is necessary to analyze the problem along several key
dimensions. Some of the main key features of a problem are given below.
Is the problem decomposable into set of sub problems?
Can the solution step be ignored or undone?
Is the problem universally predictable?
Is a good solution to the problem obvious without comparison to all the possible solutions?
Is the desire solution a state of world or a path to a state?
Is a large amount of knowledge absolutely required to solve the problem?
Will the solution of the problem required interaction between the computer and the person?
The above characteristics of a problem are called as 7-problem characteristics under which the solution
must take place.
PRODUCTION SYSTEM AND ITS CHARACTERISTICS
The production system is a model of computation that can be applied to implement search algorithms and
model human problem solving. Such problem solving knowledge can be packed up in the form of little
quanta called productions. A production is a rule consisting of a situation recognition part and an action
part. A production is a situation-action pair in which the left side is a list of things to watch for and the
right side is a list of things to do so. When productions are used in deductive systems, the situation that
trigger productions are specified combination of facts. The actions are restricted to being assertion of new
facts deduced directly from the triggering combination. Production systems may be called premise
conclusion pairs rather than situation action pair.
A production system consists of following components.
(a) A set of production rules, which are of the form A→B. Each rule consists of left hand side
constituent that represent the current problem state and a right hand side that represent an output
state. A rule is applicable if its left hand side matches with the current problem state.

(b) A database, which contains all the appropriate information for the particular task. Some part of
the database may be permanent while some part of this may pertain only to the solution of the
current problem.
(c) A control strategy that specifies order in which the rules will be compared to the database of rules
and a way of resolving the conflicts that arise when several rules match simultaneously.
(d) A rule applier, which checks the capability of rule by matching the content state with the left hand
side of the rule and finds the appropriate rule from database of rules.
The important roles played by production systems include a powerful knowledge representation scheme.
A production system not only represents knowledge but also action. It acts as a bridge between AI and
expert systems. Production system provides a language in which the representation of expert knowledge
is very natural. We can represent knowledge in a production system as a set of rules of the form
If (condition) THEN (condition)
along with a control system and a database. The control system serves as a rule interpreter and sequencer.
The database acts as a context buffer, which records the conditions evaluated by the rules and information
on which the rules act. The production rules are also known as condition – action, antecedent –
consequent, pattern – action, situation – response, feedback – result pairs.
For example,
If (you have an exam tomorrow)
THEN (study the whole night)
The production system can be classified as monotonic, non-monotonic, partially commutative and
commutative.
Figure Architecture of Production System
Features of Production System
Some of the main features of production system are:

Expressiveness and intuitiveness: In real world, many times situation comes like “if this happen-you
will do that”, “if this is so-then this should happen” and many more. The production rules essentially tell
us what to do in a given situation.
1. Simplicity: The structure of each sentence in a production system is unique and uniform as they use
“IF-THEN” structure. This structure provides simplicity in knowledge representation. This feature of
production system improves the readability of production rules.
2. Modularity: This means production rule code the knowledge available in discrete pieces.
Information can be treated as a collection of independent facts which may be added or deleted from
the system with essentially no deletetious side effects.
3. Modifiability: This means the facility of modifying rules. It allows the development of production
rules in a skeletal form first and then it is accurate to suit a specific application.
4. Knowledge intensive: The knowledge base of production system stores pure knowledge. This part
does not contain any type of control or programming information. Each production rule is normally
written as an English sentence; the problem of semantics is solved by the very structure of the
representation.
Disadvantages of production system
1. Opacity: This problem is generated by the combination ofproduction rules. The opacity is generated
because of less prioritization of rules. More priority to a rule has the less opacity.
2. Inefficiency: During execution of a program several rules may active. A well devised control strategy
reduces this problem. As the rules of the production system are large in number and they are hardly
written in hierarchical manner, it requires some forms of complex search through all the production
rules for each cycle of control program.
3. Absence of learning: Rule based production systems do not store the result of the problem for future
use. Hence, it does not exhibit any type of learning capabilities. So for each time for a particular
problem, some new solutions may come.
4. Conflict resolution: The rules in a production system should not have any type of conflict
operations. When a new rule is added to a database, it should ensure that it does not have any
conflicts with the existing rules.
ALGORITHM OF PROBLEM SOLVING
Any one algorithm for a particular problem is not applicable over all types of problems in a variety of
situations. So there should be a general problem solving algorithm, which may work for different
strategies of different problems.
Algorithm (problem name and specification)
Step 1:

Analyze the problem to get the starting state and goal state.
Step 2:
Find out the data about the starting state, goalstate
Step 3:
Find out the production rules from initial database for proceeding the problem to goal state.
Step 4:
Select some rules from the set of rules that can be applied to data.
Step 5:
Apply those rules to the initial state and proceed to get the next state.
Step 6:
Determine some new generated states after applying the rules. Accordingly make them as current state.
Step 7:
Finally, achieve some information about the goal state from the recently used current state and get the
goal state.
Step 8:
Exit.
After applying the above rules an user may get the solution of the problem from a given state to another
state. Let us take few examples.
VARIOUS TYPES OF PROBLEMS AND THEIR SOLUTIONS
Water Jug Problem
Definition:
Some jugs are given which should have non-calibrated properties. At least any one of the jugs should
have filled with water. Then the process through which we can divide the whole water into different jugs
according to the question can be called as water jug problem.
Procedure:

Suppose that you are given 3 jugs A,B,C with capacities 8,5 and 3 liters respectively but are not calibrated
(i.e. no measuring mark will be there). Jug A is filled with 8 liters of water. By a series of pouring back
and forth among the 3 jugs, divide the 8 liters into 2 equal parts i.e. 4 liters in jug A and 4 liters in jug B.
How?
In this problem, the start state is that the jug A will contain 8 liters water whereas jug B and jug C will be
empty. The production rules involve filling a jug with some amount of water, taking from the jug A. The
search will be finding the sequence of production rules which transform the initial state to final state. The
state space for this problem can be described by set of ordered pairs of three variables (A, B, C) where
variable A represents the 8 liter jug, variable B represents the 5 liter and variable C represents the 3 liters
jug respectively.
Figure
The production rules are formulated as follows:
Step 1:
In this step, the initial state will be (8, 0, 0) as the jug B and jug C will be empty. So the water of jug A
can be poured like:
(5, 0, 3) means 3 liters to jug C and 5 liters will remain in jug A.
(3, 5, 0) means 5 liters to jug B and 3 liters will be in jug A.
(0, 5, 3) means 5 liters to jug B and 3 liters to jug C and jug C and jug A will be empty.
Step2:
In this step, start with the first current state of step-1 i.e. (5, 0, 3). This state can only be implemented by
pouring the 3 liters water of jug C into jug B. so the state will be (5, 3, 0). Next, come to the second

current state of step-1 i.e. (3, 5, 0). This state can be implemented by only pouring the 5 liters water of jug
B into jug C. So the remaining water in jug B will be 2 liters. So the state will be (3, 2, 3). Finally come to
the third current state of step-1 i.e. (0, 5, 3). But from this state no more state can be implemented because
after implementing we may get (5, 0, 3) or (3, 5, 0) or (8, 0, 0) which are repeated state. Hence these
states are not considerably again for going towards goal.
So the state will be like:
5, 0, 3 5, 3, 0
3, 5, 0 3, 2, 3
0, 5, 3 X
Step 3:
In this step, start with the first current state of step-2 i.e. (5, 3, 0) and proceed likewise the above steps.
5, 3, 0 2, 3, 3
3, 2, 3 6, 2, 0
Step 4:
In this step, start with the first current state of step-3 i.e. (2, 3, 3) and proceed.
2, 3, 3 2, 5, 1
6, 2, 0 7, 0, 1
Step 5:
2, 5, 1 7, 0, 1
6, 0, 2 1, 5, 2
Step6:
7, 0, 1 7, 1, 0
1, 4, 3 1, 4, 3
Step7:
7, 1, 0 4, 1, 3

1, 4, 3 4, 4, 0 Goal
So finally the state will be (4, 4, 0) that means jug A and jug B contains 4 liters of water each which is our
goal state. One thing you have to very careful about the pouring of water from one jug to another that the
capacity of jug must satisfy the condition to contain that much of water.
The tree of the water jug problem can be like:
Figure
Comments:
This problem takes a lot of time to find the goal state.
This process of searching in this problem is very lengthy.

At each step of the problem the user have to strictly follow the production rules. Otherwise the
problem may go to infinity step.
Missionaries and Carnivals Problem
Definition:
In Missionaries and Carnivals Problem, initially there are some missionaries and some carnivals will be at
a sideof a river. They want to cross the river. But there is only one boat available to cross the river. The
capacity of the boat is 2 and no one missionary or no Carnivals can cross the river together. So for solving
the problem and to find out the solution on different states is called the Missionaries and Carnival
Problem.
Procedure:
Let us take an example. Initially a boatman, Grass, Tiger and Goat is present at the left bank of the river
and want to cross it. The only boat available is one capable of carrying 2 objects of portions at a time. The
condition of safe crossing is that at no time the tiger present with goat, the goat present with the grass at
the either side of the river. How they will cross the river?
The objective of the solution is to find the sequence of their transfer from one bank of the river to the
other using the boat sailing through the river satisfying these constraints.
Let us use different representations for each of the missionaries and Carnivals as follows.
B: Boat
T: Tiger
G: Goat
Gr: Grass
Step 1:
According to the question, this step will be (B, T, G, Gr) as all the Missionaries and the Carnivals are at
one side of the bank of the river. Different states from this state can be implemented as

`The states (B, T, O, O) and (B, O, O, Gr) will not be countable because at a time the Boatman and the
Tiger or the Boatman and grass cannot go. (According to the question).
Step 2:
Now consider the current state of step-1 i.e. the state (B, O, G, O). The state is the right side of the river.
So on the left side the state may be (B, T, O, Gr)
i.e.B, O, G, O B, T, O, Gr
(Right) (Left)
Step 3:
Now proceed according to the left and right sides of the river such that the condition of the problem must
be satisfied.
Step 4:
First, consider the first current state on the right side of step 3 i.e.
Now consider the second current state on the right side of step-3 i.e.
Step 5:
Now first consider the first current state of step 4 i.e.

Step 6:
Step 7:
Hence the final state will be (B, T, G, Gr) which are on the right side of the river.
Comments:
This problem requires a lot of space for its state implementation.
It takes a lot of time to search the goal node.
The production rules at each level of state are very strict.
Chess Problem
Definition:
It is a normal chess game. In a chess problem, the start is the initial configuration of chessboard. The final
state is the any board configuration, which is a winning position for any player. There may be multiple
final positions and each board configuration can be thought of as representing a state of the game.
Whenever any player moves any piece, it leads to different state of game.

Procedure:
Figure
The above figure shows a 3x3 chessboard with each square labeled with integers 1 to 9. We simply
enumerate the alternative moves rather than developing a general move operator because of the reduced
size of the problem. Using a predicate called move in predicate calculus, whose parameters are the
starting and ending squares, we have described the legal moves on the board. For example, move (1, 8)
takes the knight from the upper left-hand corner to the middle of the bottom row. While playing Chess, a
knight can move two squares either horizontally or vertically followed by one square in an orthogonal
direction as long as it does not move off the board.
The all possible moves of figure are as follows.
Move (1, 8) move (6, 1)
Move (1, 6) move (6, 7)
Move (2, 9) move (7, 2)
Move (2, 7) move (7, 6)
Move (3, 4) move (8, 3)
Move (3, 8) move (8, 1)
Move (4, 1) move (9, 2)
Move (4, 3) move (9, 4)
The above predicates of the Chess Problem form the knowledge base for this problem. An unification
algorithm is used to access the knowledge base.
Suppose we need to find the positions to which the knight can move from a particular location, square 2.
The goal move (z, x) unifies with two different predicates in the knowledge base, with the substitutions
{7/x} and {9/x}. Given the goal move (2, 3), the responsible is failure, because no move (2, 3) exists in
the knowledge base.

Comments:
In this game a lots of production rules are applied for each move of the square on the chessboard.
A lots of searching are required in this game.
Implementation of algorithm in the knowledge base is very important.
8- Queen Problem
Definition:
“We have 8 queens and an 8x8 Chess board having alternate black and white squares. The queens are
placed on the chessboard. Any queen can attack any other queen placed on same row, or column or
diagonal. We have to find the proper placement of queens on the Chess board in such a way that no queen
attacks other queen”.
Procedure:
Figure A possible board configuration of 8 queen problem
In figure , the possible board configuration for 8-queen problem has been shown. The board has
alternative black and white positions on it. The different positions on the board hold the queens. The
production rule for this game is you cannot put the same queens in a same row or same column or in same
diagonal. After shifting a single queen from its position on the board, the user have to shift other queens
according to the production rule. Starting from the first row on the board the queen of their corresponding
row and column are to be moved from their original positions to another position. Finally the player has
to be ensured that no rows or columns or diagonals of on the table is same.
Comments:
This problem requires a lot of space to store the board.

It requires a lot of searching to reach at the goal state.
The time factor for each queen’s move is very lengthy.
The problem is very strict towards the production rules.
8- Puzzle Problem
Definition:
“It has set off a 3x3 board having 9 block spaces out of which 8 blocks having tiles bearing number from
1 to 8. One space is left blank. The tile adjacent to blank space can move into it. We have to arrange the
tiles in a sequence for getting the goal state”.
Procedure:
The 8-puzzle problem belongs to the category of “sliding block puzzle” type of problem. The 8-puzzle is
a square tray in which eight square tiles are placed. The remaining ninth square is uncovered. Each tile in
the tray has a number on it. A tile that is adjacent to blank space can be slide into that space. The game
consists of a starting position and a specified goal position. The goal is to transform the starting position
into the goal position by sliding the tiles around. The control mechanisms for an 8-puzzle solver must
keep track of the order in which operations are performed, so that the operations can be undone one at a
time if necessary. The objective of the puzzles is to find a sequence of tile movements that leads from a
starting configuration to a goal configuration such as two situations given below.
Figure (Starting State) (Goal State)
The state of 8-puzzle is the different permutation of tiles within the frame. The operations are the
permissible moves up, down, left, right. Here at each step of the problem a function f(x) will be defined
which is the combination of g(x) and h(x).
i.e.
Where
g x: how many steps in the problem you have already done or the current state from the initial state.
h x: Number of ways through which you can reach at the goal state from the current state or
Or
F(x)=g(x) + h (x)

h xis the heuristic estimator that compares the current state with the goal state note down how many
states are displaced from the initial or the current state. After calculating the f (x) value at each step
finally take the smallest f (x) value at every step and choose that as the next current state to get the goal
state.
Let us take an example.
Figure (Initial State) (Goal State)
Step1:
f xis the step required to reach at the goal state from the initial state. So in the tray either 6 or 8 can
change their portions to fill the empty position. So there will be two possible current states namely B and
C. The f (x) value of B is 6 and that of C is 4. As 4 is the minimum, so take C as the current state to the
next state.
Step 2:
In this step, from the tray C three states can be drawn. The empty position will contain either 5 or 3 or 6.
So for three different values three different states can be obtained. Then calculate each of their f (x) and

take the minimum one.
Here the state F has the minimum value i.e. 4 and hence take that as the next current state.
Step 3:
The tray F can have 4 different states as the empty positions can be filled with b4 values i.e.2, 4, 5, 8.
Step 4:
In the step-3 the tray I has the smallest f (n) value. The tray I can be implemented in 3 different states
because the empty position can be filled by the members like 7, 8, 6.
Hence, we reached at the goal state after few changes of tiles in different positions of the trays.
Comments:
This problem requires a lot of space for saving the different trays.
Time complexity is more than that of other problems.

The user has to be very careful about the shifting of tiles in the trays.
Very complex puzzle games can be solved by this technique.
Monkey Banana Problem
Definition:
“A monkey is in a room. A bunch of bananas is hanging from the ceiling. The monkey cannot reach the
bananas directly. There is a box in the corner of the room. How can the monkey get the bananas?”
Procedure:
The solution of the problem is of course that the monkey must push the box under the bananas, then stand
on the box and grab the bananas. But the solution procedure requires a lot of planning algorithms. The
purpose of the problem is to raise the question: Are monkeys intelligent? Both humans and monkeys have
the ability to use mental maps to remember things like where to go to find shelter or how to avoid danger.
They can also remember where to go to gather food and water, as well as how to communicate with each
other. Monkeys have the ability not only to remember how to hunt and gather but they also have the
ability to learn new things, as is the case with the monkey and the bananas. Even though that monkey may
never have entered that room before or had only a box for a tool to gather the food available, that monkey
can learn that it needs to move the box across the floor, position it below the bananas and climb the box to
reach for them. Some people believe that this is part instinct, part learned behaviour. It is most probably
both.
Initially, the monkey is at location ‘A’, the banana is at location ‘B’ and the box is at location ‘C’. The
monkey and box have height “low”; but if the monkey climbs onto the box will have height “High”, the
same as the bananas.
The action available to the monkey include:
“GO” from one place to another.
“PUSH” an object from one place to another.
“Climb” onto an object.
“Grasp” an object.
Grasping results in holding the object if the monkey and the object are in the same place at the same
height.
The solution of the problem in different steps can be of followings.

1. What is the initial state description?
At (monkey, A), At (banana, B), At (box, C)
Position (monkey, low), Position (banana, high), Position (box, low)
2. What are the definitions of the different actions?
a) Go (x, y)
Precondition: At (monkey, x)
Effects: ¬At (monkey, x), At (monkey, y)
b) Push (object, x, y, height)
Pre condition: At (monkey, x), At (object, x), Position (monkey, height), Position (object,
height)
Effects: ¬ (monkey, x), ¬At (object, x), At (monkey, y), At(object, y)
c) Climb up (object, y)
Precondition: At (monkey, x), At (object, x), Position (monkey, low), Position (object, low)
Effects: ¬Position (monkey, low), Position (monkey, high), On (monkey, object)
d) Climb down (object)
Preconditions: Position (monkey, high), On (monkey, object)
Effects: ¬Position (monkey, high), ¬ On (monkey, object)
Position: (monkey, low)
e) Grasp (object, x, height)
Preconditions: At (monkey, x), At (object, x), Position (monkey, height), Position (object,
height)
Effect: Hold (object)
f) UnGrasp (object, x, height)
Preconditions: Hold (object), At (monkey, x), At (object, x), Position (monkey, height),
Position (object, height)
Effects: ¬Hold (object)
So the solution to the planning problem may be of following
GO(A,C)
PUSH (Box, C, B, Low)
Climb Up(Box , B)
Grasp(banana, B, High)
Climb down(Box)
Push(Box, B, C, Low)
Comments:
One major application of the monkey banana problem is the toy problem of computer science.

One of the specialized purposes of the problem is to raise the question: Are monkeys intelligent?
This problem is very useful in logic programming and planning.
Tower of Hanoi Problem
Definition:
“We are given a tower of eight discs (initially) four in the applet below, initially stacked in increasing size
on one of three pegs. The objective is to transfer the entire tower to one of the other pegs (the right most
one in the applet below), moving only one disc at a time and never a larger one onto a smaller”.
Procedure:
The tower of Hanoi puzzle was invented by the French mathematician Eduardo Lucas in 1883. The
puzzle is well known to students of computer science since it appears in virtually any introductory text on
data structure and algorithms.
The objective of the puzzle is to move the entire stack to another rod, obeying the following rules.
Only one disc can be moved at a time.
Each move consist of taking the upper disc from one of the rods and sliding it onto another rod,
on top of the other discs that may already be present on that rod.
No disc may be placed on the top of a smaller disk.
There is a legend about a Vietnamese temple which contains a large room with three times. Worn posts in
it surrounded by 64 golden disks. The priests of Hanoi, acting out of command of an ancient prophecy,
have been moving these disks, in accordance with the rules of the puzzle, since that time. The puzzle is
therefore also known as the tower of Brahma puzzle. According to the legend, when the last move of the
puzzle is completed, the world will end.
There are many variations on this legend. For instance, in some tellings, the temple is a monastery and the
priests are monks. The temple or monastery may be said to be in different parts of the world including
Hanoi, Vietnam and may be associated with any religion. The flag tower of Hanoi may have served as the
inspiration for the name.
The puzzle can be played with any number of disks, although many toy versions have around seven to
nine of them. The game seems impossible to many novices yet is solvable with a simple algorithm. The
following solution is a very simple method to solve the tower of Hanoi problem.
Alternative moves between the smallest piece and a non- smallest piece. When moving the
smallest piece, always move it in the same direction (to the right if starting number of pieces is
even, to the left if starting number of pieces is odd).

If there is no tower in the chosen direction, move the pieces to the opposite end, but then
continue to move in the correct direction, for example if you started with three pieces, you
would move the smallest piece to the opposite end, then continue in the left direction after that.
When the turn is to move the non-smallest piece, there is only one legal move.
Doing this should complete the puzzle using the least amount of moves to do so. Finally, the user will
reach at the goal. Also various types of solutions may be possible to solve the tower of Hanoi problem
like recursive procedure, non-recursive procedure and binary solution procedure.
Another simple solution to the problem is given below.
For an even number of disks
Make the legal move between pegs A and B.
Make the legal moves between pages A and C.
Make the legal move between pages B and C.
For an even number of disks
Make the legal move between pegs A and C.
Make the legal move between pegs A and B.
Make the legal move between pegs B and C.
Repeat until complete.
A recursive solution for tower of Hanoi problem is as follows.
A key to solving this problem is to recognize that it can be solve by breaking the problem down into the
collection of smaller problems and further breaking those problems down into even smaller problems
until a solution is reached. The following procedure demonstrates this approach.
Label the pegs A, B, C - these levels may move at different steps.
Let n be the total number of disks.
Number of disks from 1 (smallest, topmost) to n (largest, bottommost).
To move n disks from peg A to peg C.
a) Move n-1 disks from A to B. This leaves disk #n alone on peg A.
b) Move disk #n from A to C.
c) Move n-1 disks from B to C so they sit on disk #n.
To carry out steps a and c, apply the same algorithm again for n-1. The entire procedure is a finite number
of steps, since at most point the algorithm will be required for n = 1. This step, moving a single disc from
peg A to peg B, is trivial.
Comments:

The tower of Hanoi is frequently used in psychological research on problem solving.
This problem is frequently used in neuro-psychological diagnosis and treatment of executive
functions.
The tower of Hanoi is also used as backup rotation scheme when performing computer data
backups where multiple tabs/media are involved.
This problem is very popular for teaching recursive algorithm to beginning programming
students.
A pictorial version of this puzzle is programmed into emacs editor, accessed by typing M - X
Hanoi.
The tower of Hanoi is also used as a test by neuro-psychologists trying to evaluate frontal lobe
deficits.
Cryptarithmatic Problem
Definition:
“It is an arithmetic problem which is represented in letters. It involves the decoding of digit represented
by a character. It is in the form of some arithmetic equation where digits are distinctly represented by
some characters. The problem requires finding of the digit represented by each character. Assign a
decimal digit to each of the letters in such a way that the answer to the problem is correct. If the same
letter occurs more than once, it must be assigned the same digit each time. No two different letters may be
assigned the same digit”.
Procedure:
Cryptarithmatic problem is an interesting constraint satisfaction problem for which different algorithms
have been developed. Cryptarithm is a mathematical puzzle in which digits are replaced by letters of the
alphabet or other symbols. Cryptarithmatic is the science and art of creating and solving cryptarithms.
The different constraints of defining a cryptarithmatic problem are as follows.
1) Each letter or symbol represented only one and a unique digit throughout the problem.
2) When the digits replace letters or symbols, the resultant arithmetical operation must be correct.
The above two constraints lead to some other restrictions in the problem.
For example:

Consider that, the base of the number is 10. Then there must be at most 10 unique symbols or letters in
the problem. Otherwise, it would not possible to assign a unique digit to unique letter or symbol in the
problem. To be semantically meaningful, a number must not begin with a 0. So, the letters at the
beginning of each number should not correspond to 0. Also one can solve the problem by a simple blind
search. But a rule based searching technique can provide the solution in minimum time.
Now, let us solve a simple cryptarithmatic puzzle given below.
Step 1:
In the above problem, M must be 1. You can visualize that, this is an addition problem. The sum of two
four digit numbers cannot be more than 10,000. Also M cannot be zero according to the rules, since it is
the first letter.
So now you have the problem like
Step 2:
Now in the column s10, s+1 ≥ 10. S must be 8 because there is a 1 carried over from the column EON or
9. O must be 0 (if s=8 and there is a 1 carried or s = 9 and there is no 1 carried) or 1 (if s=9 and there is a
1 carried). But 1 is already taken, so O must be 0.

Step 3:
There cannot be carry from column EON because any digit +0 10, unless there is a carry from the
column NRE, and E=9; But this cannot be the case because then N would be 0 and 0 is already taken. So
E 9 and there is no carry from this column. Therefore S=9 because 9+1=10.
Step 4:
In the column EON, E cannot be equal to N. So there must be carry from the column NRE; E+1=N. We
now look at the column NRE, we know that E+1=N. Since we know that carry from this column,
N+R=1E (if there is no carry from the column DEY) or N+R+1=1E (if there is a carry from the column
DEY).
Let us see both the cases:
No carry: N R 10 N 1 N 9
R 9
But 9 is already taken, so this will not work
Carry: N R 1 9
R 9 1 8 This must be solution for R
Step 5:
Now just think what are the digits we have left? They are 7, 6, 5, 4, 3 and 2. We know there must be a
carry from the column DEY. So D E % 10.N E 1, So E cannot be 7 because then N would be 8

which is already taken. D is almost 7, so E cannot be 2 because then D E 10 and E cannot be 3
because then D E 10 and Y 0, but 0 is already taken. Also E cannot be 4 because if D % 6, D
E 10 and if D 6 or D 7 then Y 0 or Y 1, which are both taken. So E is 5 or 6. If E
6, then D 7 and Y 3. So this part will work but look the column N8E. Point that there is a carry from
the column D5Y.N 8 1 16(As there is a carry from this column). But then N=7 and 7 is taken by D
therefore E=5.
Step 6:
Now we have gotten this important digit, it gets much simpler from here. N+8+1=15, N=6
Step 7:
The digits left are 7, 4, 3 and 2. We know there is carry from the column D5Y, so the only pair that works
is D=7 and Y= 2.
Which is final solution of the problem.
Comments:
This problem requires a lot of reasoning.
Time complexity of the problem is more as concerned to the other problems.
This problem can also be solved by the evolutionary approach and mutation operations.
This problem is dependent upon some constraints which are necessary part of the problem.
Various complex problems can also be solved by this technique.
SEARCHING
Problem solving in artificial intelligence may be characterized as a systematic search through a range of
possible actions in order to reach some predefined goal or solution. In AI problem solving by search

algorithms is quite common technique. In the coming age of AI it will have big impact on the
technologies of the robotics and path finding. It is also widely used in travel planning. This chapter
contains the different search algorithms of AI used in various applications. Let us look the concepts for
visualizing the algorithms.
A search algorithm takes a problem as input and returns the solution in the form of an action sequence.
Once the solution is found, the actions it recommends can be carried out. This phase is called as the
execution phase. After formulating a goal and problem to solve the agent cells a search procedure to solve
it. A problem can be defined by 5 components.
a) The initial state: The state from which agent will start.
b) The goal state: The state to be finally reached.
c) The current state: The state at which the agent is present after starting from the initial state.
d) Successor function: It is the description of possible actions and their outcomes.
e) Path cost: It is a function that assigns a numeric cost to each path.
DIFFERENT TYPES OF SEARCHING
the searching algorithms can be various types. When any type of searching is performed, there may some
information about the searching or mayn’t be. Also it is possible that the searching procedure may depend
upon any constraints or rules. However, generally searching can be classified into two types i.e.
uninformed searching and informed searching. Also some other classifications of these searches are given
below in the figure .

Figure
UNINFORMED SEARCH
Breadth First Search (BFS)
Breadth first search is a general technique of traversing a graph. Breadth first search may use more
memory but will always find the shortest path first. In this type of search the state space is represented in
form of a tree. The solution is obtained by traversing through the tree. The nodes of the tree represent the
start value or starting state, various intermediate states and the final state. In this search a queue data
structure is used and it is level by level traversal. Breadth first search expands nodes in order of their
distance from the root. It is a path finding algorithm that is capable of always finding the solution if one
exists. The solution which is found is always the optional solution. This task is completed in a very
memory intensive manner. Each node in the search tree is expanded in a breadth wise at each level.
Concept:
Step 1: Traverse the root node
Step 2: Traverse all neighbours of root node.
Step 3: Traverse all neighbours of neighbours of the root node.
Step 4: This process will continue until we are getting the goal node.
Algorithm:
Step 1: Place the root node inside the queue.
Step 2: If the queue is empty then stops and return failure.
Step 3: If the FRONT node of the queue is a goal node then stop and return success.
Step 4: Remove the FRONT node from the queue. Process it and find all its neighbours that are in ready
state then place them inside the queue in any order.

Step 5: Go to Step 3.
Step 6: Exit.
Implementation:
Let us implement the above algorithm of BFS by taking the following suitable example.
Figure
Consider the graph in which let us take A as the starting node and F as the goal node (*)
Step 1:
Place the root node inside the queue i.e. A
Step 2:
Now the queue is not empty and also the FRONT node i.e. A is not our goal node. So move to step 3.
Step 3:
So remove the FRONT node from the queue i.e. A and find the neighbour of A i.e. B and C
B C A
Step 4:
Now b is the FRONT node of the queue .So process B and finds the neighbours of B i.e. D.
C D B
A

Step 5:
Now find out the neighbours of C i.e. E
D E C
Step 6:
Next find out the neighbours of D as D is the FRONT node of the queue
E F D
Step 7:
Now E is the front node of the queue. So the neighbour of E is F which is our goal node.
F E
Step 8:
Finally F is our goal node which is the FRONT of the queue. So exit.
F
Advantages:
In this procedure at any way it will find the goal.
It does not follow a single unfruitful path for a long time.
It finds the minimal solution in case of multiple paths.
Disadvantages:
BFS consumes large memory space.
Its time complexity is more.
It has long pathways, when all paths to a destination are on approximately the same search depth.
Depth First Search (DFS)
DFS is also an important type of uniform search. DFS visits all the vertices in the graph. This type of
algorithm always chooses to go deeper into the graph. After DFS visited all the reachable vertices from a
particular sources vertices it chooses one of the remaining undiscovered vertices and continues the search.
DFS reminds the space limitation of breath first search by always generating next a child of the deepest
unexpanded nodded. The data structure stack or last in first out (LIFO) is used for DFS. One interesting

property of DFS is that, the discover and finish time of each vertex from a parenthesis structure. If we use
one open parenthesis when a vertex is finished then the result is properly nested set of parenthesis.
Concept:
Step 1: Traverse the root node.
Step 2: Traverse any neighbour of the root node.
Step 3: Traverse any neighbour of neighbour of the root node.
Algorithm:
Step 1: PUSH the starting node into the stack.
Step 2: If the stack is empty then stop and return failure.
Step 3: If the top node of the stack is the goal node, then stop and return success.
Step 4: Else POP the top node from the stack and process it. Find all its neighbours that are in ready state
and PUSH them into the stack in any order.
Step 5: Go to step 3.
Step 6: Exit.
Implementation:
Let us take an example for implementing the above DFS algorithm.
Figure Examples of DFS

Consider A as the root node and L as the goal node in the graph figure
Step 1: PUSH the starting node into the stack i.e.
Step 2: Now the stack is not empty and A is not our goal node. Hence move to next step.
Step 3: POP the top node from the stack i.e. A and find the neighbours of A i.e. B and C.
B C A
Step 4: Now C is top node of the stack. Find its neighbours i.e. F and G.
B F G C
Step 5: Now G is the top node of the stack. Find its neighbour i.e. M
B F M G
Step 6: Now M is the top node and find its neighbour, but there is no neighbours of M in the graph so
POP it from the stack.
B F M
Step 7: Now F is the top node and its neighbours are K and L. so PUSH them on to the stack.
B K L F
Step 8: Now L is the top node of the stack, which is our goal node.
B K L
Also you can traverse the graph starting from the root A and then insert in the order C and B into the
stack. Check your answer.
Advantages:
DFSconsumes very less memory space.
It will reach at the goal node in a less time period than BFS if it traverses in a right path.
It may find a solution without examining much of search because we may get the desired solution
in the very first go.
A

Disadvantages:
It is possible that may states keep reoccurring.
There is no guarantee of finding the goal node.
Sometimes the states may also enter into infinite loops.
Brute Force or Blind Search
Brute force or blind search is a uniformed exploration of the search space and it does not explicitly take
into account either planning efficiency or execution efficiency. Blind search is also called uniform search.
It is the search which has no information about its domain. The only thing that a blind search can do is to
differentiate between a non goal state and a goal state. These methods do not need domain knowledge but
they are less efficient in result. Uniform strategies don’t use any information about how a close a node
might be to a goal. They differ in the order that the nodes are expanded. The most important brute force
techniques are breadth first search, depth first search, uniform search and bidirectional search. All brute
force techniques must take ( b0
time and use o (d) space. This technique is not as efficient as compared
to other algorithms.
Difference between BFS and DFS
BFS
It uses the data structure queue.
BFS is complete because it finds the solution if one exists.
BFS takes more space i.e. equivalent to o (b0
) where b is the maximum breath exist in a search
tree and d is the maximum depth exit in a search tree.
In case of several goals, it finds the best one.
DFS
It uses the data structure stack.
It is not complete because it may take infinite loop to reach at the goal node.
The space complexity is O (d).
In case of several goals, it will terminate the solution in any order.
Greedy Search
This algorithm uses an approach which is quite similar to the best first search algorithm. It is a simple
best first search which reduces the estimated cost of reach the goal. Basically it takes the closest node that
appears to be closest to the goal. This search starts with the initial matrix and makes very single possible
changes then looks at the change it made to the score. This search then applies the change till the greatest
improvement. The search continues until no further improvement can be made. The greedy search never

makes never makes a lateral move .It uses minimal estimated cost h (n) to the goal state as measure which
decreases the search time but the algorithm is neither complete nor optimal. The main advantage of this
search is that it is simple and finds solution quickly. The disadvantages are that it is not optimal,
susceptible to false start.
Figure Greedy Search
INFORMED SEARCH (HEURISTIC SEARCH)
Heuristic is a technique which makes our search algorithm more efficient. Some heuristics help to guide a
search process without sacrificing any claim to completeness and some sacrificing it. Heuristic is a
problem specific knowledge that decreases expected search efforts. It is a technique which sometimes
works but not always. Heuristic search algorithm uses information about the problem to help directing the
path through the search space. These searches uses some functions that estimate the cost from the current
state to the goal presuming that such function is efficient. A heuristic function is a function that maps
from problem state descriptions to measure of desirability usually represented as number. The purpose of
heuristic function is to guide the search process in the most profitable directions by suggesting which path
to follow first when more than is available.
Generally heuristic incorporates domain knowledge to improve efficiency over blind search. In AI
heuristic has a general meaning and also a more specialized technical meaning. Generally a term heuristic
is used for any advice that is effective but is not guaranteed to work in every case. For example in case of

travelling sales man (TSP) problem we are using a heuristic to calculate the nearest neighbour. Heuristic
is a method that provides a better guess about the correct choice to make at any junction that would be
achieved by random guessing. This technique is useful in solving though problems which could not be
solved in any other way. Solutions take an infinite time to compute.
Let us see some classifications of heuristic search.
Best First Search
Best first search is an instance of graph search algorithm in which a node is selected for expansion based
o evaluation function f (n). Traditionally, the node which is the lowest evaluation is selected for the
explanation because the evaluation measures distance to the goal. Best first search can be implemented
within general search frame work via a priority queue, a data structure that will maintain the fringe in
ascending order of f values. This search algorithm serves as combination of depth first and breadth first
search algorithm. Best first search algorithm is often referred greedy algorithm this is because they
quickly attack the most desirable path as soon as its heuristic weight becomes the most desirable.
Concept:
Step 1: Traverse the root node
Step 2: Traverse any neighbour of the root node, that is maintaining a least distance from the root node
and insert them in ascending order into the queue.
Step 3: Traverse any neighbour of neighbour of the root node, that is maintaining a least distance from
the root node and insert them in ascending order into the queue
Step 4: This process will continue until we are getting the goal node
Algorithm:
Step 1: Place the starting node or root node into the queue.
Step 2: If the queue is empty, then stop and return failure.
Step 3: If the first element of the queue is our goal node, then stop and return success.
Step 4: Else, remove the first element from the queue. Expand it and compute the estimated goal distance
for each child. Place the children in the queue in ascending order to the goal distance.
Step 5: Go to step-3
Step 6: Exit.

Implementation:
Let us solve an example for implementing above BFS algorithm.
Figure
Step 1:
Consider the node A as our root node. So the first element of the queue is A whish is not our goal node,
so remove it from the queue and find its neighbour that are to inserted in ascending order.
A
Step 2:
The neighbours of A are B and C. They will be inserted into the queue in ascending order.
B C A
Step 3:
Now B is on the FRONT end of the queue. So calculate the neighbours of B that are maintaining a least
distance from the roof.
F E D C B
Step 4:
Now the node F is on the FRONT end of the queue. But as it has no further children, so remove it from
the queue and proceed further.

E D C F
Step 5:
Now E is the FRONT end. So the children of E are J and K. Insert them into the queue in ascending order.
K J D C E
Step 6:
Now K is on the FRONT end and as it has no further children, so remove it and proceed further
J D C K
Step7:
Also, J has no corresponding children. So remove it and proceed further.
D C J
Step 8:
Now D is on the FRONT end and calculates the children of D and put it into the queue.
I C D
Step9:
Now I is the FRONT node and it has no children. So proceed further after removing this node from the
queue.
C I
Step 10:
Now C is the FRONT node .So calculate the neighbours of C that are to be inserted in ascending order
into the queue.
G H C

Step 11:
Now remove G from the queue and calculate its neighbour that is to insert in ascending order into the
queue.
M L H G
Step12:
Now M is the FRONT node of the queue which is our goal node. So stop here and exit.
L H M
Advantage:
It is more efficient than that of BFS and DFS.
Time complexity of Best first search is much less than Breadth first search.
The Best first search allows us to switch between paths by gaining the benefits of both breadth
first and depth first search. Because, depth first is good because a solution can be found without
computing all nodes and Breadth first search is good because it does not get trapped in dead
ends.
Disadvantages:
Sometimes, it covers more distance than our consideration.
Branch and Bound Search
Branch and Bound is an algorithmic technique which finds the optimal solution by keeping the best
solution found so far. If partial solution can’t improve on the best it is abandoned, by this method the
number of nodes which are explored can also be reduced. It also deals with the optimization problems
over a search that can be presented as the leaves of the search tree. The usual technique for eliminating
the sub trees from the search tree is called pruning. For Branch and Bound algorithm we will use stack
data structure.
Concept:
Step 1: Traverse the root node.
Step 2: Traverse any neighbour of the root node that is maintaining least distance from the root node.
Step 3: Traverse any neighbour of the neighbour of the root node that is maintaining least distance from
the root node.

Algorithm:
Step 1: PUSH the root node into the stack.
Step 2: If stack is empty, then stop and return failure.
Step 3: If the top node of the stack is a goal node, then stop and return success.
Step 4: Else POP the node from the stack. Process it and find all its successors. Find out the path
containing all its successors as well as predecessors and then PUSH the successors which are belonging to
the minimum or shortest path.
Step 5: Go to step 5.
Step 6: Exit.
Implementation:
Let us take the following example for implementing the Branch and Bound algorithm.
Figure
Step 1:
Consider the node A as our root node. Find its successors i.e. B, C, F. Calculate the distance from the root
and PUSH them according to least distance.
A

B: 0+5 = 5 (The cost of A is 0 as it is the starting node)
F: 0+9 = 9
C: 0+7 = 7
Here B (5) is the least distance.
5
Step 2:
Now the stack will be
C F B A
As B is on the top of the stack so calculate the neighbours of B.
D: 0+5+4 = 9
E: 0+5+6 = 11
The least distance is D from B. So it will be on the top of the stack.
5
4
Step 3:
A
B
A
B
C

As the top of the stack is D. So calculate neighbours of D.
C F D B
C: 0+5+4+8 = 17
F: 0+5+4+3 = 12
The least distance is F from D and it is our goal node. So stop and return success.
Step 4:
C F D
Hence the searching path will be A-B -D-F
Advantages:
As it finds the minimum path instead of finding the minimum successor so there should not be
any repetition.
The time complexity is less compared to other algorithms.
Disadvantages:
The load balancing aspects for Branch and Bound algorithm make it parallelization difficult.
The Branch and Bound algorithm is limited to small size network. In the problem of large
networks, where the solution search space grows exponentially with the scale of the network, the
approach becomes relatively prohibitive.
A* SEARCH
A* is a cornerstone name of many AI systems and has been used since it was developed in 1968 by Peter
Hart; Nils Nilsson and Bertram Raphael. It is the combination of Dijkstra’s algorithm and Best first
search. It can be used to solve many kinds of problems. A* search finds the shortest path through a search
space to goal state using heuristic function. This technique finds minimal cost solutions and is directed to
a goal state called A* search. In A*, the * is written for optimality purpose. The A* algorithm also finds
the lowest cost path between the start and goal state, where changing from one state to another requires
some cost. A* requires heuristic function to evaluate the cost of path that passes through the particular
state. This algorithm is complete if the branching factor is finite and every action has fixed cost. A*
requires heuristic function to evaluate the cost of path that passes through the particular state. It can be
defined by following formula.

f n g n h n
Where
g n: The actual cost path from the start state to the current state.
h n: The actual cost path from the current state to goal state.
f n: The actual cost path from the start state to the goal state.
For the implementation of A* algorithm we will use two arrays namely OPEN and CLOSE.
OPEN:
An array which contains the nodes that has been generated but has not been yet examined.
CLOSE:
An array which contains the nodes that have been examined.
Algorithm:
Step 1: Place the starting node into OPEN and find its f (n) value.
Step 2: Remove the node from OPEN, having smallest f (n) value. If it is a goal node then stop and return
success.
Step 3: Else remove the node from OPEN, find all its successors.
Step 4: Find the f (n) value of all successors; place them into OPEN and place the removed node into
CLOSE.
Step 5: Go to Step-2.
Step 6: Exit.
Implementation:
The implementation of A* algorithm is 8-puzzle game.
Advantages:

It is complete and optimal.
It is the best one from other techniques.
It is used to solve very complex problems.
It is optimally efficient, i.e. there is no other optimal algorithm guaranteed to expand fewer nodes
than A*.
Disadvantages:
This algorithm is complete if the branching factor is finite and every action has fixed cost.
The speed execution of A* search is highly dependant on the accuracy of the heuristic algorithm
that is used to compute h (n).
It has complexity problems.
AO* Search: (And-Or) Graph
The Depth first search and Breadth first search given earlier for OR trees or graphs can be easily adopted
by AND-OR graph. The main difference lies in the way termination conditions are determined, since all
goals following an AND nodes must be realized; where as a single goal node following an OR node will
do. So for this purpose we are using AO* algorithm.
Like A* algorithm here we will use two arrays and one heuristic function.
OPEN:
It contains the nodes that has been traversed but yet not been marked solvable or unsolvable.
CLOSE:
It contains the nodes that have already been processed.
6 7:The distance from current node to goal node.
Algorithm:
Step 1: Place the starting node into OPEN.
Step 2: Compute the most promising solution tree say T0.
Step 3: Select a node n that is both on OPEN and a member of T0. Remove it from OPEN and place it in
CLOSE

Step 4: If n is the terminal goal node then leveled n as solved and leveled all the ancestors of n as solved.
If the starting node is marked as solved then success and exit.
Step 5: If n is not a solvable node, then mark n as unsolvable. If starting node is marked as unsolvable,
then return failure and exit.
Step 6: Expand n. Find all its successors and find their h (n) value, push them into OPEN.
Step 7: Return to Step 2.
Step 8: Exit.
Implementation:
Let us take the following example to implement the AO* algorithm.
Figure
Step 1:
In the above graph, the solvable nodes are A, B, C, D, E, F and the unsolvable nodes are G, H. Take A as
the starting node. So place A into OPEN.
i.e. OPEN = CLOSE = (NULL)
Step 2:
A φ A

The children of A are B and C which are solvable. So place them into OPEN and place A into the
CLOSE.
i.e. OPEN = CLOSE =
Step 3:
Now process the nodes B and C. The children of B and C are to be placed into OPEN. Also remove B and
C from OPEN and place them into CLOSE.
So OPEN = CLOSE =
(O)
(O)
‘O’ indicated that the nodes G and H are unsolvable.
Step 4:
As the nodes G and H are unsolvable, so place them into CLOSE directly and process the nodes D and E.
i.e. OPEN = CLOSE =
*
Step 5:
Now we have been reached at our goal state. So place F into CLOSE.
A
B C
A
B C
G D E
H
A B C
B
G D
C
E H
A D E H(O)
A B C G (O)
F
D E
D E H
(O)
A B C G
(O) F

A
B C
F
D E
i.e. CLOSE =
Step 6:
Success and Exit
AO* Graph:
Figure
Advantages:
It is an optimal algorithm.
If traverse according to the ordering of nodes.
It can be used for both OR and AND graph.
Disadvantages:
Sometimes for unsolvable nodes, it can’t find the optimal path.
Its complexity is than other algorithms.
Hill Climbing
Hill climbing search algorithm is simply a loop that continuously moves in the direction of increasing
value. It stops when it reaches a “peak” where no neighbour has higher value. This algorithm is
considered to be one of the simplest procedures for implementing heuristic search. The hill climbing
comes from that idea if you are trying to find the top of the hill and you go up direction from where ever
you are. This heuristic combines the advantages of both depth first and breadth first searches into a single
method.
The name hill climbing is derived from simulating the situation of a person climbing the hill. The person
will try to move forward in the direction of at the top of the hill. His movement stops when it reaches at

the peak of hill and no peak has higher value of heuristic function than this. Hill climbing uses knowledge
about the local terrain, providing a very useful and effective heuristic for eliminating much of the
unproductive search space. It is a branch by a local evaluation function. The hill climbing is a variant of
generate and test in which direction the search should proceed. At each point in the search path, a
successor node that appears to reach for exploration.
Algorithm:
Step 1: Evaluate the starting state. If it is a goal state then stop and return success.
Step 2: Else, continue with the starting state as considering it as a current state.
Step 3: Continue step-4 until a solution is found i.e. until there are no new states left to be applied in the
current state.
Step 4:
a) Select a state that has not been yet applied to the current state and apply it to produce a new state.
b) Procedure to evaluate a new state.
i. If the current state is a goal state, then stop and return success.
ii. If it is better than the current state, then make it current state and proceed further.
iii. If it is not better than the current state, then continue in the loop until a solution is found.
Step 5: Exit.
Advantages:
Hill climbing technique is useful in job shop scheduling, automatic programming, circuit
designing, and vehicle routing and portfolio management.
It is also helpful to solve pure optimization problems where the objective is to find the best state
according to the objective function.
It requires much less conditions than other search techniques.
Disadvantages:
The question that remains on hill climbing search is whether this hill is the highest hill possible.
Unfortunately without further extensive exploration, this question cannot be answered. This technique
works but as it uses local information that’s why it can be fooled. The algorithm doesn’t maintain a search
tree, so the current node data structure need only record the state and its objective function value. It
assumes that local improvement will lead to global improvement.
There are some reasons by which hill climbing often gets suck which are stated below.

Local Maxima:
A local maxima is a state that is better than each of its neighbouring states, but not better than some other
states further away. Generally this state is lower than the global maximum. At this point, one cannot
decide easily to move in which direction! This difficulties can be extracted by the process of backtracking
i.e. backtrack to any of one earlier node position and try to go on a different event direction. To
implement this strategy, maintaining in a list of path almost taken and go back to one of them. If the path
was taken that leads to a dead end, then go back to one of them.
Figure Local Maxima
Ridges:
It is a special type of local maxima. It is a simply an area of search space. Ridges result in a sequence of
local maxima that is very difficult to implement ridge itself has a slope which is difficult to traverse. In
this type of situation apply two or more rules before doing the test. This will correspond to move in
several directions at once.
Figure Ridges
Plateau:

It is a flat area of search space in which the neighbouringhave same value. So it is very difficult to
calculate the best direction. So to get out of this situation, make a big jump in any direction, which will
help to move in a new direction this is the best way to handle the problem like plateau.
Figure Plateau
KNOWLEDGE
Knowledge is the collection of facts, inference rules etc. which can be used for a particular purpose.
Knowledge requires the use of data and information. It combines relationships, correlations, dependencies
with data and information.
The basic components of knowledge are:
1) A set of collected data
2) A form of belief or hypothesis
3) A kind of information.
Knowledge is different from data. Data is the collection of raw materials where as knowledge is the
collection of some well specified inference rules and facts. Knowledge is also different from belief and
hypothesis. Belief is any meaningful and coherent expression that can be represented. Belief may be true
or false. A hypothesis is a justified belief that is not known to be true. A hypothesis is a belief which is
backed up with some supporting evidence but it may still be false. So knowledge can be defined as true
justified knowledge.
KNOWLEDGE BASED SYSTEMS
Knowledge based systems get their power from the expert knowledge that has been coded into facts,
rules, heuristics and procedures. The knowledge is stored in a knowledge base separate from the control
and inferencing components. Knowledge is important and essential for knowledge based intelligent
behaviour.
Input, Output Unit
Inference Control
Unit
Knowledge Base

Figure A typical Knowledge based system
Any choice of representation will depend on the type of problem to be solved and the inference methods
available. Knowledge may be vague, contradictory or incomplete. Thus, knowledge is information about
objects, concepts and relationships that are assumed to exist in a particular area of interest.
TYPE OF KNOWLEDGE
The categorisation of knowledge is very much large and interesting. They can be of following types:
Declarative knowledge
It is the passive knowledge expressed as statements of facts about the world. It gives the simple facts and
ideas about any phenomenon. It means just the representation of facts or assertions. This tells the total
description about the situation. For example, the facts about an organization may be its buildings,
location, no. of departments, no. of employees etc. The facts may be of two types i.e. static and dynamic.
The static facts do not change with time where as the dynamic facts change with time. For example, the
name and location of an organization is permanent. But some additional departments may be added.
Procedural knowledge
Procedural knowledge is the compiled knowledge related to the performance of some task. For example
the steps used to solve an algebric equation can be expressed as procedural knowledge. It also eradicates
the limitations of declarative knowledge i.e. declarative knowledge tells about the organization but it
cannot tell how the employees are working in that organization and how the products are developed. But
procedural knowledge describes everything about the organization by using production rules and dynamic
attributes.
For example, If: All the employees are very hardworking
They are very punctual
They have productive ideas.
Then: Large no. of products can be produced within a very limited time period.
The advantages of using procedural knowledge are as follows:
1) Domain specific knowledge can be easily represented.
2) Extended logical inferences, such as default reasoning facilitated.

3) Side effects of actions may be modeled.
Some disadvantages of procedural knowledge are
1) Completeness: In procedural knowledge not all cases may be represented.
2) Consistency: Not all deductions may be correct.
3) Modularity: Changes in knowledge base might have far-reaching effects.
Inheritable knowledge
There are many situations in the world, where the object of an event inherits some properties of that
particular event or any other event.
For example, consider a college. A college has certain features like classrooms, teachers, play ground,
furniture, students etc. Besides these, there will be some general concepts regarding the functioning of the
college, like it will have time table for each class, a fee deposit plan, examination pattern, course module
etc. It can have many more deep concepts like placement of students etc. Now, if we say “A is a College”,
then A will automatically inherits all the features of the college. It may be possible that X has some
additional features. The inheritable knowledge is diagrammatically represented below. Here, the
relationship ‘has’ indicates the silent features or attributes and ‘is a’ represents the variable or instance of
that type. A inherits all the properties of college and has one additional feature of having male students. In
this type of knowledge, data must be organized into a hierarchy of classes. The arrows represent the point
from object to its value in the diagram. Boxed nodes represent the objects and values of attributes of
objects.
Figure College Attribute representation

Relational Knowledge
Relational knowledge is made up of objects consisting of attributes and corresponding associated values.
In this type of knowledge, the facts are represented as set of relations in a tabular form. The table stores or
captures all the hidden attributes of objects.
For example the knowledge about doctors may be as mentioned in figure .
Department Qualification Height Age
Eye P.HD 5.0 35
Kidney P.HD 5.10 32
Surgery P.HD 6.3 28
Medicine P.HD 6.1 44
Figure Knowledge about Doctor
This form of representation is the simplest and can be used in database systems. But this representation
cannot store any semantic, information. For example, from this information we cannot answer the
questions like “What is the name of the doctor”? or “How many doctors are in eye department”?
Inferential Knowledge
The knowledge, which can use inference mechanism to use this knowledge is called inferential
knowledge. The inheritance property is a very powerful form of inferential knowledge. The inference
procedures implement the standard logic rules of inference. There are two types of inference procedures
like forward inference and backward inference. Forward inference moves from start state to goal state
whereas backward inference moves from goal state to start state. In this type of knowledge several
symbols are generally used like ∀ (universal quantifier), ∃(existential quantifier), → (arrow indicator) etc.
For example: All cats have tails
∀ X: cat (x) → has tail (x)
Advantages:
1) A set of strict rules are defined which can be used to derive more facts.
2) Truths of new statements can be verified.
3) It gives guarantee about the correctness.
4) Many inference procedures available to implement standard rules of logic.

Heuristic Knowledge
This type of knowledge is fully experimental. This knowledge requires some judgments about any
performance. One can guess a good thing and also one can think bad thing. But good performances are
generally taken in heuristic knowledge. For example, suppose it is asked that “Ram will score how much
percentage in his final semester?” Then the answer might be 80%, 70%, 30% or 95%. The individual
answers of this question based on the heuristic knowledge. The answer would be based on various factors
such as past performance, his talent etc. If his previous semester percentage was 78%, then if one will say
he will secure 10% in this semester then obviously he has not any knowledge about Ram.
Tacit Knowledge
This kind of knowledge is acquired by experience. Tacit knowledge is subconsciously understood and
applied, difficult to articulate and formalize. This type of knowledge is developed from direct experience
and action. This knowledge is usually shared through highly interactive conversation, story telling and
experience. It also includes cognitive skills such as intuition as well as technical skills such as craft and
know-how. Tacit knowledge cannot be transmitted before it is converted into words, models or numbers
that can be understood. Tacit knowledge can be defined in two dimensions, such as technical dimension
and cognitive dimension. In technical dimension highly subjective and personal insights, intuitions and
inspirations derived from long experience. The dimensions such as beliefs, ideals, principles, values and
emotions fall in the category of cognitive dimension.
Explicit Knowledge
This knowledge is formalized, coded in several natural languages (English, Italian and Spanish) or
artificial languages (UML, Mathematics etc). This knowledge can be easily transmitted. It includes
theoretical approaches, problem solving, manuals and database. As explicit knowledge, it was the first to
be or, at least, to be archived. Tacit and explicit knowledge are not totally separate, but mutually
complementary entities. Without any experience, we cannot truly understand. Explicit knowledge is
playing an increasingly large role in organization and it is considered by some to be the most important
factor of production in the knowledge economy. Imagine an organization without procedure manuals
product literature or computer software. Also with explicit knowledge, some tacit knowledge is required
to run the business in an organization. Without explicit knowledge, the organization is simply has a zero
performance.
Research Knowledge
There are many standards for the generation and critical appraisal of research knowledge, but judging the
quality of knowledge in this source is not without difficulty. There are disputes about the nature and
content of standards in areas such as qualitative research, and the implementation of standards is

sometimes weak so that conformity with them is not necessarily a guarantee of quality. This type of
knowledge is very useful for researchers to improve the research quality.
KNOWLEDGE ACQUISITION
Knowledge acquisition is the gathering or collecting knowledge from various sources. It is the process of
adding new knowledge to a knowledge base and refining or improving knowledge that was previously
acquired. Acquisition is the process of expanding the capabilities of a system or improving its
performance at some specified task. So it is the goal oriented creation and refinement of knowledge.
Acquired knowledge may consist of facts, rules, concepts, procedures, heuristics, formulas, relationships,
statistics or any other useful information. Source of these knowledges may be experts in the domain of
interest, text books, technical papers, database reports, journals and the environments. The knowledge
acquisition is a continuous process and is spread over entire lifetime. Example of knowledge acquisition
is machine learning. It may be process of autonomous knowledge creation or refinements through the use
of computer programs. The newly acquired knowledge should be integrated with existing knowledge in
some meaningful way. The knowledge should be accurate, non-redundant, consistent and fairly complete.
Knowledge acquisition supports the activities like entering the knowledge and maintaining knowledge
base. The knowledge acquisition process also sets dynamic data structures for existing knowledge to
refine the knowledge.
The role of knowledge engineer is also very important with respect to develop the refinements of
knowledge. Knowledge engineers may be the professionals who elicit knowledge from experts. They
integrate knowledge from various sources like creates and edits code, operates the various interactive
tools, build the knowledge base etc.

Figure Knowledge Engineer’s Roles in Interactive Knowledge Acquisition
Knowledge Acquisition Techniques
Many techniques have been developed to deduce knowledge from an expert. They are termed as
knowledge acquisition techniques. They are:
a) Diagram Based Techniques
b) Matrix Based Techniques
c) Hierarchy-Generation Techniques
d) Protocol Analysis Techniques
e) Protocol Generation Techniques
f) Sorting Techniques
In diagram based techniques the generation and use of concept maps, event diagrams and process maps.
This technique captures the features like “why, when, who, how and where”. The matrix based techniques
involve the construction of grids indicating such things as problems encountered against possible

solutions. Hierarchical techniques are used to build hierarchical structures like trees. Protocol analysis
technique is used to identify the type of knowledge like goals, decisions, relationships etc. The protocol
generation techniques include various types of interviews like structured, semi-structured and
unstructured.
The most common knowledge acquisition technique is face-to-face interview. Interview is a very
important technique which must be planned carefully. The results of an interview must be verified and
validated. Some common variations of an unstructured interview are talk through, teach through and read
through. The knowledge engineer slowly learns about the problem. Then can build a representation of the
knowledge. In unstructured interviews, seldom provides complete or well-organized descriptions of
cognitive processes because the domains are generally complex. The experts usually find it very difficult
to express some more important knowledge. Data acquired are often unrelated, exists at varying levels of
complexity, and are difficult for the knowledge engineer to review, interpret and integrate. But on the
other hand structured interviews are systematic goal oriented process. It forces an organized
communication between the knowledge engineer and the expert. In structured interview, inter personal
communication and analytical skills are important.
KNOWLEDGE REPRESENTATION
Knowledge representation is probably, the most important ingredient for developing an AI. A
representation is a layer between information accessible from outside world and high level thinking
processes. Without knowledge representation it is impossible to identify what thinking processes are,
mainly because representation itself is a substratum for a thought.
The subject of knowledge representation has been messaged for a couple of decades already. For many
applications, specific domain knowledge is required. Instead of coding such knowledge into a system in a
way that it can never be changed (hidden in the overall implementation), more flexible ways of
representing knowledge and reasoning about it have been developed in the last 10 years.
The need of knowledge representation was felt as early as the idea to develop intelligent systems. With
the hope that readers are well conversant with the fact by now, that intelligent requires possession of
knowledge and that knowledge is acquired by us by various means and stored in the memory using some
representation techniques. Putting in another way, knowledge representation is one of the many critical
aspects, which are required for making a computer behave intelligently. Knowledge representation refers
to the data structures techniques and organizing notations that are used in AI. These include semantic
networks, frames, logic, production rules and conceptual graphs.
Properties for knowledge Representation
The following properties should be possessed by a knowledge representation system.

a. Representational Adequacy: It is the ability to represent the required knowledge.
b. Inferential Adequacy: It is the ability to manipulate the knowledge represented to produce new
knowledge corresponding to that inferred from the original.
c. Inferential Efficiency: The ability to direct the inferential mechanisms into the most productive
directions by storing appropriate guides.
d. Acquisitional Efficiency: The ability to acquire new knowledge using automatic methods
wherever possible rather than reliance on human intervention.
Syntax and semantics for Knowledge Representation
Knowledge representation languages should have precise syntax and semantics. You must know exactly
what an expression means in terms of objects in the real world. Suppose we have decided that “red 1”
refers to a dark red colour, “car1” is my car, car2 is another. Syntax of language will tell you which of the
following is legal: red1 (car1), red1 car1, car1 (red1), red1 (car1 car2)?
Semantics of language tell you exactly what an expression means: for example, Pred (Arg) means that the
property referred to by Pred applies to the object referred to by Arg. E.g., properly “dark red” applies to
my car.
Types of Knowledge Representation
Knowledge can be represented in different ways. The structuring of knowledge and how designers might
view it, as well as the type of structures used internally are considered. Different knowledge
representation techniques are
a. Logic
b. Semantic Network
c. Frame

d. Conceptual Graphs
e. Conceptual Dependency
f. Script
Logic
A logic is a formal language, with precisely defined syntax and semantics, which supports sound
inference. Different logics exist, which allow you to represent different kinds of things, and which allow
more or less efficient inference. The logic may be different types like propositional logic, predicate logic,
temporal logic, description logic etc. But representing something in logic may not be very natural and
inferences may not be efficient.
Figure
Semantic Network
A semantic network is a graphical knowledge representation technique. This knowledge representation
system is primarily on network structure. The semantic networks were basically developed to model
human memory. A semantic net consists of nodes connected by arcs. The arcs are defined in a variety of
ways, depending upon the kind of knowledge being represented.
The main idea behind semantic net is that the meaning of a concept comes, from the ways in which it is
connected to other concepts. The semantic network consists of different nodes and arcs. Each node should
contain the information about objects and each arc should contain the relationship between objects.
Semantic nets are used to find relationships among objects by spreading activation about from each of
two nodes and seeing where the activation met this process is called intersection search.
For example: Ram is a boy.

Figure
Semantic network by using Instances
The semantic network based knowledge representation mechanism is useful where an object or concept is
associated with many attributes and where relationships between objects are important. Semantic nets
have also been used in natural language research to represent complex sentences expressed in English.
The semantic representation is useful because it provides a standard way of analyzing the meaning of
sentence. It is a natural way to represent relationships that would appear as ground instances of binary
predicates in predicate logic. In this case we can create one instance of each object. In instance based
semantic net representations some keywords are used like: IS A, INSTANCE, AGENT, HAS-PARTS etc.
Consider the following examples:
1. Suppose we have to represent the sentence “Sima is a girl”.
Figure
2. Ram is taller than Hari

It can also be represented as
(b)
3. “Mouse is a Rodent and Rodent is a mammal. Mouse has teeth and etas grass”. Check whether the
sentence mammal has teeth is valid or not. ]
(c)
Partitioned Semantic Network
Some complex sentences are there which cannot be represented by simple semantic nets and for this we
have to follow the technique partitioned semantic networks. Partitioned semantic net allow for
1. Propositions to be made without commitment to truth.
2. Expressions to be quantified.

In partitioned semantic network, the network is broken into spaces which consist of groups of nodes and
arcs and regard each space as a node.
Let us consider few examples.
Draw the partitioned semantic network structure for the followings:
a) Sima is eating an apple.
Figure
b) All Sima are eating an apple.
Figure
c) All Sima are eating some apple.

Figure
d) All men are mortal
Figure
e) Every dog has bitten a shopkeeper

Figure
f) Every dog in town has bitten a shopkeeper.
Figure
NOTE: On the above semantic network structures, the instance “IS A” is used. Also two terms like
assailant and victim are used. Assailant means “by which the work is done” and that of victim refers to
“on which the work is applied”. Another term namely GS, which refers to General Statement. For GS,
make a node g which is an instance of Gs. Every element will have at least two attributes. Firstly, a form
that states which a relation is being asserted. Secondly, one or more for all (∀) or there exists (∃)
connections which represent universally quantifiable variables.
FRAME
A frame is a collection of attributes and associated values that describe some entity in the world. Frames
are general record like structures which consist of a collection of slots and slot values. The slots may be of
any size and type. Slots typically have names and values or subfields called facets. Facets may also have
names and any number of values. A frame may have any number of slots, a slot may have any number of
facets, each with any number of values. A slot contains information such as attribute value pairs, default
values, condition for filling a slot, pointers to other related frames and procedures that are activated when
needed for different purposes. Sometimes a frame describes an entity in some absolute sense, sometimes
it represents the entity from a particular point of view. A single frame taken alone is rarely useful. We
build frame systems out of collection of frames that are connected to each other by virtue of the fact that
the value of an attribute of one frame may be another frame. Each frame should start with an open
parenthesis and closed with a closed parenthesis.
Syntax of a frame

Let us consider the below examples.
1) Create a frame of the person Ram who is a doctor. He is of 40. His wife name is Sita. They
have two children Babu and Gita. They live in 100 kps street in the city of Delhi in India. The
zip code is 756005.
(Ram
(PROFESSION (VALUE Doctor))
(AGE (VALUE 40))
(WIFE (VALUE Sita))
(CHILDREN (VALUE Bubu, Gita))
(ADDRESS
(STREET (VALUE 100 kps))
(CITY(VALUE Delhi))
(COUNTRY(VALUE India))
(ZIP (VALUE 756005))))
2) Create a frame of the person Anand who is a chemistry professor in RD Women’s College.
His wife name is Sangita having two children Rupa and Shipa.

(Anand
(PROFESSION (VALUE Chemistry Professor))
(ADDRESS (VALUE RD Women’s College))
(WIFE (VALUE Sangita))
(CHILDREN(VALUE RupaShipa)))
3) Create a frame of the person Akash who has a white maruti car of LX-400 Model. It has 5
doors. Its weight is 225kg, capacity is 8, and mileage is 15 km /lit.
(Akash
(CAR (VALUE Maruti))
(COLOUR (VALUE White))
(MODEL (VALUE LX-400))
(DOOR (VALUE 5))
(WEIGHT (VALUE 225kg))
(CAPACITY (VALUE 8))
(MILAGE (VALUE 15km/lit)))
The frames can be attached with another frame and can create a network of frames. The main task of
action frame is to provide the facility for procedural attachment and help in reasoning process. Reasoning
using frames is done by instantiation. Instantiation process begins, when the given situation is matched
with frames that are already in existence. The reasoning process tries to match the current problem state
with the frame slot and assigns them values.The valuesassigned to the slots depict a particular situation
and by this, the reasoning process moves towards a goal. The reasoning process can be defined as filling
slot values in frames.
Conceptual Graphs
It is a knowledge representation technique which consists of basic concepts and the relationship between
them. As the name indicates, it tries to capture the concepts about the events and represents them in the
form of a graph. A concept may be individual or generic. An individual concept has a type field followed
by a reference field. For example person : Ram. Here person indicates type and Ram indicates reference.

An individual concept should be represented within a rectangle in graphical representation and within a
square bracket in linear representation. The generic concept should be represented within an oval in
graphical representation and within a parenthesis in linear representation. Conceptual graph is a basic
building block for associative network. Concepts like AGENT, OBJECT, INSTRUMENT, PART are
obtained from a collection of standard concepts. New concepts and relations can be defined from these
basic ones. These are also basic building block for associative network. A linear conceptual graph is an
elementary form of this structure. A single conceptual graph is roughly equivalent to a graphical diagram
of a natural language sentence where the words are depicted as concepts and relationships.
Consider an example
“Ram is eating an apple “
Figure Graphical Representation
Conceptual Dependency
It is an another knowledge representation technique in which we can represent any kind of knowledge. It
is based on the use of a limited number of primitive concepts and rules of formation to represent any
natural language statement. Conceptual dependency theory is based on the use of knowledge
representation methodology was primarily developed to understand and represent natural language
structures. The conceptual dependency structures were originally developed by Roger C SChank in 1977.
If a computer program is to be developed that can understand wide phenomenon represented by natural
languages, the knowledge representation should be powerful enough to represent these concepts. The
conceptual dependency representation captures maximum concepts to provide canonical form of meaning
of sentences. Generally there are four primitives from which the conceptual dependency structure can be
described. They are
a. ACTS : Actions
b. PPs : Objects (Picture Producers)
c. AAs : Modifiers of Actions (Action Aiders)

d. PAs : Modifiers of PPs (Picture Aiders)
e. TS : Time of action
Conceptual dependency provides both a structure and a specific set of primitives at a particular level of
granularity, out of which representation of particular pieces of information can be constructed.
For example
Where ←: Direction of dependency
Double arrow indicates two way link between actor and action.
P: Past Tense
ATRANS: One of the primitive acts used by the theory
O: The objective case relation
R: Recipient case Relation
In CD, representation of actions are built from a set of primitive acts.
1) ATRANS: Transfer of an abstract relationship (give, accept, take)
2) PTRANS: Transfer the physical location of an object ( Go, Come, Run, Walk)
3) MTRANS: Transfer the mental information (Tell)
4) PROPEL: Application of physical force to an object (push, pull, throw)
5) MOVE: Movement of a body part by its owner (kick).
6) GRASP: Grasping of an object by an action (clutch)
7) INGEST: Ingestion of an object by an animal (eat)
8) EXPEL: Expel from an animal body (cry)
9) MBUILD: Building new information out of old (decide)
10) SPEAK: Production of sounds (say)
11) ATTEND: Focusing of a sense organ towards a stimulus (Listen)
The main goal of CD representation is to capture the implicit concept of a sentence and make it explicit.
In normal representation of the concepts, besides actor and object, other concepts of time, location, source
and destination are also mentioned. Following conceptual tenses are used in CD representation.

1) O : Object case relationship
2) R : Recipient case relationship
3) P : Past
4) F : Future
5) Nil : Present
6) T : Transition
7) Ts : Start Transition
8) Tf : Finisher Transition
9) K : Continuing
10) ? : Interrogative
11) / : Negative
12) C : Conditional
Also there are several rules in conceptual dependency
Rule 1: PP ACT
It describes the relationship between an actor and an event, he/she causes.
E.g. Ram ran
Ram PTRANS
Where P: Past Tense
Rule 2: PP PA
It describes the relationship between a PP and PA where the PA indicates one characteristics of PP. E.g.
Ram is tall
Ram Tall or Ram Height ( Average)
Rule 3: PP PP
It describes the relationship between two PPs where one PP is defined by other.
E.g. Ram is a doctor
Ram Doctor
Rule 4: PP or PA
PA PP

It describes the relationship between the PP and PA, where PA indicates one attributes of PP.
E.g. A nice boy is a doctor
Boy Doctor
Nice
Rule 5: PP
PP
It describes the relationship between 3 PP’s where one PP is the owner of another PP.
E.g. Ram’s Cat
Cat
Ram
Rule 6: Act PP Where O: Object
It describes the relationship between the PP and ACT. Where PP indicates the object of that action. E.g.
Ram is eating an apple.
Ram INGEST
O
Apple
Rule 7: ACT
(R: Recipient)
Here one PP describes the recipient and another PP describes the donner
E.g. Rahul gave a book to sourav.

Rule 8: (I: Instrument used in the action)
Here PP1 indicates the agent and PP2 indicates the object that is used in the action.
E.g. Tapash ate the ice cream with the spoon.
Rule 9:
Here D indicates destination, PP1 indicates destination and PP2 indicates the source.
E.g. the bucket is filled with milk.
xindicates the average milk and the source i.e. bucket is dry which is hidden.
Rule 10:
(T: Time)
It describes the relationship between a conceptualization and the time at which the event is described
occurs.

E.g. Sita ate the apple yesterday.
Rule 11:
It describes the relationship between a conceptualization and the place at which it is occurred.
E.g. Shanu ate the apple at VRS hotel yesterday
Rule 12:
It describes the relationship between one conceptualization with another.
E.g. while I was going to college, I saw a snake
(Where CP: Conscious Processor i.e. the combination of all sense organs like eye, ear, nose etc.)
By using the above rules we can represent any sentence. Let us visualize few examples on conceptual
dependency.

1) Sima gave a book to Niki
Where O: Object, P: Past Tense, R: Recipient, Sima: PP, Book: PP, Niki: PP, ATRANS: give
2) Bhabani cuts an apple with a knife
3) Sanjay drove the car fast
4) The rose was given by Rupa to Anand
5) Shruti pushed the door.
6) The man took a book
Here man is the doctor and book is the object of the action took.
7) My grandfather told me a story

8) Ira gave the man a dictionary
SCRIPT
It is an another knowledge representation technique. Scripts are frame like structures used to represent
commonly occurring experiences such as going to restaurant, visiting a doctor. A script is a structure that
describes a stereotyped sequence of events in a particular context. A script consist of a set of slots.
Associated with each slot may be some information about what kinds of values it may contain as well as a
default value to be used if no other information is available. Scripts are useful because in the real world,
there are no patterns to the occurrence of events. These patterns arise because of clausal relationships
between events. The events described in a script form a giant casual chain. The beginning of the chain is
the set of entry conditions which enable the first events of the script to occur. The end of the chain is the
set of results which may enable later events to occur. The headers of a script can all serve as indicators
that the script should be activated.
Once a script has been activated, there are a variety of ways in which it can be useful in interpreting a
particular situation. A script has the ability to predict events that has not explicitly been observed. An
important use of scripts is to provide a way of building a single coherent interpretation from a collection
of observation. Scripts are less general structures than are frames and so are not suitable for representing
all kinds of knowledge. Scripts are very useful for representing the specific kinds of knowledge for which
they were designed.
A script has various components like:
1) Entry condition: It must be true before the events described in the script can occur. E.g. in a
restaurant script the entry condition must be the customer should be hungry and the customer has
money.
2) Tracks: It specifies particular position of the script e.g. In a supermarket script the tracks may be
cloth gallery, cosmetics gallery etc.
3) Result: It must be satisfied or true after the events described in the script have occurred.
e.g. In a restaurant script the result must be true if the customer is pleased.
The customer has less money.
4) Probs: It describes the inactive or dead participants in the script e.g. In a supermarket script, the
probes may be clothes, sticks, doors, tables, bills etc.

5) Roles: It specifies the various stages of the script. E.g. In a restaurant script the scenes may be
entering, ordering etc.
Now let us look on a movie script description according to the above component.
a) Script name : Movie
b) Track : CINEMA HALL
c) Roles : Customer(c), Ticket seller(TS), Ticket Checker(TC), Snacks
Sellers (SS)
d) Probes : Ticket, snacks, chair, money, Ticket, chart
e) Entry condition : The customer has money
The customer has interest to watch movie.
6) Scenes:
a. SCENE-1 (Entering into the cinema hall)
C PTRANS C into the cinema hall
C ATTEND eyes towards the ticket counter
C PTRANS C towards the ticket counters
C ATTEND eyes to the ticket chart
C MBUILD to take which class ticket
C MTRANS TS for ticket
C ATRANS money to TS
TS ATRANS ticket to C
b. SCENE-2 (Entering into the main ticket check gate)
C PTRANS C into the queue of the gate
C ATRANS ticket to TC
TC ATTEND eyes onto the ticket
TC MBUILD to give permission to C for entering into the hall
TC ATRANS ticket to C
C PTRANS C into the picture hall.

c. SCENE-3 (Entering into the picture hall)
C ATTEND eyes into the chair
TC SPEAK where to sit
C PTRANS C towards the sitting position
C ATTEND eyes onto the screen
d. SCENE-4 (Ordering snacks)
C MTRANS SS for snacks
SS ATRANS snacks to C
C ATRANS money to SS
C INGEST snacks
e. SCENE-5 (Exit)
C ATTEND eyes onto the screen till the end of picture
C MBUILD when to go out of the hall
C PTRANS C out of the hall
7) Result:
The customer is happy
The customer has less money
Example 2: Write a script of visiting a doctor in a hospital
1) SCRIPT_NAME : Visiting a doctor
2) TRACKS : Ent specialist
3) ROLES : Attendant (A), Nurse(N), Chemist (C),
Gatekeeper(G), Counter clerk(CC), Receptionist(R), Patient(P),
Ent specialist Doctor (D), Medicine
Seller (M).
4) PROBES : Money, Prescription, Medicine, Sitting chair,
Doctor’s table, Thermometer, Stetho scope, writing pad, pen,
torch, stature.

5) ENTRY CONDITION: The patient need consultation.
Doctor’s visiting time on.
6) SCENES:
a. SCENE-1 (Entering into the hospital)
P PTRANS P into hospital
P ATTEND eyes towards ENT department
P PTRANS P into ENT department
P PTRANS P towards the sitting chair
b. SCENE-2 (Entering into the Doctor’s Room)
P PTRANS P into doctor’s room
P MTRANS P about the diseases
P SPEAK D about the disease
D MTRANS P for blood test, urine test
D ATRANS prescription to P
P PTRANS prescription to P.
P PTRANS P for blood and urine test
c. SCENE-3 (Entering into the Test Lab)
P PTRANS P into the test room
P ATRANS blood sample at collection room
P ATRANS urine sample at collection room
P ATRANS the examination reports
d. SCENE-4 (Entering to the Doctor’s room with Test reports)
P ATRANS the report to D
D ATTEND eyes into the report
D MBUILD to give the medicines
D SPEAK details about the medicine to P

P ATRANS doctor’s fee
P PTRANS from doctor’s room
e. SCENE-5 (Entering towards medicine shop)
P PTRANS P towards medicine counter
P ATRANS Prescription to M
M ATTEND eyes into the prescription
M MBUILD which medicine to give
M ATRANS medicines to P
P ATRANS money to M
P PTRANS P from the medicine shop
7) RESULT:
The patient has less money
Patient has prescription and medicine.
Advantages And Disadvantages Of Different Knowledge Representation
Sl.
No.
Scheme Advantages Disadvantages
1 Production
rules
• Simple syntax
• Easy to understand
• Simple interpreter
• Highly Modular
• Easy to add or modify
• Hard to follow Hierarchies
• Inefficient for large systems
• Poor at representing structured
descriptive knowledge.
2 Semantic • Easy to follow hierarchy
• Easy to trace associations
• Flexible
• Meaning attached to nodes might
be ambiguous
• Exception handling is difficult
• Difficult to program
3 Frame • Expressive Power
• Easy to set up slots for new
properties and relations
• Easy to create specialized
• Difficult to program
• Difficult for inference
• Lack of inexpensive software

procedures
4 Script • Ability to predict events
• A single coherent
interpretation may be build
up from a collection of
observations
• Less general than frames
• May not be suitable to represent all
kinds of knowledge
5 Formal Logic • Facts asserted independently
of use
• Assurance that only valid
consequence are asserted
• Completeness
• Separation of representation and
processing
• Inefficient with large data sets
• Very slow with large knowledge
bases
HUMAN ASSOCIATIVE MEMORY (HAM)
This model was developed by John Anderson and Gordon Bower (1973). This memory is organized as a
network of propositional binary trees. When an informant asserts a statement to HAM, the system parses
the sentence and builds a binary tree representation. As HAM is informed of new sentences, they are
parsed and formed into new tree like structures with existing ones. When HAM is posed with a query it is
formed into a tree structure called a probe. This structure is then matched against memory structures for
the best match. The structure with the closest match is used to formulate an answer to the query. Matching
is accomplished by first locating the leaf nodes in memory that match leaf nodes in the probe. The
corresponding links are then checked to see if they have the same labels and in the same order. The search
process is constrained by searching only node groups that have the same relation links. Access to nodes in
HAM is accomplished through word indexing in LISP.
In HAM, nodes in the tree are assigned with unique numbers, while links are labeled with some functions.
They are given below:
C : Context for free fact
E : Set membership
F : a fact
L : a location
O : An object
P : Predicate

R : Relation
S : Subject
T : Time (Present, past, future)
On the basis of above function, we can represent various sentences in HAM. Let us look some example
by using the above functions.
1) Sony met Rahul
Figure
On the above Ham Structure the time is the past (met), as Sonly did the work so sonly is the subject and
Rahul will be the object and the relation is met.
2) Ram is eating an apple at Nico Park.
Figure

3) Sima is eating ice cream as well as chips at Gandhi Park.
Figure
4) He came to me.
Figure
5) In a room Muna touched Lily.
Figure

Module 2
MIN-MAX Search
Games have always been an important application area for heuristic algorithms. In playing games whose
state space may be exhaustively delineated, the primary difficulty is in accounting for the actions of the
opponent. This can be handled easily by assuming that the opponent uses the same knowledge of the state
space as us and applies that knowledge in a consistent effort to win the game. Minmax implements game
search under referred to as MIN and MAX.
The min max search procedure is a depth first, depth limited search procedure. The idea is to start at the
current position and use the plausible move generator to generate the set of possible successor positions.
To decide one move, it explores the possibilities of winning by looking ahead to more than one step. This
is called a ply. Thus in a two ply search, to decide the current move, game tree would be explored two
levels farther.
Consider the below example
Figure Tree showing two ply search
In this tree, node A represents current state of any game and nodes B, C and D represent three possible
valid moves from state A. similarly E, F, G represents possible moves from B, H, I from C and J, K, L,
from D. to decide which move to be taken from A, the different possibilities are explored to two next
steps. 0, -3, 3, 4, 5, 6, -5, 0 represent the utility values of respective move. They indicate goodness of a
move. The utility value is back propagated to ancestor node, according to situation whether it is max ply
or min ply. As it is a two player game, the utility value is alternatively maximized and minimized. Here as

the second player’s move is maximizing, so maximum value of all children of one node will be back
propagated to node. Thus, the nodes B, C, D, get the values 4, 5, 6 respectively. Again as ply 1 is
minimizing, so the minimum value out of these i.e. 4 is propagated to A. then from A move will be taken
to B.
MIN MAX procedure is straightforward recursive procedure that relies on two auxiliary procedures that
are specific to the game being played.
1. MOVEGEN (position, player): the move generator which returns a list of nodes representing the
moves that can be made by player in position. We may have 2 players namely PLAYER-TWO in a
chess problem.
2. STATIC (position, player): the static evaluation function, which returns a number representing the
goodness of position from the standpoint of player.
We assume that MIN MAX returns a structure containing both results and that we have two functions,
VALUE and PATH that extract the separate components. A function LAST PLY is taken which is
assumed to evaluate all of the factors and to return TRUE if the search should be stopped at the current
level and FALSE otherwise.
MIN MAX procedure takes three parameters like a board position, a current depth of the search and the
players to move. So the initial call to compute the best move from the position CURRENT should be
MIN MAX (CURRENT, 0, PLAYER-ONE)
(If player is to move)
Or
MIN MAX (CURRENT, 0, PLAYER-TWO)
(If player two is to move)
Let us follow the algorithm of MIN MAX
Algorithm: MINMAX (position, depth, player)
1. If LAST PLY (position, depth)
Then RETURN VALUE = STATIC (position, player)
PATH = nil.
2. Else, generate one more ply of the tree by calling the function MOVE_GEN (position, player)
and set SUCCESORS to the list it returns.
3. If SUCESSORS is empty,

THEN no moves to be made
RETURN the same structure that would have been returned if LAST_PLY had returned TRUE.
4. If SUCCESORS is not empty,
THEN examine each element in turn and keep track of the best one.
5. After examining all the nodes,
RETURN VALUE = BEST- SCORE
PATH = BEST- PATH
When the initial call to MIN MAX returns, the best move from CURRENT is the first element in the
PATH.
Alpha- Beta (α-β) Pruning
When a number of states of a game increase and it cannot be predicted about the states, then we can use
the method pruning. Pruning is a method which is used to reduce the no. of states in a game. Alpha- beta
is one such pruning technique. The problem with minmax search is that the number of game states it has
to examine is exponential in the number of moves. Unfortunately we cannot eliminate the exponent, but
we can effectively cut it in half. Alpha-beta pruning is one of the solutions to the problem of minmax
search tree. When α-β pruning is applied to a standard minmax tree, it returns the same move as minmax
would, but prunes away branches that cannot possibly influence the final decision.
The idea of alpha beta pruning is very simple. Alpha beta search proceeds in a depth first fashion rather
than searching the entire space. Generally two values, called alpha and beta, are created during the search.
The alpha value is associated with MAX nodes and the beta value is with MIN values. The value of alpha
can never decrease; on the other hand the value of beta never increases. Suppose the alpha value of A
MAX node is 5. The MAX node then need not consider any transmitted value less than or equal to 5
which is associated with any MIN node below it. Alpha is the worst that MAX can score given that MIN
will also do its best. Similarly, if a MIN has a beta value of 5, it need not further consider any MAX node
below it that has a value of 6 or more.
The general principal is that: consider a node η somewhere in the search tree, such that player has a
choice of moving to that node. If player has a better choice К either at the parent node of η or at any
choice point further up, then η will never be reached in actual play. So once we have found out enough
about η (by examining some of its descendents) to reach this conclusion, we can prune it.
We can also say that “α” is the value of the best choice we have found so far at any choice point along the
path for MAX. Similarly “β” is the value of the best choice we have found so far at any choice point
along the path for MIN. Consider the following example

Figure
Here at MIN ply, the best value from three nodes is - 4, 5, 0. These will be back propagated towards root
and a maximizing move 5 will be taken. Now the node E has the value 8 is far more, then accepted as it is
minimizing ply. So, further node E will not be explored. In the situation when more plies are considered,
whole sub tree below E will be pruned. Similarly if α=0, β=7, all the nodes and related sub trees having
value less than 0 at maximizing ply and more than 7 at minimizing ply will be pruned.
Alpha beta search updates the value of α and β as it goes along and prunes the remaining branches at a
node as soon as the value of the current node is known to be worse than the current α and β value for
MAX or MIN respectively. The effectiveness of alpha- beta pruning is highly dependent on the order in
which the successors are examined suppose in a search tree the branching factor is x and depth d. the α-β
search needs examining only x0/9
nodes to pick up best move, instead of x0
for MINMAX.
Constraint Satisfaction Search
A constraint search does not refer to any specific search algorithm but to a layer of complexity added to
existing algorithms that limit the possible solution set. Heuristic and acquired knowledge can be
combined to produce the desired result a constraint satisfaction problem is a special kind of search
problem in which states are defined by the values of a set of variables and the goal state specifies a set of
constraints that the value must obey. There are many problems in AI in which the goal state is not
specified in the problem and it requires to be discovered according to some specific constraint. Examples
of some constraint satisfaction search include design problem, labeling graphs, robot path planning and
cryptarithmatic problem etc.
A constraint satisfaction problem (CSP) is defined by a set of variables (x:, x9----x; ) and a set of
constraints (c:, c9-----c). Each variable x= has a non empty domain d= of possible values. Each
constraint c= in values some subset of the variables and specifies the allowable combination of values for

that subset. The search space of CSPS is often exponential. Therefore a number of different approaches to
the problem have been proposed to reduce the search space and find a feasible solution in a reasonable
time based on the search space exploring and variable selection heuristics different algorithms and can be
developed for a CSP problem. The algorithms can be divided into two major categories such as complete
and incomplete algorithm.
Complete algorithms seek any solution or solutions of a CSP or they try to prove that no solution into the
categories like constraint propagation techniques which tries to eliminate values that are consistent with
some constraints and systematic search techniques. Which explores systematically the whole search
space. But on the other hand incomplete search methods do not explore the whole search space. They
search the space either non-systematically or in a systematic manner, but with a limit on some resource.
They may not provide a solution but their computational time is reasonably reduced. They cannot be
applied to find all solutions or to prove that no solution exists. Let us look an algorithm to solve a
constraint satisfaction problem.
Algorithm:
1) Open all objects that must be assigned values in a complete solution.
2) Repeat until all objects assigned valid values.
3) Select an object and strengthen as much as possible. The set of constraints that apply to object.
4) If set of constraints is different from previous set then open all objects that share any of these
constraints. Remove selected objects.
5) If union of constraints discovered above defines a solution, return solution.
6) If union of constraints discovered above defines a contradiction, return failure.
7) Make a guess in order to proceed. Repeat until a solution is found.
8) Select an object with a number assigned value and try strengthen its constraints.
PLANNING
The process of doing a sequence of actions to achieve a goal is called planning. A plan is a representation
of the crude structure of the input scene by the various object labels. The process of planning is a bottom
up process to provide clues concerning which knowledge can be applied to different parts of the scene.
The knowledge of the task world is represented by sets of productions rules. Each rule in the bottom up
process has a fuzzy predicate which describes the properties of relations between objects. Generally there
are various agents who act to plan. The environments for an agent may be deterministic, finite, static in
which change happens only when the agent acts. The discrete environment includes the time factor,
objects, effects etc. These environments are called classical planning environments. On the other hand,
the non classical planning environments are partially observable and involves a different set of algorithms

and agent designs. Planning refers to the process of computing several steps of a problem solving
procedure before evaluation of that problem.
Computer cannot solve any problem without planning it. For example, in 8-puzzle game, the computer
can’t replace the tiles onto their positions without the planning procedure of that problem. When we
discuss the computer solution of the 8-puzzle game, what we are really doing was outlining the way the
computer might generate a plan for solving it. A computer could look for a solution plan in the same way
as a person who was actually trying to solve the problem by moving tiles on a board. If solution steps in
the real world cannot be ignored or undone, though planning becomes extremely important. Although real
world steps may be irrevocable, computer simulation of those steps is not. So we can circumvent the
constraints of the real world by looking for a complete solution in a simulated world in which
backtracking is allowed. After we find a solution, we can execute it in the real world. The fact that we can
leave out properties of world states that are irrelevant to the problem at hand or that are not known is one
of the powerful aspects of using a feature based approach. This aspect is particularly important is
describing the goal condition that we want the agent to achieve by its actions.
Basic Components of a Planning System
When a particular problem will be solved, at that time some specific rules regarding to that problem are to
be applied. Then apply the choosen rule to compute the new problem state that arises from its application.
Detect when a solution has been found and calculate the active and inactive ends of that problem. Various
components of a planning system are described as follows.
(a) States: For a planning process, the planners decompose the world into some environments. Then
environments are defined by some logical conditions and states. The problems can be viewed as
the task of finding a path from a given starting state to some desirable goal state. The state can be
viewed as a conjunction of positive literals. For example, Rich A famous might represent the
state of a best agent.
(b) Goal: A goal is a specified state. To find a solution to a problem using a search procedure is to
generate moves through the problem space until a goal state is reached. In the context of game
playing programs, a goal state is one in which we win. Unfortunately, for interesting games like
chess, it is not usually, possible, even with a good plausible move generator, to search until a
goal state is found.
(c) Actions: An action is specified en terms of the pre-conditions that must hold before it can be
executed and then the effects that ensue when it is executed. For example, an action for running a
tiger from one location to another is
Action Run T, from, to,
PRECONDITION: At T, from ? Tiger T ? Jungle from ? Jungle To

EFFECT: ~ At T, from ? At T, to
(d) Precondition: The precondition is a conjunction of function free positive literals stating what
must be true in a state before the action can be executed.
(e) Effect: It is a conjunction of function free literals describing how the state changes when the
action is executed.
(f) Finding a solution: A planning system has succeeded in finding a solution to a problem when it
has found a sequence of operators that transforms the initial problem state into the goal state. The
way it can be solved depends on the way that state descriptions are represented.
(g) Calculating the Dead State: As a planning system is searching for a sequence of operators to
solve a particular problem, it must be able to detect when it is exploring a path that can never
lead to a solution. The same reasoning methods that can be used to detect a solution can often be
used for detecting a dead path. If the search process is reasoning in forward direction from the
initial state, it can prune any path that leads to a state from which the goal state cannot be
reached. If the search process is reasoning backward from the goal state, it can also terminate a
path either because it is sure that the starting state cannot be reached.
Planning in State Space Search
Problem solving in AI may be characterized as a systematic search through a range of possible actions in
order to reach some predefined goal or solution. The problem solving agents decide what to do by finding
sequence of action that lead to desirable states. The simplest agents which have been described below are
the reflex and goal based agents. The reflex agents use direct mapping from states to actions and are
unsuitable for very large mappings. Problem solving agents find action sequence that lead to desirable
state.
A state space is represented by four components like steps involved in a problem solving process, the start
state of the problem and the corresponding goal state. Search algorithms should track the paths from the
start node to the goal node because these paths contain a series of operations that lead to the solution of
the problem. A programmer should analyze and predict the behaviour of search algorithms to successfully
designed and implement them in a proper manner. The problems can be characterized as spaces consisting
of a set of states and a set of operators that map from one state to another state. The states may be
distinguished as containing one of the following: one or more initial or starting states, a number of
intermediate states and one or more goal states. A sequence of operators that map an initial state to a goal
state will provide the solution to a problem. A best solution is one that requires the fewest no. of
operations while mapping from an initial state to the goal state. The amount of time and memory space
required to complete the mapping measures the performance of a particular solution method. The state
space search can be in forward and backward direction. The forward state space planning is also known as

progression planning in which searching always takes place in forward direction. In backward search, it
finds only the relevant actions. An agent with several immediate options of unknown values can decide
what to do by first examining the different possible sequences of actions that lead to states of known
values and then choosing the best one.
A state space search can be searched in two directions like from the inputs towards the goal or from the
goals towards the inputs. In data driven search, one starts with the given facts of a problem and uses a set
of legal moves or rules to change the states. This process is continued until it generates a path that
satisfies the goal condition. In goal driven search, first determine the rules or legal moves that can be used
to generate the goal and identifies the condition that can be applied to use these rules. These conditions
form the new goals or sub goals for the search. One must continue the search process by working
backwards through successive sub goals until it returns of moves or rules leading from the data to a goal,
even if it performs this process backwards. Data driven search is suggested if one is provided with almost
all the data at the time of formulation of the problem statement.
Data driven search uses the knowledge and constraints present in the given data of a problem to guide the
search along a path. The main work in the area of search strategies is to find the correct search
performance measures like time complexity, space complexity, completeness and optimality help to judge
the fitness of a particular search algorithm.
Various Planning Techniques
Several planning techniques are described below.
(1) Hierarchical Planning: In hierarchical planning, at each level of hierarchy the objective
functions are reduced to a small number of activities at the next lower level. So the computational
cost of finding the correct way to arrange these activities for the current problem is small.
Hierarchical methods can result in linear time. The initial plan of hierarchical planning describes
the complete problem which is a very high level description. The plans are refined by applying
action decompositions. Each action decomposition reduces a high level description to some of the
individual lower level descriptions. The action decomposers describe how to implement the
actions.
(2) Conditional Planning: It deals with the planning by some appropriate conditions. The agents
plan first and then execute the plan that was produced. The agents find out which part of the plan
to execute by including sensing actions in the plan to test for the appropriate conditions.
(3) Exact Planning: It is also called as conformation planning. It ensures that the plan achieves the
goal in all possible circumstances regardless of the true initial state and the actual actions
outcome. This planning is based on the idea that the world can be forced into a given state even
when the agent has only partial information about the current state.

(4) Replanning: It occurs when there is any wrong information regarding with the planning. The
agent can plan the same plan as the conditional planner or some new steps.
(5) Continuous Planning: In this planning, the planner at first achieves the goal and then only can
stop. A continuous planner is designed to persist over a lifetime. It can handle any unfavorable
circumstances in the environment.
(6) Multiagent Planning: In multiagent planning some other new agents may involved with our
single agent in the environment. This may lead to a poor performance because dealing with other
agents is not the same as dealing with the nature. It is necessary when there are other agents in the
environment with which to cooperate, compete or coordinate.
(7) Multibody Planning: This planning constructs joint plans, using an efficient decomposition of
joint action descriptions, but must be augmented with some form of co-ordination of two
cooperative agents are to agree on which joint plan to execute.
UNDERSTANDING
Understanding is the simplest procedure of all human beings. Understanding means ability to determine
some new knowledge from a given knowledge. For each action of a problem, the mapping of some new
actions is very necessary. Mapping the knowledge means transferring the knowledge from one
representation to another representation. For example, if you will say “I need to go to New Delhi” for
which you will book the tickets. The system will have “understood” if it finds the first available plane to
New Delhi. But if you will say the same thing to you friends, who knows that your family lives in “New
Delhi”, he/she will have “understood” if he/she realizes that there may be a problem or occasion in your
family. For people, understanding applies to inputs from all the senses. Computer understanding has so
far been applied primarily to images, speech and typed languages. It is important to keep in mind that the
success or failure of an “understanding” problem can rarely be measured in an absolute sense but must
instead be measured with respect to a particular task to be performed. There are some factors that
contribute to the difficulty of an understanding problem.
(a) If the target representation is very complex for which you cannot map from the original
representation.
(b) There are different types of mapping factors may arise like one-to-one, one-to-many and many-
to-many.
(c) Some noise or disturbing factors are also there.
(d) The level of interaction of the source components may be complex one.
(e) The problem solver might be unknown about some more complex problems.
(f) The intermediary actions may also be unavailable.
Consider an example of an English sentence which is being used for communication with a keyword-
based data retrieval system. Suppose I want to know all about the temples in India. So I would need to be
translated into a representation such as

SEARCH KEYWORDS TEMPLE INDIA
The above sentence is a simple sentence for which the corresponding representation may be easy to
implement. But what for the complex queries?
Consider the following query.
“Ram told Sita he would not eat apple with her. He has to go to the office”. This type of complex
queries can be modeled with the conceptual dependency representation which is more complex than that
of simple representation. Constructing these queries is very difficult since more informationare to be
extracted. Extracting more information will require some more knowledge.
Also the type of mapping process is not quite easy to the problem solver. Understanding is the process of
mapping an input from its original form to a more useful one. The simplest kind of mapping is “one-to-
one”. In one-to-one mapping each different problems would lead to only one solution. But there are very
few inputs which are one-to-one. Other mappings are quite difficult to implement. Many-to-one mappings
are frequent is that free variation is often allowed, either because of the physical limitations of that
produces the inputs or because such variation simply makes the task of generating the inputs. Many-to-
one mapping require that the understanding system know about all the ways that a target representation
can be expressed in the source language. One-to-many mapping requires a great deal of domain
knowledge in order to make the correct choice among the available target representation. The mapping
process is simplest if each component can be mapped without concern for the other components of the
statement. If the number of interactions increases, then the complexity of the problem will increase. In
many understanding situations the input to which meaning should be assigned is not always the input that
is presented to the under stander. Because of the complex environment in which understanding usually
occurs, other things often interfere with the basic input before it reaches the under stander. Hence the
understanding will be more complex if there will be some sort of noise on the inputs.
NATURAL LANGUAGE PROCESSING
Natural language processing is a subfield of computer science and in artificial intelligence that is
concerned with computational processing of natural languages, emulating cognitive capabilities without
being committed to a true simulation of cognitive processes. It is a theoretically motivated range of
computational techniques for analyzing and representing naturally occurring texts at one or more levels of
linguistic analysis for the purpose of achieving human like language processing for a range of tasks or
applications. It is a computerized approach to analyzing text that is based on both a set of theories and a
set of technologies. NLP is a very active area of research and development. Naturally occurring texts can
be of any language, mode and genre etc. The text can be oral or written. The only requirement is that they
be in a language used by humans to communicate to one another. Also, the text being analyzed should not

be specifically constructed for the purpose of analysis, but rather that the text is gathered from actual
usage.
The notion of levels of linguistic analysis refers to the fact that there are multiple types of language
processing known to be at work when humans produce or comprehend language. The humans use
generally various types of sentences for expressing their feelings. Sentences are classified by structure
and usage. A simple sentence has one independent clause comprised of a subject and predicate. A
compound sentence consists of two or more independent clauses connected by a conjunction or a
semicolon. The way a sentence is used determines its mood, declarative, imperative, interrogative or
exclamatory. A word functions in a sentence as a part of speech. Parts of speech for the English language
are nouns, pronouns, verbs, adjectives, adverbs, prepositions, conjuctions etc.
Generally NLP is the means for accomplishing a particular task. It is a combination of computational
linguistics and artificial intelligence. The natural language processing uses the tools of AI such as:
algorithms, data structures, formal models for representing knowledge, models or reasoning processes etc.
There are two ways through which the natural languages are being processed. First parsing technique and
the second is the transition network. The architecture of NLP is given figure .
Figure Architecture of a NLP
In NLP, to interact with the database in natural languages, computer is required to have knowledge of
basic alphabet, lexicon, grammar and words formation etc. The inputs are in the form of natural language
given by the user. Finally after parsing process the output in the language is being understood by the
application program.
GOALS OF NLP
The goal of natural language processing is to specify a language comprehension and production theory to
such a level of detail that a person is able to write a computer program which can understand and produce
natural language. The basic goal of NLP is to accomplish human like language processing. The choice of
word “processing” is very deliberate and should not be replaced with “understanding”. For although the
field of NLP was originally referred to as Natural Language Understanding (NLU), that goal has not yet
been accomplished. A full NLU system would be able to:
→ Paraphrase an input text.
Inputs in form of
Natural language User
Interface for
natural language
Parsing
Process
Application
Program

→ Translate the text into another language.
→ Answer questions about the contents of the text.
→ Draw inferences from the text.
While NLP has made serious inroads into accomplishing goals from first to third, the fact that NLP
system can not, of themselves, draw inferences from text, NLU still remains the goal of NLP. Also there
are some practical applications of NLP. An NLP-based IR system has the goal of providing more precise,
complete information in response to a user’s real information need. The goal of the NLP system is to
represent the true meaning and intent of the user’s query, which can be expressed as naturally in everyday
language.
APPLICATIONS OF NLP
NLP lie in a number of disciplines like computer and information sciences, linguistics, mathematics,
electrical and electronic engineering, artificial intelligence and robotics, psychology etc. Applications of
NLP include a number of fields of studies such as machine translation, natural language text processing,
summarization, user interfaces multilingual and Gross language information retrieval (CLIR), speech
recognition, artificial intelligence and expert system. Research on NLP is regularly published in a number
of conferences such as the annual proceedings of ACL (Association of Computational Linguistics) and its
European counter part EACL, biennial proceedings of the Message Understanding Conferences (MUCS),
Text Retrieval Conferences (TRECS) and ACM-SIGIR (Association of Computing Machinery-Special
Interest Group on Information Retrieval) conferences.
As natural language processing technology matures, it is increasingly being used to support other
computer applications. Such use naturally falls into two areas, one in which linguistic analysis merely
serves as an interface to the primary program and the second one in which natural language
considerations are central to the application. Natural language interfaces into a request in a formal
database query language, and the program then proceeds as it would without the use of natural language
processing techniques. The design of question answering systems is similar to that for interfaces to
database management systems. One difference however, is that the knowledge base supporting the
question answering system does not have the structure of a database. Similarly in message understanding
systems, a fairly complete linguistic analysis may be required but the messages are relatively short and
the domain is often limited. Also some more application areas include information and text
categorization. In both applications, natural language processing imposes a linguistic representation on
each document being considered. In text categorization a collection of documents is inspected and all
documents are grouped into several categories based on the characteristics of the linguistic
representations of the documents. In information filtering documents satisfying some criterion are singled
out from a collection.

Discourse Knowledge
While syntax and semantics work with sentence-length units, the discourse level of NLP works with units
of text longer than a sentence i.e. it does not interpret multi-sentence texts as just concatenated sentences,
each of which can be interpreted singly. Discourse focuses on the properties of the text as a whole that
convey meaning by making connections between component sentences. Several types of discourse
processing can occur at this level like anaphora resolution and discourse/text structure recognition.
Anaphora resolution is the replacing of words such as pronouns which are semantically vacant with the
appropriate entity to which they refer. For example, newspaper articles can be deconstructed into
discourse components such as: lead, main story, previous events, evaluation etc. A discourse is a
sequence of sentences. Discourse has structure much like sentences do. Understanding discourse structure
is extremely important for dialog system.
For example: The dialog may be
When does the bus to Bhubaneswar leave?
There is one at 10 a.m. and one at 1 p.m.
Give me two tickets for the earlier one, please.
The problems with discourse analysis may be non-sentential utterances, cross-sentential anaphora.
Pragmatic Knowledge
This level is concerned with the purposeful use of language in situations and utilizes context over and
above the contents of the text for understanding. The goal is to explain how extra meaning is read into
texts without actually being encoded in them. This requires much world knowledge including the
understanding of intentions, plans and goals. Some NLP applications may utilize knowledge bases and
inferencing modules. Pragmatic is the study of how more gets communicated than is said. Speech acts in
the pragmatic processing is the illocutionary force, the communicative force of an utterance, resulting
from the function associated with it. For example: Suppose the sentence is I will see you later.
Prediction: I predict that I will see you later.
Promise: I promise that I will see you later.
Warning: I warn you that I will see you later.

EXAMPLES OF SOME NLP SYSTEMS
In early 1950s, few NLP systems had been developed. Except the theoretical developments many
practical systems were developed to demonstrate the effectiveness of particular principles. Weizenbaum’s
ELIZA was built to replicate the conversation between a psychologist and a patient; simply by permuting
the user input. Winograd’s SHRDLU simulated a robot that manipulated blocks on a table top. Also
LUNAR was developed by Woods as an interface system to a database. In the late 1970’s, McKeown’s
discourse planner TEXT and McDonald’s response generator MUMMBLE used theoretical predicates to
produce declarative descriptions in the form of short texts, usually paragraphs. Some of the earliest NLP
systems are described below.
ELIZA System
ELIZA was an early natural language processing system capable of carrying on a limited form of
conversation with a user. It was developed by Weizenbaum in 1966. It is a remarkably simple program
that makes use of pattern matching to process the input and translate it into suitable outputs. Consider the
following session with a version of ELIZA that limited the responses of a psychotherapist.
User 1:
You are like my brother in some ways.
ELIZA 1:
WHAT RESEMBLANCE DO YOU SEE?
User 2:
You are not very intelligent but I think you don’t want me to notice that.
ELIZA 2:
WHAT MAKES YOU THINK I AM NOT INTELLIGENT?
User 3:
You don’t agree with me.
ELIZA 3:
WHY DO YOU THINK I DON’T AGREE WITH YOU?
User 4:
You are afraid of me.

ELIZA 4:
DOES IT PLEASE YOU TO BELIEVE I AM AFRAID OF YOU?
User 5:
I need some help, that much seems certain.
ELIZA 5:
WHAT WOULD IT MEAN TO YOU IF YOU GOT SOME HELP?
This program is one of the most popular AI programs of all time since it carried the keyword based
analysis approach to an absurd (but impressive) extreme. The success of this simple technique in this
domain is due to the fact that ELIZA doesn’t actually need to know anything to a mimic a
psychotherapist. As Weizenbaum notes, this is one of the few dialogue genres where the listener can act if
they know nothing of the world. ELIZA’s deep relevance to Turing’s ideas is that many people who
interacted with ELIZA cam to believe that it really understood them and their problems. Indeed,
Weizenbaum (1976) notes that many of these people continued to believe in ELIZA’s abilities even after
the program’s operation was explained to them.
LUNAR System
It was developed by Woods in 1970. It is one of the largest and most successful question-answering
system using AI techniques. This system had a separate syntax analyzer and a semantic interpreter. Its
parser was written in ATN (Augmented Transition Network) form. The system was used in various tests
and responded successfully to queries like followings:
→ How many oak trees have height greater than 15 inches?
→ What is the average concentration of hydrogen and oxygen in water?
→ Which one is the oldest material between Iron, Bauxite and Aluminum?
The LUNAR system is mainly deal with queries. But the performance of the system is very good than
other systems.
HAL System
HAL is an artificial agent capable of such advanced language processing behaviour as speaking and
understanding English. The HAL system was developed by Arthur C. Clarke. Generally HAL system is
useful for language and speech recognition. By speech and language processing we have in mind those
computational techniques that process spoken and written human language. HAL require much broader
and deeper knowledge of language. To determine what the user is saying, HAL must be capable of
analyzing an incoming audio signal and recovering the exact sequence of words user used to produce that

signal. Similarly, in generating its response HAL must be able to take a sequence of words and generate
an audio signal that the user can recognize. Both of these tasks require knowledge about phonetics and
phonology which can help model how words are pronounced in colloquial speech.
SHRDLU System
It was developed by Winograd in 1970. It was a dialogue system which could converse with a human user
about simple world containing building blocks. It is a simulation based programming system involving of
a hand and eye. It is a syntax based system which is a combination of deep and surface structure. It
contains a syntactic parser with a fairly wide coverage which builds surface structures that are not simply
of trivial category labeling. It performs the combination and integration of many components which will
create a total system. For example: User 1: Chose a green pen.
SHRDLU: Ok
User 2: Write with the pen.
SHRDLU: I DON’T UNDERSTAND WHICH PEN YOU MEAN.
The Chomsky Hierarchy of Grammars
A hierarchy of classes of languages viewed as sets of strings, ordered by their “complexity”. The higher
the language is in the hierarchy, the more complex it is. In particular the class of languages in one class
properly includes the languages in lower classes. There exists a correspondence between the class of
languages and the format of phrase structure rules necessary for generating all its languages. Noam
Chomsky defined a hierarchy of grammars called Type 0, Type 1, Type 2 and Type 3. The outline of
Chomsky hierarchy of languages is given in figure .

Figure Chomsky
The Chomsky hierarchy of languages reflects a certain order of complexity in some sense, the lower the
language class is in the hierarchy; the simplest are its possible constructions.
1. Type 0 (Recursively Enumerable Languages):
that cannot be empty string in the rewrite form abc
than others.
2. Type 1 (Context Sensitive Languages):
on the right hand side of the rewrite rule must be at least as long as the strings on the left side. In
production of the form in
empty string.
For example:
Figure Chomsky Hierarchy of Languages
language class is in the hierarchy; the simplest are its possible constructions.
Type 0 (Recursively Enumerable Languages): It is obtained by making the simple restrictions
that cannot be empty string in the rewrite form abc → adc. It is the much more simplest grammar
Type 1 (Context Sensitive Languages): They have added restriction that the length of the string
e right hand side of the rewrite rule must be at least as long as the strings on the left side. In
production of the form in , b must be a single non terminal symbol and d is a non
Recursively-
Enumerable
languages
Context-Sensitive
languages
Context force
language
Regular Language
s obtained by making the simple restrictions
adc. It is the much more simplest grammar
They have added restriction that the length of the string
e right hand side of the rewrite rule must be at least as long as the strings on the left side. In
, b must be a single non terminal symbol and d is a non

aA ab
aA aa
ba K AB
3. Type 2 (Context Free Grammar): This type can be represented as
LMNOPQR % LMNOPQR 1 % LMNOPQR 2 % LMNOPQRS % TUVWVS X 1 Also in
type 2 grammars the left hand side is a single non terminal symbol which is very important. Its
computational device is push down automata.
For example: S as
S asa
S aA
S aAB
A a
B b
4. Type 3 (Regular Languages): It is the most restrictive type grammar than others. The rules of
type 3 grammar is A aB
A a.
Its computational device is finite state automata. Example: B bA
B ba
So we can say that always a terminal symbol should start first (like ‘b’) then any no. of symbols
may come.
Some symbols are used to represent the grammars and they are described as follows.
• Sentences → S
• Verb Phrase → VP
• Noun Phrase → NP
• Preposition Phrase → PP
• Auxiliary → AUX
• Preposition → PREP
• Adverb → ADV
• Adjectives → ADJ
• Determiners → DEJ

• Article → ART
• Noun → N
• Verb → V and so on.
Also while representing a sentence we have to follow some rules like as follows.
1. S → NP VP
Or
S → NP VP PP
2. VP → AUX VP
VP → V NP
VP → V PP
VP →V ADJ
VP →V NP PP
VP → V
VP →AUX V NP
3. NP → NP ART
NP → N ADJ
NP → N PP
NP → N
4. PP → PREP NP
By taking into consideration the above rules any types of syntactic grammars are possible. Always you
have to remember a sentence must have at least two clauses like verb phrase (VP) and noun phrase (NP).
Now let us see some other types of grammars used in language processing like transformational grammar,
case grammar, systemic grammar and semantic grammar etc.
Transformational Grammar
These are the grammars in which the sentence can be represented structurally into two stages. Obtaining
different structures from sentences having the same meaning is undesirable in language understanding
systems. Sentences with the same meaning should always correspond to the same internal knowledge
structures. In one stage the basic structure of the sentence is analyzed to determine the grammatical
constituent parts and in the second stage just the vice versa of the first one. This reveals the surface
structure of the sentence, the way the sentence is used in speech or in writing. Alternatively, we can also
say that application of the transformation rules can produce a change from passive voice to active voice
and vice versa. Let us see the structure of a sentence as given below.
1. Ram is eating an apple (In Active Voice)

Figure (Transformational grammar Tree representation of Active voice to sentence formation)
2. An apple is being eaten by Ram (In Passive Voice)
Figure Passive voice to sentence formation
Both of the above sentences are two different sentences but they have same meaning. Thus it is an
example of a transformational grammar. These grammars were never widely used in computational
models of natural language. The applications of this grammar are changing of voice (Active to
Passive and Passive to Active) change a question to declarative form etc.
Case Grammars (FILLMORE’s Grammar)
Case grammars use the functional relationships between noun phrases and verbs to conduct the more
deeper case of a sentence. Generally in our English sentences, the difference between different forms of a
sentence is quite negligible. In early 1970’s Fillmore gave some idea about different cases of a English
sentence. He extended the transformational grammars of Chomsky by focusing more on the semantic
aspects of view of a sentence. In case grammars a sentence id defined as being composed of a preposition
P, a modality constituent M, composed of mood, tense, aspect, negation and so on. Thus we can represent
a sentence like
Where P Set of relationships among verbs and noun phrases i.e. P = (C=Case)
M Modality constituent
For example consider a sentence “Ram did not eat the apple”.
S → M + P

Figure Case Grammar Tree Representation
The tree representation for a case grammar will identify the words by their modality and case. The cases
may be related to the actions performed by the agents, the location and direction of actions. The cases
may also be instrumental and objective. For example “Ram cuts the apple by a knife”. Here knife is an
instrumental case. In fig 8.5 the modality constituent is the negation part, eat is the verb and Ram, apple
are nouns which are under the case C1 and C2 respectively. Case frames are provided for verbs to identify
allowable cases. They give the relationships which are required and which are optional.
Semantic Grammars
Semantic grammars encode semantic information into a syntactic grammar. They use context free rewrite
rules with non terminal semantic constituents. Generally a semantic error occurs when the meaning of the
knowledge is not properly communicated. Semantic errors occur if the human expert misinterprets the
knowledge engineer’s question or answers inappropriately. The proper constituents of semantic grammar
are noun phrase (NP), verb phrase (VP), Noun (N), Verb (V), Preposition phrase (PP), Adverb (ADV)
and so on. One of the successful applications of semantic grammar is the LIFER system (A database
query system). In the LIFER system, there are several rules to handle the wh-queries such as:
What is the name of your country?
Which department of your country is must efficient?
Who guide them?
Where they are being guided?
These sentences are analyzed and words matched to the symbols contained in the lexicon entries.
Semantic grammars are suitable for use in systems with restricted grammars since its power of
computation is limited.

Context Free Grammar (CFG)
The grammar in which each production has exactly one terminal symbol in its left hand side and at least
one symbol at the right hand side is called context free grammar. A CFG is a four tuple Σ, V, S, P where
Σ: Finite non empty set of terminals, the alphabet.
V: Finite non empty set of grammar variables (categories or non terminal symbols)
Such as Σ V ]
S: Starting symbol S ^ V
P: Finite set of production rules, each of the form A α, where
A ^ V and α ^ V ` ΣK
Each terminal symbol in a grammar denotes a language. The non terminals are written in capital letters
and terminals are written in small letters. Some properties of CFG formalism are
→ Concatenation is the only string combination operation.
→ Phrase structure is the only syntactic relationship.
→ The terminal symbols have no properties.
→ Non terminal symbols are atomic.
→ Most of the information encoded in a grammar lies in the production rules.
→ Any attempt of extending the grammar with semantics requires extra means.
→ Concatenation is not necessarily the only way by which phrases may be combined to yield other
phrases.
→ Even if concatenation is the sole string operation, other syntactic relationships are being put
forward.
For example we can write the followings:

PARSING PROCESS
Parsing is the term used to describe the process of automatically building syntactic analysis of a sentence
in terms of a given grammar and lexicon. The resulting syntactic analysis may be used as input to a
process of semantic interpretation. Occasionally, parsing is also used to include both syntactic and
semantic analysis. The parsing process is done by the parser. The parsing performs grouping and labeling
of parts of a sentence in a way that displays their relationships to each other in a proper way.
The parser is a computer program which accepts the natural language sentence as input and generates an
output structure suitable for analysis. The lexicon is a dictionary of words where each word contains some
syntactic, some semantic and possibly some pragmatic information. The entry in the lexicon will contain
a root word and its various derivatives. The information in the lexicon is needed to help determine the
function and meanings of the words in a sentence. The basic parsing technique is shown in figure .
Figure Parsing Technique
Generally in computational linguistics the lexicon supplies paradigmatic information about words
including part of speech labels, irregular plurals and sub categorization information for verbs.
Traditionally, lexicons were quite small and were constructed largely by hand. The additional information
being added to the lexicon increase the complexity of the lexicon. The organization and entries of a
lexicon will vary from one implementation to another but they are usually made up of variable length data
structures such as lists or records arranged in alphabetical order. The word order may also be given in
terms of usage frequency so that frequently used words like “a”, “the” and “an” will appear at the
beginning of the list facilitating the search. The entries in a lexicon could be grouped and given word
category (by articles, nouns, pronouns, verbs, adjectives, adverbs and so on) and all words contained
within the lexicon listed within the categories to which they belong. The entries are like a, an
(determiner), be (verb), boy, stick, glass (noun), green, yellow, red (adjectives), I, we, you, he, she, they
(pronouns) etc.
Input Sentence Parser Output Sentence
Lexicon

In most contemporary grammatical formalisms, the output of parsing is something logically equivalent to
a tree, displaying dominance and precedence relations between constituents of a sentence. Parsing
algorithms are usually designed for classes of grammar rather than tailored towards individual grammars.
Types of Parsing
The parsing technique can be categorized into two types such as
1. Top down Parsing
2. Bottom up Parsing
Let us discuss about these two parsing techniques and how they will work for input sentences.
1 Top down Parsing
Top down parsing starts with the starting symbol and proceeds towards the goal. We can say it is the
process of construction the parse tree starting at the root and proceeds towards the leaves. It is a strategy
of analyzing unknown data relationships by hypothesizing general parse tree structures and then
considering whether the known fundamental structures are compatible with the hypothesis. In top down
parsing words of the sentence are replaced by their categories like verb phrase (VP), Noun phrase (NP),
Preposition phrase (PP), Pronoun (PRO) etc. Let us consider some examples to illustrate top down
parsing. We will consider both the symbolical representation and the graphical representation. We will
take the words of the sentences and reach at the complete sentence. For parsing we will consider the
previous symbols like PP, NP, VP, ART, N, V and so on. Examples of top down parsing are LL (Left-to-
right, left most derivation), recursive descent parser etc.
Example 1: Rahul is eating an apple.
Symbolical Representation
S → NP VP
N VP (∴ NP N)
N AUX VP ( VP AUX VP)
N AUX V NP ( VP V NP)
  UX V ART N (P ART N)
 UX V ART apple
N AUX V an apple
N AUX eating an apple

 is eating an apple
Rahul is eating an apple.
Graphical Representation
Figure Example of Top down Parsing
Example 2: The small tree shades the new house by the stream.
Symbolical Representation
SNP VP
ART NP VP
The ADJ N VP
The small N V NP
The small tree V ART NP
The small tree shades ART ADJ NP
The small tree shades the ADJ N NP
The small tree shades the new N PREP N
The small tree shades the new house PREP ART N
The small tree shades the new house by ART N
The small tree shades the new house by the N

The small tree shades the new house by the stream.
Figure Top down Parsing
Bottom up Parsing
In this parsing technique the process begins with the sentence and the words of the sentence is replaced by
their relevant symbols. This process was first suggested by Yngve (1955). It is also called shift reducing
parsing. In bottom up parsing the construction of parse tree starts at the leaves and proceeds towards the
root. Bottom up parsing is a strategy for analyzing unknown data relationships that attempts to identify
the most fundamental units first and then to infer higher order structures for them. This process occurs in
the analysis of both natural languages and computer languages. It is common for bottom up parsers to
take the form of general parsing engines that can wither parse or generate a parser for a specific
programming language given a specific of its grammar.
A generalization of this type of algorithm is familiar from computer science LR (k) family can be seen as
shift reduce algorithms with a certain amount (“K” words) of look ahead to determine for a set of possible
states of the parser which action to take. The sequence of actions from a given grammar can be pre-
computed to give a ‘parsing table’ saying whether a shift or reduce is to be performed and which state to
go next. Generally bottom up algorithms are more efficient than top down algorithms, one particular

phenomenon that they deal with only clumsily are “empty rules”: rules in which the right hand side is the
empty string. Bottom up parsers find instances of such rules applying at every possible point in the input
which can lead to much wasted effort. Let us see some examples to illustrate the bottom up parsing.
Example-1: Rahul is eating an apple.
 is eating an apple.
N AUX eating an apple.
N AUX V an apple.
N AUX V ART apple.
N AUX V ART N
N AUX V NP
N AUX VP
N VP
NP VP
S
Figure Examples of Bottom up Parsing
Example-2:
The small tree shades the new house by the stream

ART small tree shades the new house by the stream
ART ADJ tree shades the new house by the stream
ART ADJ N shades the new house by the stream
ART ADJ N V the new house by the stream
ART ADJ N V ART new house by the stream
ART ADJ N V ART ADJ house by the stream
ART ADJ N V ART ADJ N by the stream
ART ADJ N V ART ADJ N PREP the stream
ART ADJ N V ART ADJ N PREP ART stream
ART ADJ N V ART ADJ N PREP ART N
ART ADJ N V ART ADJ N PREP NP
ART ADJ N V ART ADJ N PP
ART ADJ N V ART ADJ NP
ART ADJ N V ART NP
ART ADJ N V NP
ART ADJ N VP
ART NP VP
NP VP
S

Figure Example of Bottom up Parsing
Deterministic Parsing
A deterministic parser is one which permits only one choice for each word category. That means there is
only one replacement possibility for every word category. Thus, each word has a different test conditions.
At each stage of parsing always the correct choice is to be taken. In deterministic parsing back tracking to
some previous positions is not possible. Always the parser has to move forward. Suppose the parser some
form of incorrect choice, then the parser will not proceed forward. This situation arises when one word
satisfies more than one word categories, such as noun and verb or adjective and verb. The deterministic
parsing network is shown in figure.
Figure A deterministic Network

Non-Deterministic Parsing
The non deterministic parsing allows different arcs to be labeled with the some test. Thus, they can
uniquely make the choice about the next arc to be taken. In non deterministic parsing, the back tracking
procedure can be possible. Suppose at some extent of point, the parser does not find the correct word,
then at that stage it may backtracks to some of its previous nodes and then start parsing. But the parser has
to guess about the proper constituent and then backtrack if the guess is later proven to be wrong. So
comparative to deterministic parsing, this procedure may be helpful for a number of sentences as it can
backtrack at any point of state. A non deterministic parsing network is shown in figure.
TRANSITION NETWORK
It is a method to represent the natural languages. It is based on applications of directed graphs and finite
Figure Non-Deterministic Parsing Network
state automata. The transition network can be constructed by the help of some inputs, states and outputs.
A transition network may consist of some states or nodes, some labeled arcs from one state to the next
state through which it will move. The arc represents the rule or some conditions upon which the transition
is made from one state to another state. For example, a transition network is used to recognize a sentence
consisting of an article, a noun, an auxiliary, a verb, an article, a noun would be represented by the
transition network as follows.
Figure Transition Network

The transition from N1 to N2 will be made if an article is the first input symbol. If successful, state N2 is
entered. The transition from N2 to N3 can be made if a noun is found next. If successful, state N3 is
entered. The transition from N3 to N4 can be made if an auxiliary is found and so on. Suppose consider a
sentence “A boy is eating a banana”. So if the sentence is parsed in the above transition network then,
first ‘A’ is an article. So successful transition to the node N1 to N2. Then boy is a noun (so N2 to N3), “is”
is an auxiliary (N5 to N6) and finally “banana” is a noun (N6 to N7) is done successfully. So the above
sentence is successfully parsed in the transition network.
TYPES OF TRANSITION NETWORK
There are generally two types of transition networks like
1. Recursive Transition networks (RTN)
2. Augmented Transition networks (ATN)
Let us focus on these two transition networks and their structure for parsing a sentence.
Recursive Transition Networks (RTN)
RTNs are considered as development for finite state automata with some essential conditions to take the
recursive complexion for some definitions in consideration. A recursive transition network consists of
nodes (states) and labeled arcs (transitions). It permits arc labels to refer to other networks and they in
turn may refer back to the referring network rather than just permitting word categories. It is a modified
version of transition network. It allows arc labels that refer to other networks rather than word category. A
recursive transition network can have 5 types of arcs (Allen’s, JM’s) like
1) CAT: Current word must belong to category.
2) WORD: Current word must match label exactly.
3) PUSH: Named network must be successfully traversed.
4) JUMP: Can always be traversed.
5) POP: Can always be traversed and indicates that input string has been accepted by the network.
In RTN, one state is specified as a start state. A string is accepted by an RTN if a POP arc is reached and
all the input has been consumed. Let us consider a sentence “The stone was dark black”.
Here The: ART
Stone: ADJ NOUN
Was: VERB
Dark: ADJ
Black: ADJ NOUN

The RTN structure is given in figure
Figure RTN Structure
So we can parse the sentence through the RTN structure as follows.
Current Word Return Arc Backup States
S 1 ------- NP ------------------
NP 1 S1 ART ------------------
NP1 2 S1 NOUN (NP1, 3, S1)
NP2 3 S1 POP (NP1, 3, S1) (NP2, 4, S1)
S1 3 ------- VERB (NP1, 3, S1) (NP2, 4, S1)
S2 4 ------- NP (NP1, 3, S1) (NP2, 4, S1)
NP 4 S3 NONE (NP1, 3, S1) (NP2, 4, S1)
NP1 3 S1 NOUN (NP2, 4, S1)
NP2 4 S1 POP (NP2, 4, S1)
S1 4 ------- VERB (NP2, 4, S1)
S2 5 ------- NP (NP2, 4, S1)
NP 5 S3 ADJ (NP2, 4, S1)
NP1 6 S3 NOUN (NP2, 4, S1) (NP1, 7, S3)
NP2 7 S3 POP (NP2, 4, S1) (NP1, 7, S3)
NP2 7 S3 POP (NP2, 4, S1) (NP1, 7, S3)
S3 7 ------- POP (NP2, 4, S1) (NP1, 7, S3)
Finally as there are no words left so the parse is successful.
Also there is an another structure of RTN is described by William Woods (1970) is illustrated in figure.
He described the total RTN structure into three parts like sentence (S), Noun Phrase (NP), Preposition
Phrase (PP).

Figure RTN Structure
The number of sentences accepted by an RTN can be extended if backtracking is permitted when a failure
occurs. This requires that states having alternative transitions be remembered until the parse progresses
past possible failure points. In this way, if a failure occurs at some point, the interpreter can backtrack and
try alternative paths. The disadvantage with this approach is that parts of a sentence may be parsed more
than time resulting in excessive computations. During the traversal of an RTN, a record must be
maintained of the word position in the input sentence and the current state and return nodes to be used as
return points when control has been transformed to a lower level network.
Augmented Transition Network (ATN)
An ATN is a modified transition network. It is an extension of RTN. The ATN uses a top down parsing
procedure to gather various types of information to be later used for understanding system. It produces the
data structure suitable for further processing and capable of storing semantic details. An augmented
transition network (ATN) is a recursive transition network that can perform tests and take actions during
arc transitions. An ATN uses a set of registers to store information. A set of actions is defined for each arc
and the actions can look at and modify the registers. An arc may have a test associated with it. The arc is
traversed (and its action) is taken only if the test succeeds. When a lexical arc is traversed, it is put in a
special variable (*) that keeps track of the current word. The ATN was first used in LUNAR system. In

ATN, the arc can have a further arbitrary test and an arbitrary action. The structure of ATN is illustrated
in figure. Like RTN, the structure of ATN is also consisting of the substructures of S, NP and PP.
Figure ATN Structure
The ATN collects the sentence features for further analysis. The additional features that can be captured
by the ATN are; subject NP, the object NP, the subject verb agreement, the declarative or interrogative
mood, tense and so on. So we can conclude that ATN requires some more analysis steps compared to that
of RTN. If these extra analysis tests are not performed, then there must some ambiguity in ATN. The
ATN represents sentence structure by using a slot filter representation, which reflects more of the
functional role of phrases in a sentence. For example, one noun phrase may be identified as “subject”
(SUBJ) and another as the “object” of the verb. Within noun phrases, parsing will also identify the
determiner structure, adjectives, the noun etc. For the sentence “Ram ate an apple”, we can represent as in
figure.

Figure Representation of sentence in ATN
The ATN maintains the information by having various registers like DET, ADJ and HEAD etc. Registers
are set by actions that can be specified on the arcs. When the arc is followed, the specified action
associated with it is executed. An ATN can recognize any language that a general purpose computer can
recognize. The ATNs have been used successfully in a number of natural language systems as well as
front ends for databases and expert systems.
Module 3
LEARNING
Learning process is the basis of knowledge acquisition process. Knowledge acquisition is the expanding
the capabilities of a system or improving its performance at some specified task. So we can say
knowledge acquisition is the goal oriented creation and refinement of knowledge. The acquired
knowledge may consist of various facts, rules, concepts, procedures, heuristics, formulas, relationships or
any other useful information. Knowledge can be acquired from various sources like, domain of interests,
text books, technical papers, databases, reports. The terms of increasing levels of abstraction, knowledge
includes data, information and Meta knowledge. Meta knowledge includes the ability to evaluate the
knowledge available, the additional knowledge required and the systematic implied by the present rules.
Learning involves generalization from experience. Computer system is said to have learning if it is able to
not only do the “repetition of same task” more effectively, but also the similar tasks of the related domain.
Learning is possible due to some factors like the skill refinement and knowledge acquisition. Skill
refinement refers to the situation of improving the skill by performing the same task again and again. If
machines are able to improve their skills with the handling of task, they can be said having skill of
learning. On the other hand, as we are able to remember the experience or gain some knowledge by

The image part with relationship ID rId115 was not found in the file.
handling the task, so we can improve our skill. We would like our learning algorithms to be efficient in
three respects:
(1) Computational: Number of computations during training and during recognition.
(2) Statistical: Number of examples required for good generalization, especially labeled data.
(3) Human Involvement: Specify the prior knowledge built into the model before training.
A similar machine learning architecture is given in figure .
Figure Architecture of Machine Learning
Design of learning element is dictated by the followings.
(1) What type of performance element is used?
(2) Which functional component is to be learned?
(3) How that functional component is represented?
(4) What kind of feedback is available?
(5) How can be compared between the existing feedbacks with the new data?
(6) What are the levels of comparisons? Etc.
Any system designed to create new knowledge and thereby improve its performance must include a set of
data structures that represents the system’s present level of expertise and a task algorithm that uses the

rules to guide the system’s problem solving activity. The architecture of a general learning procedure is
given in figure .
Figure A general Learning Procedure
Hence the inputs may be any types of inputs, those are executed for solution of a problem. Those inputs
are processed to get the corresponding results. The learning element learns some sort of knowledges by
the knowledge acquisition techniques. The acquired knowledge may be required for a same problem in
future, for which that problem can be easily solved.
Every learning model must contain implicit or explicit restrictions on the class of functions that can learn.
Among the set of all possible functions, we are particularly interested in a subset that contains all the
tasks involved in intelligent behaviour. Examples of such tasks include visual perception, auditory
perception, planning, control etc. The set does not just include specific visual perception tasks, but the set
of all the tasks that an intelligent agent should be able to learn. Although we may like to think that the
human brain is some what general purpose, it is extremely restricted in its ability to learn high
dimensional functions.
CLASSIFICATION OF LEARNING
The process of learning may be of various types. One can develop learning taxonomies based on the type
of knowledge representation used (predicate calculus, rules, frames, scripts etc), the type of knowledge
learned (game playing, problem solving) or by areas of application (medical diagnosis, engineering etc).
Generally learning may be of two types like single agent learning and multi-agent learning. A general
architecture of learning process is given figure .
The image part with relationship ID rId116 was not found in the file.

Figure Learning Classification
Single Agent Learning
Over the last four decades, machine learning’s primary interest has been single agent learning. Single
agent learning involves improving the performance or increasing the knowledge of a single agent. An
improvement in performance or an increase in knowledge allows the agent to solve past problems with
better quality or efficiency. An increase in knowledge may also allow the agent to solve new problems.
An increase in performance is not necessarily due to an increase in knowledge. It may be brought about
simply by rearranging the existing knowledge or utilizing it in a different manner. Single agent learning
systems may be classified according to their underlying learning strategies. These strategies are classified
as follows.
Learning
Single Agent Learning Multi Agent Learning
Control Learning
Organization
Learning
Communication
Learning
Group observation and
discovery Learning
Rote or Base Learning
Learning from instruction
Learning by deduction
Learning by Analogy
Learning from examples
Learning from observation
and discovery
Learning by induction
Learning from advices
Learning by clustering

Rote Learning
This strategy does not require the learning system to transform or infer knowledge. It is the simplest form
of learning. It requires the least amount of inference and is accomplished by simply copying the
knowledge in the same form that it will be used directly into the knowledge base. It includes learning my
imitation, simple memorization and learning by being performed. For example we may use this type of
learning when we memorize multiplication tables. In this method we store the previous computed values,
for which we do not have for recomputed them later. Also we can say rote learning is one type of existing
or base learning. For example, in our childhood, we have the knowledge that “sun rises in the east”. So in
our later stage of learning we can easily memorize the thing. Hence in this context, a system may simply
memorize previous solutions and recall them when confronted with the same problem. Generally access
of stored value must be faster than it would be to recompute. Methods like hashing, indexing and sorting
can be employed to enable this. One drawback of rote learning is it is not very effective in a rapidly
changing environment. If the environment does change then we must detect and record exactly what has
changed. Also this technique must not decrease the efficiency of the system. We must be able to decide
whether it is worth storing the value in the first place.
Learning from Instruction
This strategy also known as learning by being told or learning by direct instruction. It requires the
learning system to select and transform knowledge into a usable form and then integrate it into the
existing knowledge of the system. It is a more complex form of learning. This learning technique requires
more inference than rote learning. It includes learning from teachers and learning by using books,
publications and other types of instructions.
Learning by Deduction
This process is accomplished through a sequence of deductive inference steps using known facts. From
the known facts, new facts or relationships are logically derived. Using this strategy, the learning system
derives new facts from existing information or knowledge by employing deductive inference. It requires
more inferences than other techniques. The inference method used is a deductive type, which is a valid
form of inference. For example we can say x is the cousin of y if we have the knowledge of x’s and y’s
parents and the rules for cousin relationships. The learner draws deductive inferences from the knowledge
and reformulates them in the form of useful conclusions which preserve the information content of the
original data. Deductive learning includes knowledge reformulation, compilation and organizational
procedures that preserve the truth of the original formulation.
Learning by Analogy
It is a process of learning a new concept or solution through the use of similar known concepts or
solutions. We make frequent use of analogical learning. The first step is inductive inference, required to
find a common substructure between the problem domain and one of the analogous domains stored in the

learner’s existing knowledge base. This form of learning requires the learning system to transform and
supplement its existing knowledge from one domain or problem area into new domain. This strategy
requires more inferencing by the learning system than previous strategies. Relevant knowledge must be
found in the systems existing knowledge by using induction strategies. This knowledge must then be
transformed to the new problem using deductive inference. Example of learning by analogy may include
the driving technique of vehicles. If we know the driving procedure of a bike, then when we will drive a
car then some sort of previous learning procedures we may employ. Similarly for driving a bus or truck,
we may use the procedure for driving a car.
Learning from Examples
In this process of learning it includes the learning through various interactive and innovative examples.
This strategy, also called concept acquisition. It requires the learning system to induce general class or
concept descriptions from examples. Since the learning system does not have prior or analogous
knowledge of the concept area, the amount of inferencing is greater than both learning by deduction and
analogy. For solving a newly designed problem we may use its corresponding old examples.
Learning from Observations and Discovery
Using this strategy, the learning system must either induce class descriptions from observing the
environment or manipulate the environment to acquire class descriptions or concepts. This is an
unsupervised learning technique. It requires the greatest amount of inferencing among all of the different
forms of learning. From an existing knowledge base, some new forms of discovery of knowledge may
formed. The learning discovery process is very important in the respect of constructing new knowledge
base.
Learning by Induction
Inductive learning is the system that tries to induce a general rule based on observed instances. In other
words, the system tries to infer an association between specific inputs and outputs. In general, the input of
the program is a set of training instance where the output is a method of classifying subsequent instance.
For example, the input of the program may be color of types of fruits where the output may be the types
of fruits those are useful for protein. Induction method involves the learning by examples,
experimentation, observation and discovery. The search spaces encountered in learning tend to be
extremely large, even by the standards of search based problem solving.
This complexity of problems is cleared by choosing a problem among the different generalizations
supported by any given training data. Inductive bias refers to any method a learning program uses to
constrain the space of possible generalizations. A no. of inductive learning algorithms have been
developed like Probably Approximately Correct (PAC), Version spaces etc. Probably Approximately
Correct learning was proposed concerning the situation that cannot be deductive. Approximately correct

is recognized whenever the program can get most of the problems right. In order to increase performance
of the program, learning algorithms should restrict the size of hypothesis space. On the other hand, the
goal of version space is to produce a description that uses only positive examples. The program in
practice, produce a description of all acceptable concepts. In detail, we may conclude that there are two
sets of concepts that are produced during learning. Firstly, the most specific concept describes what the
target set should be. Secondly, the least specific concept describes what should not be in the target group.
Generally inductive learning is frequently used by humans. This form of learning is more powerful than
the others. We use this learning when we formulate a general concept after seeing a number of instances.
For example, we can say the taste of sugar is sweet if we have the knowledge about sweetness.
Learning from Advices
In this process we can learn through taking advice from others. The idea of advice taking learning was
proposed in early 1958 by McCarthy. In our daily life, this learning process is quite common. Right from
our parents, relatives to our teachers, when we start our educational life, we take various advices from
others. All most all the initial things and all type of knowledges we acquire through the advices of others.
We know the computer programs are written by programmers. When a programmer writes a computer
program he or she gives many instructions to computer to follow, the same way a teacher gives his/her
advice to his students. The computer follows the instructions given by the programmer. Hence, a kind of
learning takes place when computer runs a particular program by taking advice from the creator of the
program.
Learning by Clustering
This process is similar to the inductive learning. Clustering is a process of grouping or classifying objects
on the basis of a close association or shared characteristics. The clustering process is essentially required
in a learning process in which similarity patterns are found among a group of objects. The program must
discover for itself the natural classes that exist for the objects, in addition to a method for classifying
instances. AUTOCLASS (Cheeseman et al., 1988) is one program that accepts a number of training cases
and hypothesizes a set of classes. For any given case, the program provides a set of probabilities that
predict into which classes the case is likely to be fall.
Multi Agent Learning
Distributed artificial intelligence (DAI) systems solve problems using multiple, cooperative agents. In
these systems, control and information are often distributed among the agents. This reduces the
complexity of each agent and allows agents to work in parallel and increases problem solving speed. Also
each agent has resource limitations which could limit the ability of a single agent system to solve large,
complex problems. Allowing multiple agents to work on these types of problems may be the only way to
realistically solve them. In general, multiple agent learning involves improving the performance of the

group of agents as a whole or increasing the domain knowledge of the group. It also includes increasing
communication knowledge. An increase in communication knowledge can lead to an increase in
performance by allowing the agents to communicate in a more efficient manner. In the context of
improving the performance of a group of agents, allowing individual agents to improve their performance
may not be enough to improve the performance of the group. To apply learning to the overall group
performance, the agents need to adapt and learn to work with the each other. The agents may not need to
learn more about the domain, as in the traditional sense of machine learning, to improve group
performance. In fact to improve the performance of the group, the agents may only need to learn to work
together and not necessarily improve their individual performance. In addition, not all the agents must be
able to learn or adapt to allow the group to improve.
Control Learning
Learning and adapting to work with other agents involves adjusting the control of each agent’s problem
solving plan. Different tasks may have to be solved in a specific sequence. If the tasks are assigned to
separate agents, the agents must work together to solve the tasks. Learning which agents are typically
assigned different types of tasks will allow each agent to select other agents to work with on different
tasks. Teams can be formed based on the type of task to be solved. Some of the issues involved are the
type, immediacy and importance of task, as well as each agent’s task solving ability, capability, reliability
and past task assignments. Each team member’s plan would be adjusted according to the other agent’s
plans.
Organization Learning
Learning what type of information and knowledge each agent possesses allows for an increase in
performance by specifying the long term responsibilities of each agent. By assigning different agents
different responsibilities, the group of agents can improve group performance by providing a global
strategy. Organizing the responsibilities reduces the working complexity of each agent.
Communication Learning
Learning what type of information, knowledge reliability and capability each agent possesses allows for
an increase in performance by allowing improved communication. Directly addressing the best agent for
needed information or knowledge allows for more efficient communication among he agents.
Group Observation and Discovery Learning
Individual agents incorporate different information and knowledge. Combining this differing information
and knowledge may assist in the process of learning new class descriptions or concepts that could not
have been learned by the agents separately. This type of learning is more effective than the others. The
observation towards the procedure will be focused by a group of agents. When a group of different

visions will reach, at that point of view a newly interactive procedure will be found out; which is the
discovery of all the agents.
Explanation based Learning
Explanation based learning has ability to learn from a single training instance. Instead of taking more
examples the explanation based learning is emphasized to learn a single, specific example. For example,
consider the Ludoo game. In a Ludoo game, there are generally four colors of buttons. For a single color
there are four different squares. Suppose the colors are red, green, blue and yellow. So maximum four
members are possible for this game. Two members are considered for one side (suppose green and red)
and other two are considered for another side (suppose blue and yellow). So for any one opponent the
other will play his game. A square sized small box marked by symbols one to six is circulated among the
four members. The number one is the lowest number and the number six is the highest for which all the
operations are done. Always any one from the 1st
side will try to attack any one member in the 2nd
side
and vice versa. At any instance of play the players of one side can attack towards the players of another
side. Likewise, all the buttons may be attacked and rejected one by one and finally one side will win the
game. Here at a time the players of one side can attack towards the players of another side. So for a
specific player, the whole game may be affected. Hence we can say that always explanation based
learning is concentrated on the inputs like a simple learning program, the idea about the goal state, the
idea about the usable concepts and a set of rules that describes relationships between the objects and the
actions.
Explanation based generalization (EBG) is an algorithm for explanation based learning, described in
Mitchell at al. (1986). It has two steps first, explain method and secondly, generalize method. During the
first step, the domain theory is used to prune away all the unimportant aspects of training examples with
respect to the goal concept. The second step is to generalize the explanation as far as possible while still
describing the goal concept. Consider the problem of learning the concept bucket. We want to generalize
from a single example of a bucket. At first collect the following informations.
1. Input Examples:
Owner object, X ? has part object, Y ? isobject, Deep ? Color Object, Green
? … … Where Y is any thin material
2. Domain Knowledge:
is a, Deep ? has part a, b ? is ab, handle liftable a
has part a, b ? is a b, Bottom ? is b, elat Stable a
has part a, b ? is a b, Y ? is b, Upward pointing Open vessel a
3. Goal: Bucket
B is a bucket if B is liftable, stable and open-vessel.

4. Description of Concept: These are expressed in purely structural forms like Deep, Flat, rounded
etc.
Figure An explanation of BUCKET Object
Given a training example and a functional description, we want to build a general structural description of
a bucket. In practice, there are two reasons why the explanation based learning is important.
GENETIC ALGORITHM
Genetic algorithms are based on the theory of natural selection and work on generating a set of random
solutions and making them compete in an area where only the fittest survive. Each solution in the set is
equivalent to a chromosome. Genetic algorithm learning methods are based on models of natural adaption
and evolution. These learning methods improve their performance through processes which model
population genetics and survival of the fittest. In the field of genetics, a population is subjected to an
environment which places demands on the members. The members which adapt well are selected for
matting and reproduction. Generally genetic algorithm uses three basic genetic operators like
reproduction, crossover and mutation. These are combined together to evolve a new population. Starting
from a random set of solutions the algorithm uses these operators and the fitness function to guide its
search for the optimal solution. The fitness function guesses how good the solution in question is and
provides a measure to its capability. The genetic operators copy the mechanisms based on the principles
of human evolution. The main advantage of the genetic algorithm formulation is that fairly accurate
results may be obtained using a very simple algorithm. The genetic algorithm is a method of finding a
good answer to a problem, based on the feedback received from its repeated attempts at a solution. The

fitness function is a judge of the GA’s attempts for a problem. GA is incapable to derive a problem’s
solution, but they are capable to know from the fitness function.
Genetic algorithm starts with a fixed size population of data structure which is used to perform some
given tasks. After the structure performs the given task or problem, they are rated on their performance by
some utility value and a new generation of data structure then created. The new generation is created by
mapping with the high performing structure to produce offspring. The offsprings or the children and their
parents are retained for the next generation while the poorer performers are not included. Mutations are
also performed on the best programming structures to ensure that the full space of possible structure is
reachable. This process is repeated for a number of generations until the resultant population consists of
only the highest performing structures. Matting between two strings is accomplished with the crossover
operation which randomly selects a bit position in the eight bit string and concatenates the head of one
parent to the tail of the second parent to produce the off string. Inversion is another type of genetic
operation which is applied to a single string.
The GA goes through the following cycle: Generate, Evaluate, Assignment of values, Mate and Mutate.
One criteria is to let the GA run for a certain number of cycles. A second one is to allow the GA to run
until a reasonable solution is found. Also mutation is a operation, which is used to ensure that all
locations of the rule space are reachable, that every potential rule in the rule space is available for
evaluation. The mutation operator is typically used only infrequently to prevent random wondering in the
search space. Let us focus on the genetic algorithm described as follows.
Step 1:
Generate the initial population.
Step 2:
Calculate the fitness function of each individuals.
Step 3:
Some sort of performance utility values or the fitness values are assigned to individuals.
Step 4:
New populations are generated from the best individuals by the process of selection.
Step 5:
Perform the crossover and mutation operation.
Step 6:

Replace the old population with the new individuals.
Step 7:
Perform step-2 until the goal is reached.
In its simplest form, the standard genetic algorithm is a method of stochastic optimization for discrete
programming problems of the form.
In this case f: Ω R is called the fitness function and the n-dimensional binary vectors in Ω are called
strings. The most noticeable difference between the standard genetic algorithm and the methods of
optimization is that at each stage of the computation, genetic algorithms maintain a collection of samples
from the search space Ω rather than a single point. This collection of samples is called a population of
strings. To start the genetic search, an initial population of say, B binary strings P0 gP:, P9 … . Phi j
Ω; each with n bits, is created. Usually, this initial population is created randomly because it is not
known apriori where the globally optimal strings in Ω are likely to be found. If such information is given
though it may be used to bias the initial population towards the most promising regions ofΩ. From this
initial population, subsequent population P1, P2 … … … … Pm … … .. will be computed by employing
the three genetic operators of selection, crossover and mutation.
Applications and Advantages of Genetic Algorithm
Some of the applications and characteristics of genetic algorithm as described as following.
→ GA is a randomized search and optimization technique guided by the principle of natural genetic
systems.
→ The GA is being applied to a wide range of optimization and learning problems in many domains.
→ GAs also lend themselves well to power system optimization problems.
→ GAs solve problems using principles inspired by natural population genetics.
→ GAs can provide globally optimal solutions.
→ GAs work with a coding of the parameter set, not the parameters themselves. Therefore, they can
easily handle integral variables.
→ GAs use probabilistic transition rules, non deterministic rules.
→ GAs are used to solve the problems like job shop problem, optimization problems, skill based
employee allocation problems, scheduling of job shop problems etc.
Maximize f P
subject to P ^ Ωg0, 1i;

→ GA performs always multidirectional search by maintaining a population of potential solutions.
NEURAL NETWORK
A neural network consists of inter connected processing elements called neurons that work together to
produce an output function. The output of a neural network relies on the cooperation of the individual
neurons within the network to operate. Well designed neural networks are trainable systems that can often
“learn” to solve complex problems from a set of exemplars and generalize the “acquired knowledge” to
solve unforeseen problems, i.e. they are self-adaptive systems. A neural network is used to refer to a
network of biological neurons. A neural network consists of a set of highly interconnected entities called
nodes or units. Each unit accepts a weighted set of inputs and responds with an output.
Mathematically let I I:, I9, … … I; represent the set of inputs presented to the unit U. Each input has
an associated weight that represents the strength of that particular connection. Let
W W:, W9, … … W; represent the weight vector corresponding to the input vector X. By applying to
V, these weighted inputs produce a net sum at U given by
S SUM W= K I=
A neural network is first and foremost a graph, with patterns represented in terms of numerical values
attached to the nodes of the graph and transformations between patterns achieved via simple message-
passing algorithms. The graph contains a number of units and weighted unidirectional connections
between them. The output of one unit typically becomes an input for another. There may also be units
with external inputs and outputs. The nodes in the graph are generally distinguished as being input nodes
or output nodes and the graph as a whole can be viewed as a representation of a multivariate functions
linking inputs to outputs. Numerical values (weights) are attached to the links of the graphs,
parameterizing the input/ output function and allowing it to be adjusted via a learning algorithm. A
broader view of a neural network architecture involves treating the network as a statistical processor
characterized by making particular probabilistic assumptions about data. Figure illustrates one example
of a possible neural network structure.

Figure An example of a Neural network structure
Patterns appearing on the input nodes or the output nodes of a network are viewed as samples from
probability densities and a network is viewed as a probabilistic model that assigns probabilities to
patterns. Biologically, we can also define a neuron. The human body is made up of a vast array of living
cells. Certain cells are interconnected in a way that allows them to communicate pain or to actuate fibres
or tissues. Some cells control the opening and closing of minuscule valves in the veins and arteries. These
specialized communication cells are called neurons. Neurons are equipped with long tentacle like
structures that stretch out from the cell body, permitting them to communicate with other neurons. The
tentacles that take in signals from other cells and the environment itself are called dendrites, while the
tentacles that carry signals from the neuron to other cells are called axons.
Dendrites
Nucleus
Cell Body Mitochondria
Axons
Figure A Neuron

FEATURES OF ARTIFICIAL NETWORK (ANN)
Artificial neural networks may by physical devices or simulated on conventional computers. From a
practical point of view, an ANN is just a parallel computational system consisting of many simple
processing elements connected together in a specific way in order to perform a particular task. There are
some important features of artificial networks as follows.
(1) Artificial neural networks are extremely powerful computational devices (Universal computers).
(2) ANNs are modeled on the basis of current brain theories, in which information is represented by
weights.
(3) ANNs have massive parallelism which makes them very efficient.
(4) They can learn and generalize from training data so there is no need for enormous feats of
programming.
(5) Storage is fault tolerant i.e. some portions of the neural net can be removed and there will be only a
small degradation in the quality of stored data.
(6) They are particularly fault tolerant which is equivalent to the “graceful degradation” found in
biological systems.
(7) Data are naturally stored in the form of associative memory which contrasts with conventional
memory, in which data are recalled by specifying address of that data.
(8) They are very noise tolerant, so they can cope with situations where normal symbolic systems would
have difficulty.
(9) In practice, they can do anything a symbolic/ logic system can do and more.
(10) Neural networks can extrapolate and intrapolate from their stored information. The neural networks
can also be trained. Special training teaches the net to look for significant features or relationships of
data.
TYPES OF NEURAL NETWORKS
There are a number of models of neural networks have been developed, each defined at a different level
of abstraction and trying to model different aspects of neural systems. They range from models of the
short term behaviour of individual neurons, through models of how the dynamics of neural circuitry arise
from interactions between individual neurons, to models of how behaviour can arise from abstract neural
modules that represent complete sub systems. Some neural networks have been illustrated below.
Single Layer Network
A single layer neural network consists of a set of units organized in a layer. Each unit U; receives a
weighted input Imwith weight Wm;. Figure shows a single layer neural network with j inputs and n outputs.

Figure A single Layer neural Network
Let I ni:, i9 … … . imo be the input vector and let the activation function f be simply, so that the activation
value is just the net sum to a unit. The jxn weight matrix is calculated as follows.
W p
W:: W:9 … … … W:;
W9: W99 … … . . W9;
q q q
Wm: Wm9 … … . . Wm;
r
Thus the output Ox at unit Ux is
Os nW:t, W9t … … . . Wm=o u
I:
I9
q
Im
v
Multilayer Network
A multilayer network has two or more layers of units, with the output from one layer serving as input to
the next. Generally in a multilayer network there are 3 layers present like, input layer, output layer and
hidden layer. The layer with no external output connections are referred to as hidden layers. A multilayer
neural network structure is given in figure.

Figure A multilayer neural network
Any multilayer system with fixed weights that has a linear activation function is equivalent to a single
layer linear system, for example, the case of a two layer system. The input vector to the first layer is Iwthe
output O W: K I and the second layer produces output O9 W9 K O.HenceO9 W9 K W: K I
W9 K W: K I
So a linear system with any number n of layers is equivalent to a single layer linear system whose weight
matrix is the product of the n intermediate weight matrices. A multilayer system that is not linear can
provide more computational capability than a single layer system. Generally multilayer networks have
proven to be very powerful than single layer neural network. Any type of Boolean function can be
implemented by such a network. At the output layer of a multilayer neural network the output vector is
compared to the expected output. If the difference is zero, no changes are made to the weights of
connections. If the difference is not zero, the error is calculated and is propagated back through the
network.
Feed Forward neural network
The neural networks consist of multiple layers of computational units, usually interconnected in a feed
forward way. The feed forward neural networks are the first, simplest type of artificial neural networks
devised. In this network, the information moves in only one direction, forward from the input nodes,
through the hidden nodes and to the output nodes. There are no cycles or loops in the network. In other
way we can say the feed forward neural network is one that does not have any connections from output to
input. All inputs with variable weights are connected with every other node. A single layer feed forward
network has one layer of nodes, whereas a multilayer feed forward network has multiple layers of nodes.
The structure of a feed forward multilayer network is given in figure.

Figure Multilayer Feed Forward Neural Network
Data are introduced into the system through an input layer. This is followed by processing in one or more
intermediate (hidden layers). Output data emerge from the network’s final layer. The transfer functions
contained in the individual neurons can be almost anything. The input layer is also called as Zeroth layer,
of the network serves to redistribute input values and does no processing. The output of this layer is
described mathematically as follows.
Ox
y
i Where m 1,2 … . Ny
Ny
representes the no. of neurons in the input or zeroth layer.
The input to each neuron in the first hidden layer in the network is a summation all weighted connections
between the input or Zeroth layer and the neuron in the first hidden layer. We will write the weighted sum
as net sum or net input. We can write the net input to a neuron from the first layer as the product of that
input vector im and weight factor wm plus a bias term θ. The total weighted input to the neuron is a
summation of these individual input signals described as follows.
net sum z WX θ
|
}:
Where N represents the number of neurons in the input layer.
The net sum to the neuron is transformed by the neuron’s activation or transfer function, f to produce a
new output value for the neuron. With back propagation, this transfer function is most commonly either a
sigmoid or a linear function. In addition to the net sum, a bias term θ is generally added to offset the
input. The bias is designed as a weight coming from a unitary valued input and denoted as W0. So, the
final output of the neuron is given by the following equation.

Output f net sum
f ~ z wi θ
|
}:

f €z wi
y
wy
|
}:
‚
But one question may arise in reader’s mind. Why we are using the hidden layer between the input and
output layer? The answer to this question is very silly. Each layer in a multilayer neural network has its
own specific function. The input layer accepts input signals from the outside world and redistributes these
signals to all neurons in the hidden layer. Actually, the input layer rarely includes computing neurons and
thus does not process input patterns. The output layer accepts output signals, or in other words a stimulus
patterns, from the hidden layer and established the output patterns of the entire network. Neurons in the
hidden layer detect the features, the weights of the neurons represent the features hidden in the input
patterns. These features are then used by the output layer in determining the output patterns. With one
hidden layer we can represent any continuous function of the input signals and with two hidden layers
even discontinuous functions can be represented. A hidden layer hides its desired output. Neurons in the
hidden layer cannot be observed through the input/ output behaviour of the network. The desired output
of the hidden layer is determined by the layer itself. Generally, we can say there is no obvious way to
know what the desired output of the hidden layer should be.
Back Propagation neural network
Multilayer neural networks use a most common technique from a variety of learning technique, called the
back propagation algorithm. In back propagation neural network, the output values are compared with the
correct answer to compute the value of some predefined error function. By various techniques the error is
then fed back through the network. Using this information, the algorithms adjust the weights of each
connection in order to reduce the value of the error function by some small amount. After repeating this
process for a sufficiently large number of training cycles the network will usually converge to some state
where the error of the calculation is small.
The goal of back propagation, as with most training algorithms, is to iteratively adjust the weights in the
network to produce the desired output by minimizing the output error. The algorithm’s goal is to solve
credit assignment problem. Back propagation is a gradient-descent approach in that it uses the
minimization of first-order derivatives to find an optimal solution. The standard back propagation
algorithm is given below.
Step1:

Build a network with the choosen number of input, hidden and output units.
Step2:
Initialize all the weights to low random values.
Step3:
Randomly, choose a single training pair.
Step4:
Copy the input pattern to the input layer.
Step5:
Cycle the network so that the activation from the inputs generates the activations in the hidden and output
layers.
Step6:
Calculate the error derivative between the output activation and the final output.
Step7:
Apply the method of back propagation to the summed products of the weights and errors in the output
layer in order to calculate the error in the hidden units.
Step8:
Update the weights attached the each unit according to the error in that unit, the output from the unit
below it and the learning parameters, until the error is sufficiently low.
To derive the back propagation algorithm, let us consider the three layer network shown in figure .

Figure Three layer back-propagation neural network
To propagate error signals, we start at the output layer and work backward to the hidden layer. The error
signal at the output of neuron k at iteration x is defined as
Where is the desired output of neuron k at iteration x.
Generally, computational learning theory is concerned with training classifiers on a limited amount of
data. In the context of neural networks a simple heuristic, called early stopping often ensures that the
network will generalize well to examples not in the training set. There are some problems with the back
propagation algorithm like speed of convergence and the possibility of ending up in a local minimum of
the error function. Today there are a variety of practical solutions that make back propagation in
multilayer perceptrons the solution of choice for many machine learning tasks.
CLUSTERING
Clustering is a division of data into groups of similar objects. Representing the data by fewer clusters
necessarily loses certain fine details, but achieves simplification. It models data by its clusters. Data
modeling puts clustering in a historical perspective rooted in mathematics, statistics and numerical
analysis. From a machine learning perspective clusters correspond to hidden patterns, the search for
clusters in unsupervised learning and the resulting system represents a data concept. From a practical
perspective, clustering plays an outstanding role in data mining applications such as scientific data
exploration, information retrieval and text mining, spatial database applications, web analysis, marketing,
medical diagnostics, computational biology and many others.

Clustering is the subject of active research in several fields such as statistics, pattern recognition and
machine learning. Clustering is the classification of similar objects into different group. We can also
define clustering is the unsupervised learning of a hidden data concept. Besides the term data clustering,
there are a number of terms with similar meanings, including cluster analysis, automatic classification,
numerical taxonomy, and typological analysis.
Types of Clustering
Categorization of clustering algorithms is neither straight forward nor canonical. Data clustering
algorithms can be hierarchical or partitional. Two-way clustering, co-clustering or bi-clustering are the
names for clustering where not only the objects are clustered but also the features of the objects. We
provide a classification of clustering algorithms listed below.
Clustering Algorithms
Different Clustering Algorithms

PATTERN RECOGNITION
Pattern recognition as a field of study developed significantly in the 1960’s. It was very much
interdisciplinary subject, covering developments in the areas of statistics, engineering, artificial
intelligence, computer science, psychology and physiology, among others. Some people entered the field
with a real problem to solve. The large numbers of applications ranging from the classical ones such as
automatic character recognition and medical diagnosis to the more recent ones in data mining (such as
credit scoring, consumer sales analysis and credit card transaction analysis), have attracted considerable
research effort, with many methods developed and advances made. Pattern recognition of the objects into
a lot of categories or classes. It is an integral part in most machine intelligence system built for decision
making. The nature of the pattern recognition is engineering. But the final aim of pattern recognition is to
design machines to solve the gap between application and theory. It is a process of identifying a stimulus.
This process is often accomplished with incomplete of ambiguous information. The basic model of
pattern classifiers is shown in figure.
Figure A common Pattern Classifier
A complete pattern recognition system consists of:
(a) A sensor: It gathers the information to be classified.
(b) A feature selector or Extractor: Feature selection is the process of selecting a subset of a given set
of variables. The feature extractor mechanism takes a possible non linear combination of the
original variables to form new variables.
(c) A classifier: It classifies or describes the observations relying on the extracted features.
To understand the problem of designing a pattern recognition system, we must understand the problems
that each of these components must solve. Different components of the pattern recognition system are
sensing, segmentation, feature extraction, classification, post processing. The input to a pattern
recognition system is some kind of a transducer, such as camera or a microphone array. Sensing is used to
eliminate the noise. A sensor converts images or sounds or other physical into signal data. The segmentor
isolates sensed objects from the back ground or from other objects. A feature extractor measures object
properties that are useful for classification. The goal of feature extractor is to characterize on object to be
recognized by measurements whose values are very similar for objects in the same category and for
objects in different categories very different. The task of feature extractor is domain dependent and
requires the knowledge of the domain. The task of the classifier component proper of a full system is to

use the feature vector provided by the feature extractor to assign the object to a category. The classifier
uses the features to assign the sensed object to a category. The post processor uses the output of the
classifier to decide on the recommended action.
Figure Components of a Pattern Recognition System
The design cycle of a pattern recognition system entails the repetition of a number of different activities
training and evaluation. A structure of the design cycle of a pattern recognition system is illustrated in
figure .

Figure Design cycle of a Pattern Recognition System
Data collection can account for large part of the cost of developing a pattern recognition system. The
selection of some separate distinguishable features depends on the characteristics of the problem domain.
In selecting features, we would like to find features that are simple to extract, invariant to irrelevant
transformation and useful to discriminating patterns in different categories. The selections of models are
done by the different descriptions, which are in mathematical form. The training process uses some or all
of the data to determine the system parameters. Different types of training protocols are stochastic, batch
and online. Evaluation is needed to measure the performance of the system and to identify the need for
improvements in its components. There are two main divisions of classification in pattern recognition like
supervised classification and unsupervised classification. In supervised classification, we have a set of
data samples (each consisting of measurements on a set of variables) with associated labels, the class
types. These are used as exemplars in the classifier design. In unsupervised classification, the data are not
labeled and we seek to fine groups in the data and the features that distinguish one group from another.
EXPERT SYSTEM:
An expert system may be viewed as a computer simulation of a human expert. It can also be
defined as a computer program that simulates the judgment and behaviour of a human or an
organization that has expert knowledge and experience in a particular field. Typically such a
system contains a knowledge base containing accumulated experience and a set of rules for
applying the knowledge base to each particular situation that is described to the program. Expert

systems also use human knowledge to solve problems that normally would require human
intelligence. These expert systems represent the expertise knowledge as data or rules within the
computer. These rules and data can be called upon when needed to solve problems. Books and
manual guides have a tremendous amount of knowledge but a human has to read and interpret
the knowledge for it to be used.
A system that uses human knowledge captured in a computer to solve problems that ordinarily
require human expertise. A computer program designed to model the problem solving ability of a
human expert. Expert systems make extensive use of specialized knowledge to solve problems at
the level of a human expert. An expert is a person who has expertise in a certain area i.e. the
expert has knowledge or special skills that are not known or available to most people. An expert
can solve problems that most people cannot solve them much more efficiently. Thus expert
system technology may include special expert system languages, programs and hardware
designed to aid in the development and execution of expert systems. The knowledge in expert
systems may be either expertise or knowledge that is generally available from books, magazines
and knowledgeable persons.
DIFFERENCE BETWEEN EXPERT SYSTEM AND CONVENTIONAL SYSTEM
The principle distinction between expert systems and traditional problem solving programs is the
way in which the problem related expertise is coded. In conventional applications, problem
expertise is encoded in both program and data structures. In the expert system approach all of the
problem related expertise is encoded in data structures only, none is in programs. Generally in
expert systems, the use of knowledge is vital. But in conventional system data is used more
efficiently than knowledge. Conventional systems are not capable of explaining a particular
conclusion for a problem. These systems try to solve in a straight forward manner. But expert
systems are capable of explaining how a particular conclusion is reached and why requested
information is needed during a process. However, the problems are solved more efficiently than
a conventional system by an expert system. Generally in an expert system, it uses the symbolic
representations for knowledge i.e. the rules, different forms of networks, frames, scripts etc. and
performs their inference through symbolic computations. But conventional systems are unable to
express these terms. They just simplify the problems in a straight forward manner and are
incapable to express the “how, why” questions. Also the problem solving tools those are present
in expert system are purely absent in conventional systems. The various types of problems are
always solved by the experts in an expert system. So the solution of the problem is more accurate
than a conventional system.

THE DEVELOPMENT PROCESS OF AN EXPERT SYSTEM
By the definition, an expert system is a computer program that simulates the thought process of a human
expert to solve complex decision problems in a specific domain. The expert system’s knowledge is
obtained from expert sources which are coded into most suitable form. The process of building an expert
system is called knowledge engineering and is done by a knowledge engineer. The knowledge engineer is
a human with a background in computer science and AI and he knows how to build expert systems. A
knowledge engineer also decides how to represent the knowledge in an expert system and helps the
programmers to write the code. Knowledge engineering is the acquisition of knowledge from a human
expert or any other source. The different stages in the development of an expert system are illustrated in
figure.
Figure Hierarchy of expert system development process
Some latest developments in the expert system area are as follows:
1. Availability of many tools that are designed to expedite the construction of expert system at a reduced
cost.

2. Increased use of expert systems in many tasks ranging from help desks to complex military and space
shuttle applications.
3. Use of multiple knowledge bases.
4. Improvements in knowledge acquisition.
5. Use of the internet to disseminate software and expertise.
6. Increased use of object oriented programming approach in knowledge representation.
7. The multiple use of heuristic knowledge in several applications.
8. Enables the user to think about hypothetical reasoning.
CHARACTERISTICS OF AN EXPERT SYSTEM
The growth of expert system is expected to continue for several years. With the continuing growth, many
new and exciting applications will emerge. An expert system operates as an interactive system that
responds to questions, asks for clarification, makes recommendations and generally aids the decision
making process. Expert system provides expert advice and guidance in a wide variety of activities from
computer diagnosis to delicate medical surgery.
An expert system is usually designed to have the following general characteristics.
1. High level Performance: The system must be capable of responding at a level of competency equal
to or better than an expert system in the field. The quality of the advice given by the system should be
in a high level integrity and for which the performance ratio should be also very high.
2. Domain Specificity: Expert systems are typically very domain specific. For ex., a diagnostic expert
system for troubleshooting computers must actually perform all the necessary data manipulation as a
human expert would. The developer of such a system must limit his or her scope of the system to just
what is needed to solve the target problem. Special tools or programming languages are often needed
to accomplish the specific objectives of the system.
3. Good Reliability: The expert system must be as reliable as a human expert.
4. Understandable: The system should be understandable i.e. be able to explain the steps of reasoning
while executing. The expert system should have an explanation capability similar to the reasoning
ability of human experts.
5. Adequate Response time: The system should be designed in such a way that it is able to perform
within a small amount of time, comparable to or better than the time taken by a human expert to reach
at a decision point. An expert system that takes a year to reach a decision compared to a human
expert’s time of one hour would not be useful.
6. Use symbolic representations: Expert system use symbolic representations for knowledge (rules,
networks or frames) and perform their inference through symbolic computations that closely resemble
manipulations of natural language.

7. Linked with Metaknowledge: Expert systems often reason with metaknowledge i.e. they reason
with knowledge about themselves and their own knowledge limits and capabilities. The use of
metaknowledge is quite interactive and simple for various data representations.
8. Expertise knowledge: Real experts not only produce good solutions but also find them quickly. So,
an expert system must be skillful in applying its knowledge to produce solutions both efficiently and
effectively by using the intelligence human experts.
9. Justified Reasoning: This allows the users to ask the expert system to justify the solution or advice
provided by it. Normally, expert systems justify their answers or advice by explaining their reasoning.
If a system is a rule based system, it provides to the user all the rules and facts it has used to achieve
its answer.
10. Explaining capability: Expert systems are capable of explaining how a particular conclusion was
reached and why requested information is needed during a consultation. This is very important as it
gives the user a chance to access and understand the system’s reasoning ability, thereby improving
the user’s confidence in the system.
11. Special Programming Languages: Expert systems are typically written in special programming
languages. The use of languages like LISP and PROLOG in the development of an expert system
simplifies the coding process. The major advantage of these languages, as compared to conventional
programming languages is the simplicity of the addition, elimination or substitution of new rules and
memory management capabilities. Some of the distinguishing characteristics of programming
languages needed for expert system work are as follows:
a) Efficient mix of integer and real variables.
b) Good memory management procedures.
c) Extensive data manipulation routines.
d) Incremental compilation.
e) Tagged memory architecture.
f) Efficient search procedures.
g) Optimization of the systems environment.
STRUCTURE OF AN EXPERT SYSTEM
The structure of expert systems reflect the knowledge engineers understanding of the methods of
representing knowledge and of how to perform intelligent decision making tasks with the support of a
computer based system. Complex decisions involve intricate combination of factual and heuristic
knowledge. In order for the computer to be able to retrieve and effectively use heuristic knowledge, the
knowledge must be organized in an easily accessible format that distinguishes among data, knowledge
and control structures. For this reason expert systems are organized in three distinct levels like:
a) Knowledge Base: It consists of problem solving rules, procedures and intrinsic data relevant to the
problem domain. The knowledge base constitutes the problem solving rules, facts or intuition that a

human expert might use in solving problems in a given problem domain. The knowledge base is
usually stored in terms of if-then rules. The working memory represents relevant data for the current
problem being solved.
b) Working Memory: It refers to task specific data for the problem under consideration. This is the
dynamic module of the system. It consists of an essential component called database. In general, the
workspace contains a set called rule base, i.e. it contains a set of rules that to be used by a system at a
given moment.
c) Inference Engine: This is a generic control mechanism that applies the axiomatic knowledge in the
knowledge base to the task specific data to arrive at some solution or conclusion. Inference in
production systems is accomplished by a process of chaining through the rules recursively, either in a
forward or in a backward direction until a conclusion is reached.
These three pieces may very well come from different sources. The inference engine, such as VP-Expert,
may come from a commercial vendor. The knowledge base may be a specific diagnostic knowledge base
compiled by a consulting firm, and the problem data may be supplied by the end user. A knowledge base
is the nucleus of the expert system structure. A knowledge base is created by knowledge engineers, who
translate the knowledge of real human experts into rules and strategies. These rules and strategies can
change depending on the prevailing problem scenario. The knowledge base provides the expert system
with the capability to recommend directions for user inquiry. The system also instigates further
investigation into areas that may be important to a certain line of reasoning but not apparent to the user.
The general structure of an expert system is given in figure .
The

modularity of an expert system is an important distinguishing characteristics compared to a conventional
computer program. Modularity is affected in an expert system by the use of three distinct components as
shown in fig 6.2. A good expert system is expected to grow as it learns from user feedback. Feedback is
incorporated into the knowledge base as appropriate to make the expert system smarter. The dynamism of
the application environment for expert systems is based on the individual dynamism of the components.
This can be classified into three categories as follows.
a) Most dynamic: The most dynamic part of an expert system is always the working memory. The
content of the working memory, sometimes called the data structure, changes with each problem
situation. Consequently, it is the most dynamic component of an expert system assuming, of
course that it is kept current.
b) Moderately dynamic: This part in the expert system is the knowledge base. The knowledge base
need not change unless a new piece of information arises that indicates a change in the problem
solution procedure. Changes in the knowledge base should be carefully evaluated before being
implemented. In effect, changes should not be based on just one consultation experience.
c) Least dynamic: The least dynamic part is the inference engine. As the control and coding
structure of an inference engine is very strict, so changes are made only if absolutely necessary to
correct a bug or enhance the inferential process. Commercial inference engines, in particular,
change only at the discretion of the developer. Since frequent updates can be disruptive and costly
to clients, most commercial software developers try to minimize the frequency of updates.
RULE BASED ARCHITECTURE OF AN EXPERT SYSTEM
The most common form of architecture used in expert and other types of knowledge based systems is the
production system or it is called rule based systems. This type of system uses knowledge encoded in the
form of production rules i.e. if-then rules. The rule has a conditional part on the left hand side and a
conclusion or action part on the right hand side. For example if: condition1 and condition2 and
condition3
Then: Take action4
Each rule represents a small chunk of knowledge to the given domain of expertise. When the known facts
support the conditions in the rule’s left side, the conclusion or action part of the rule is then accepted as
known. The rule based architecture of an expert system consists of the domain expert, knowledge
engineer, inference engine, working memory, knowledge base, external interfaces, user interface,
explanation module, database spreadsheets executable programs s mentioned in figure .

Integration of Expert systems Components
The components of the rule based architecture are as follows.
1. User Interface: It is the mechanism by which the user and the expert system communicate with each
other i.e. the use interacts with the system through a user interface. It acts as a bridge between user
and expert system. This module accepts the user queries and submits those to the expert system. The
user normally consults the expert system for following reasons.
a) To get answer of his/her queries.
b) To get explanation about the solution for psychological satisfaction.
The user interface module is designed in such a way that at user level it accepts the query in a
language understandable by expert system. To make the expert system user friendly, the user
interface interacts with the user in natural language. The user interface provides as much facilities as
possible such as menus, graphical interfaces etc. to make the dialog user friendly and more attractive.
2. Explanation Module: The explanation module explains the reasoning of the system to a user. It
provides the user with an explanation of the reasoning process when requested. The credibility of

expert system will be established only when it is able to explain “how and why” a particular
conclusion is drawn. This explanation increases the belief of user in the expert system.
a) Explanation(How): To respond to a how query, the explanation module traces the chain of rules
fired during a consolation with the user. This explanation mode can be activated once the process
is over. It explains how a fact was deduced by the system and similarly how a rule was/wasn’t
used. The simplest way to specify this is to explain the rule which allows the deduction. For e.g.
If the system (S) will give information about the parent-child relationship to the user (U) then the
followings can be possible.
S: My diagnosis is “A is the father of B”
U: How?
S: The result was obtained by the addition of following facts and rules.
Fact no 11: A is the parent of Hari.
Fact no 15: A is a male.
Fact no 110: X is father of Y:
X is parent of Y, X is male.
So A is the father of B.
b) Explanation (Why)? To respond to a why query, the explanation module must be able to explain
why certain information is needed by the inference engine to complete a step in the reasoning
process. This mode of explanation is dynamic and is used when the system functions in backward
chaining. If the user doesn’t understand the reason behind the question. The system is capable of
explaining why the question was asked.
For example S: Is the following true?
A is the father of B.
U: Why?
S: I need the fact:
A is the father of B to establish the following fact “B is the son of A”.
By using the rule no. 4:
A is the father of B:
B is the son of A.
3. Working Memory: It is a global database of facts used by the rules.
4. Knowledge Engineering: The primary people involved in building an expert system are the
knowledge engineer, the domain expert and the end user. Once the knowledge engineer has obtained
a general overview of the problem domain and gone through several problem solving sessions with

the domain expert, he/she is ready to begin actually designing the system, selecting a way to represent
the knowledge, determining the search strategy (backward or forward) and designing the user
interface. After making complete designs, the knowledge engineer builds a prototype. The prototype
should be able to solve problems in a small area of the domain. Once the prototype has been
implemented, the knowledge engineer and domain expert test and refine its knowledge by giving it
problems to solve and correcting its disadvantages.
5. Knowledge Base: In rule based architecture of an expert system, the knowledge base is the set of
production rules. The expertise concerning the problem area is represented by productions. In rule
based architecture, the condition actions pairs are represented as rules, with the premises of the rules
(if part) corresponding to the condition and the conclusion (then part) corresponding to the action.
Case-specific data are kept in the working memory. The core part of an expert system is the
knowledge base and for this reason an expert system is also called a knowledge based system. Expert
system knowledge is usually structured in the form of a tree that consists of a root frame and a
number of sub frames. A simple knowledge base can have only one frame, i.e. the root frame whereas
a large and complex knowledge base may be structured on the basis of multiple frames.
Inference Engine: The inference engine accepts user input queries and responses to questions through
the I/O interface. It uses the dynamic information together with the static knowledge stored in the
knowledge base. The knowledge in the knowledge base is used to derive conclusions about the current
case as presented by the user’s input. Inference engine is the module which finds an answer from the
knowledge base. It applies the knowledge to find the solution of the problem. In general, inference engine
makes inferences by deciding which rules are satisfied by facts, decides the priorities of the satisfied rules
and executes the rule with the highest priority. Generally inferring process is carried out recursively in 3
stages like match, select and execute. During the match stage, the contents of working memory are
compared to facts and rules contained in the knowledge base. When proper and consistent matches are
found, the corresponding rules are placed in a conflict set.
APPLICATIONS OF EXPERT SYTEM
There are several major application areas of expert system such as agriculture, education, environment,
law manufacturing, medicine power system etc. Expert system is used to develop a large number of new
products as well as new configurations of established products. When established products are modified
to include an expert system as a component or when an established product item is replaced with an
expert system, the expert system supported entity is called intelligent. Expert systems are designed and
created to facilitate tasks in the fields of accounting, medicine, process control, financial service,
production, education etc. The foundation of a successful expert system depends on a series of technical
procedures and development that may be designed by certain related experts.

Expert Systems are for everyone
Everyone can find an application potential in the field of expert systems. Contrary to the belief that expert
systems may pose a threat to job security, expert systems can actually help to create opportunities for new
job areas. No matter which is of business one is engages in, expert systems can fulfill the need for higher
productivity and reliability of decisions. Some job opportunities offered by the expert system are listed
below:
→ Basic Research
→ Applied Research
→ Knowledge Engineering
→ The development of Inference engine
→ Training
→ Sales and marketing
Expert System in Education
In the field of education, many of the expert system’s application are embedded inside the Intelligent
Tutoring System (ITS) by using techniques from adaptive hypertext and hypermedia. Most of the system
usually will assist student in their learning by using adaptation techniques to personalize with the
environment prior knowledge of student and student’s ability to learn. Expert system in education has
expanded very consistently from micro computer to web based and agent based technology. Web based
expert system can provide an excellent alternative to private tutoring at any time from any place where
internet is provided. Agent based expert system will help users by finding materials from the web based
on the user’s profile. Expert system also had tremendous changes in the applying of methods and
techniques. Expert system are beneficial as a teaching tools because it has equipped with the unique
features which allow users to ask question on how, why and what format. When it is used in the class
environment, surely it will give many benefit to student as it prepare the answer without referring to the
teacher. Beside that, expert system is able to give reasons towards the given answer. Expert system had
been used in several fields of study including computer animation, computer science and engineering,
language teaching business study etc.
Expert system in Agriculture
The expert system for agriculture is same as like other fields. Here also the expert system uses the rule
based structure and the knowledge of a human expert is captured in the form of IF-THEN rules and facts
which are used to solve problems by answering questions typed at a keyboard attached to a computer. For
example, in pest control, the need to spray, selection of a chemical to spray, mixing and application etc.
The early, state of developing the expert systems are in the 1960’s and 1970’s were typically written on a
mainframe computer in the programming language based on LISP. Some examples of these expert

systems are MACSYMA developed at the Massachusetts Institute of Technology (MIT) for assisting
individuals in solving complex mathematical problems. Other examples may be MYCIN, DENDRAL,
and CALEX etc. The rises of the agricultural expert system are to help the farmers to do single point
decisions, which to have a well planning for before start to do anything on their land. It is used to design
an irrigation system for their plantation use. Also some of the other functions of agricultural expert
system are:
→ To predict the extreme events such as thunderstorms and frost.
→ To select the most suitable crop variety.
→ Diagnosis of liver stock disorder and many more.
Expert System for a particular decision problem
The expert system can be used as a stand alone advisory system for the specific knowledge domain. It
also can provide decision support for a high level human expert. The main purposes, the rises of the
expert system are as a delivery system for extension information, to provide management education for
decision makers and for dissemination of up-to-date scientific information in a readily accessible and
easily understood form, to agricultural researchers, advisers and farmers. By the help of an expert system,
the farmers can produce a more high quality product to the citizen.
Expert System for Text Animation (ESTA)
The idea behind creating an expert system is that it can enable many people to benefit from the
knowledge of one person – the expert. By providing it with a knowledge base for a certain subject area,
ESTA can be used to create an expert system for the subject:
ESTA + Knowledge base = Expert System
Each knowledge base contains rules for a specific domain. A knowledge base for an expert system to give
tax advice might contain rules relating marital status, mortgage commitments and age to the advisability
of taking out a new life insurance policy. ESTA has all facilities to write the rules that will make up a
knowledge base. ESTA has an inference engine which can use the rules in the knowledge base to
determine which advice is to be given to the expert system user. ESTA also features the ability for the
expert system user to obtain answers to questions such as “how” and “why”. ESTA is used by a
knowledge engineer to create a knowledge base and by the expert system user to consult a knowledge
base. Knowledge representation in ESTA is based on the items like sections, parameters, title.

Ai notes

More Related Content

What's hot

Similar to Ai notes

Recently uploaded

In this document

Ai notes