ENGINEERING COLLEGES

2015 – 16 Even Semester

MINIMUM STUDY MATERIAL (MSM)

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

SUBJECT CODE : CS6659

SUBJECT NAME : ARTIFICIAL INTELLIGENCE

Regulation: 2013 Year and Semester :III/06

Prepared by

Sl.
Name of the Faculty Designation Affiliating College
No.
1 Mr. Balaji Prof SCADCET

2 Mr.K.Vinoth raja AP SMTEC

3 Ms.S.Agnes Joshy AP FXEC

4 Mr.K.Sivakumar AP FXEC

5 Mr.S.Franson Varun Richo AP FXEC

6 Ms.Fanny Arul AP SCADEC

Verified by DLI, CLI and Approved by the Centralised Monitoring Team dated 18.12.2015

Copyright © SCAD

ANNA UNIVERSITY, CHENNAI-25

 REGULATION 2013

CS6659 ARTIFICIAL INTELLIGENCE 3 0 0 100

UNIT I INTRODUCTION TO Al AND PRODUCTION SYSTEMS 9

Introduction to AI-Problem formulation, Problem Definition -Production systems,
Control strategies, Search strategies. Problem characteristics, Production system
characteristics -Specialized production system- Problem solving methods - Problem graphs,
Matching, Indexing and Heuristic functions -Hill Climbing-Depth first and Breath first,
Constraints satisfaction - Related algorithms, Measure of performance and analysis of search
algorithms.

UNIT II REPRESENTATION OF KNOWLEDGE 9

Game playing - Knowledge representation, Knowledge representation using Predicate
logic, Introduction to predicate calculus, Resolution, Use of predicate calculus, Knowledge
representation using other logic-Structured representation of knowledge.

UNIT III KNOWLEDGE INFERENCE 9

Knowledge representation -Production based system, Frame based system. Inference -
Backward chaining, Forward chaining, Rule value approach, Fuzzy reasoning - Certainty
factors, Bayesian Theory-Bayesian Network-Dempster - Shafer theory.

UNIT IV PLANNING AND MACHINE LEARNING 9

Basic plan generation systems - Strips -Advanced plan generation systems – K strips -
Strategic explanations -Why, Why not and how explanations. Learning- Machine learning,
adaptive Learning.

UNIT V EXPERT SYSTEMS 9

Expert systems - Architecture of expert systems, Roles of expert systems -
Knowledge Acquisition –Meta knowledge, Heuristics. Typical expert systems - MYCIN,
DART, XOON, Expert systems shells. TOTAL : 45
TEXT BOOKS

1. Kevin Night and Elaine Rich, Nair B., “Artificial Intelligence (SIE)”, McGraw Hill-
2008. (Unit-1,2,4,5).
2. Dan W. Patterson, “Introduction to AI and ES”, Pearson Education, 2007. (Unit-III)
REFERENCES:
1. Peter Jackson, “Introduction to Expert Systems”, 3rd Edition, Pearson Education,
2007.
2. Stuart Russel and Peter Norvig “AI – A Modern Approach”, 2nd Edition, Pearson
Education 2007.
3. Deepak Khemani “Artificial Intelligence”, Tata Mc Graw Hill Education 2013.
4. Nils.J.Nilsson “Principles of Atrificial Intelligence”, Springer –Verlag 1982.
TABLE OF CONTENTS
S.NO TITLE PAGE NO.

1
a Aim and Objective of the subject
2
b Detailed Lesson Plan
5
c UNIT I - INTRODUCTION TO AI PART A
8
d UNIT I - INTRODUCTION TO AI PART B
8
1 Informed search strategies
18
2 Uninformed search strategies
23
3 Problem characteristics
29
4 Types of production system
31
5 Constraint satisfaction problem
33
6 Hill climbing
36
7 Problem solving methods
38
8 Means end analysis
41
e UNIT II REPRESENTATION OF KNOWLEDGE PART A
44
f UNIT II REPRESENTATION OF KNOWLEDGE PART B
44
9 Predicate logic, resolution, unification

10 Issues in knowledge representation 51

53
11 Alpha beta cutoff & minmax algorithm
56
12 Conversion of predicate to causal form
58
13 Approaches to knowledge representation
60
g UNIT III KNOWLEDGE INFERENCE PART A
63
h UNIT III KNOWLEDGE INFERENCE PART B
63
14 Forward and backward chaining
67
15 Production and frame based system
73
16 Fuzzy set and fuzzy logic
77
17 Bayesian network
 79
18 Rule based approach & dempster shafer theory
82
i UNIT IV PLANNING AND MACHINE LEARNING PART A
85
j UNIT IV PLANNING AND MACHINE LEARNING PART A
85
19 Advanced plan generation system
88
20 Components of planning system
90
21 Machine learning methods
98
22 Adaptive learning
101
23 Goal stack planning
107
24 Strips problem solving
108
25 Steps in design of learning system
111
k UNIT V EXPERT SYSTEMS PART A
111
l UNIT V EXPERT SYSTEMS PART A
115
26 Expert systems components
119
27 DART
122
28 MYCIN
125
29 Pitfalls in expert system
126
30 Xcon
128
31 Expert system shells
131
32 Knowledge acquisition
132
33 Characteristics, role and advantages of expert system
137
m Industrial / Practical Connectivity of the subject
 AIM AND OBJECTIVE OF THE SUBJECT

Aim of the subject

 To make machines do a variety of useful tasks.

 To design computer programs that mimics human behavior and expertise so that
we can utilize human expertise even if expert is unavailable.
 To solve real world problems that can only be solved by human intelligence.

Objective of the subject

The student should be made to:

 Study the concepts of Artificial Intelligence.

 Learn the methods of solving problems using Artificial Intelligence.
 Introduce the concepts of Expert Systems and machine learning.

1
 DETAILED LESSON PLAN

Text Book
1. S Kevin Night and Elaine Rich, Nair B., “Artificial Intelligence (SIE)”, McGraw Hill-
2008. (Unit-1,2,4,5). (Copies available in Library : Yes)
2. Dan W. Patterson, “Introduction to AI and ES”, Pearson Education, 2007. (Unit-III)
(Copies available in Library : Yes)
REFERENCES:
1. Peter Jackson, “Introduction to Expert Systems”, 3rd Edition, Pearson Education,
2007. (Copies available in Library : Yes)
2. Stuart Russel and Peter Norvig “AI – A Modern Approach”, 2nd Edition, Pearson
Education 2007. (Copies available in Library : Yes)
3. Deepak Khemani “Artificial Intelligence”, Tata Mc Graw Hill Education 2013.
(Copies available in Library : No)
4. Nils.J.Nilsson “Principles of Atrificial Intelligence”, Springer –Verlag 1982. (Copies
available in Library : Yes)
5. www.nptel.ac.in
Hours
Sl. Cumulative Books
Unit Topic / Portions to be Covered Required
No Hrs Referred
/ Planned

UNIT I - INTRODUCTION TO Al AND PRODUCTION SYSTEMS

Introduction to AI-Problem formulation, Problem

1 1 1 1 T1 R3
Definition
2 1 Production systems, Control & Search strategies. 1 2 T1 R3
3 1 Problem characteristics 1 3 T1 R3
4 1 Production system characteristics 1 4 T1
5 1 Specialized production system 1 5 R4
Problem solving methods - Problem graphs,
6 1 1 6 T1 R3
Matching, Indexing and Heuristic functions
7 1 Hill Climbing 1 7 T1 R3
Depth first and Breath first, Constraints
8 1 1 8 T1 R3
satisfaction
9 1 Related algorithms 1 9 T1
Measure of performance and analysis of search
10 1 1 10 T1
algorithms.
UNIT II – REPRESENTATION OF KNOWLEDGE

11 2 Game playing 1 11 T1 R3
12 2 Knowledge representation 1 12 T1 T2

2
 Hours
Sl. Cumulative Books
Unit Topic / Portions to be Covered Required
No Hrs Referred
/ Planned

13 2 Knowledge representation using Predicate logic 1 13 T1

14 2 Introduction to predicate calculus 1 14 T1
15 2 Resolution 1 15 T1
16 2 Use of predicate calculus 1 16 T1
17 2 Knowledge representation using other logic 1 17 T1
18 2 Knowledge representation using other logic 1 18 T1
19 2 Structured representation of knowledge. 1 19 T1, R3
20 2 Structured representation of knowledge. 1 20 T1, R3

UNIT III- KNOWLEDGE INFERENCE

21 3 Knowledge representation 1 21 T1
22 3 Production based system 1 22 T1
23 3 Frame based system 1 23 T1 T2
24 3 Inference - Backward chaining 1 24 T1
25 3 Forward chaining 1 25 R3
26 3 Rule value approach, Fuzzy reasoning 1 26 T1 R3
27 3 Certainty factors 1 27 T1 T2
28 3 Bayesian Theory 1 28 T1 T2
29 3 Bayesian Network 1 29 T1 T2
30 3 Dempster - Shafer theory. 1 30 T1 T2

UNIT IV – PLANNING AND MACHINE LEARNING

T1 R3
31 4 Basic plan generation systems 1 31
R4
T1 R3
32 4 Strips 1 32
R4
T1 R3
33 4 Advanced plan generation systems 1 33
R4
34 4 R strips 1 34 T1 R4
35 4 Strategic explanations 1 35 T1
36 4 Why, Why not and how explanations. 1 36 T1
37 4 Learning, Machine learning 1 37 T1 R2
38 4 Adaptive Learning. 1 38 T1

3
 Hours
Sl. Cumulative Books
Unit Topic / Portions to be Covered Required
No Hrs Referred
/ Planned

UNIT V - EXPERT SYSTEMS

39 5 Expert systems 1 39 T1
40 5 Architecture of expert systems 1 40 R2
41 5 Roles of expert systems 1 41 T1
42 5 Knowledge Acquisition, Meta knowledge 1 42 T1
43 5 Heuristics 1 43 R3
44 5 Typical expert systems - MYCIN 1 44 R3
45 5 DART 1 45 R3
46 5 XCON 1 46 R3
47 5 Expert systems shells. 1 47 T1

4
 UNIT I INTRODUCTION TO Al AND PRODUCTION SYSTEMS

Introduction to AI- PART A

Problem formulation,
1. List down the characteristics of intelligent agent.
Problem Definition -
Production systems, Apr/May 2011
Control strategies, Search 1) The intelligent agent must learn and improve
strategies. Problem through interaction with the environment.
characteristics, 2) The intelligent agent must adapt online and in the
Production system real time situation.
characteristics - 3) The intelligent agent must accommodate new
Specialized production problem solving rules incrementally.
system- Problem solving
4) The intelligent agent must have memory which
methods - Problem
graphs, Matching, must exhibit storage and retrieval capacities.
Indexing and Heuristic 2. What do you mean by local maxima with respect
functions -Hill Climbing- to search technique?
Depth first and Breath
first, Constraints May-11
satisfaction - Related Local maximum is the peak that is higher than
algorithms, Measure of
each of its neighbor states, but lower than the global
performance and analysis
maximum. i.e. a local maxima is a tiny hill on the
of search algorithms.
surface whose peak is not as high as the main peak.
Hill climbing fails to find optimum solution when it encounters local maxima. Any
small moves from here also make things worse. At local maxima all the search
procedure turns out to be wasted here. It is like a dead end.
3. What are the four components to define a problem? Define them.
Apr/May 2013
The four components to define a problem are,
1. Initial state- It is the state in which agent starts in.
2. A description of possible actions- It is the description of possible actions which
are available to the agent.
3. The goal test- It is the test determines whether a given state is goal state.
4. A path cost function- It is the function that assigns a numeric cost to each path.
The problem solving agent is expected to choose a cost function that reflects its
own performance measure.
4. What is the advantage of heuristic function? Dec-11
Heuristic function ranks alternative paths in various search algorithms at each
branching step, based on the available information, so that a better path is chosen.
The main advantage of heuristic function is that it guides for which state to
explore now, while searching. It makes use of problem specific knowledge like
constraints to check the goodness of a state to be explored. This drastically reduces
the required searching time.
5. What is AI?

5
 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 involves
borrowing characteristics from human intelligence, and applying them as algorithms in a
computer friendly way. This word was coined by McCarthy in the year 1956.

6. Define the term process operationalization.

The process of creating a formal description of the problem using the knowledge about the
given problem, so as to create a program for solving a problem is called process
operationalization. The steps are

 Define a state space.

 Identify the initial states.
 Specify goal states
 Specify set of rules.
7. What is heuristic? Define heuristic function? What is the purpose of heuristic
function?
Heuristics can be said to be estimates of how far the goal state is. Heuristics basically
predict how far the goal state maybe or how much it will cost to get to the goal state from
a particular node.

A heuristic function is a function that maps from problem state description to measures
desirability, usually represented as number weights. The value of a heuristic function at a
given node in the search process gives a good estimate of that node being on the desired
path to solution. Heuristics which will be used to increase the efficiency of Uniform Cost
Search algorithm.

8. Define Constraint satisfaction. (May 2012)

 It is a search procedure that operates in a space of constrained sets.
 The initial states contain the constraints that are given in problem description.
 A goal state is any state that has been constrained to a tolerable level, where
the level of tolerance depends upon the problem.
 CS is a two step process.
i. Discover the constraints involved in the problem and propagate
them throughout the system.
ii. Any hypotheses that strengthen the constraints is formulated. If a
solution is found, then we do nothing else, we backtrack and try a
different guess.
9. List the various types of searching available in AI.
Uninformed searching

1. Breadth first search

2. Depth first search
3. Brute force or blind search
4. Greedy search
Informed searching or Heursitic search

6
 1. Best first search
2. Branch and Bound Search
3. A* Search
4. AO* Search
5. Hill Climbing
6. Constraint satisfaction
7. Means end analysis
10. What are the various problem solving methods?
The various methods used are

 Problem graph
 Matching
 Indexing
 Heuristic function
11. Define Production system.
A production system is one that consists of

 A set of rules.
 One or more knowledge/database.
 A control strategy
 A rule applier.
Production system also incorporates a family of general production system
interpreters that include

i. Basic production system languages.

ii. Expert System shells.
12. List the various classes of production system.
The various classes are

a. A Monotonic production system: A production system in which

applications of a rule never prevents the future application of another rule
that could also been applied at the time the first rule was selected.
b. A nonmonotonic production system: A production system in which the
rule of the monotonic production system doesn’t apply.
c. A partially commutative production system: A production system which
has the property of transforming x into state y on the application of
specific sequence of rules.
d. A commutative production system: A production system that is both
monotonic and partly commutative.
13. What are the various building blocks to solve a problem?
The various things to be done to solve a problem is

a. The problem should be defined precisely.

b. Analyze the problem.

7
 c. Isolate and represent the task knowledge that is necessary to solve the
problem.
d. Choose the best problem solving technique and apply it to the particular
problem.
14. What are the requirements of a good control strategy?
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.
 Control strategies that do not cause motion will never lead to a solution.
 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.
 The third requirement is it is efficient. It finds a good answer to the problem.
15. What is an AI technique? (June 2013)
An AI technique is a method that exploits knowledge that should be represented in
such a way that

a. Generalization is captured through use of knowledge.

b. People who provide the knowledge should also understand it.
c. It is easily modifiable to correct errors so they refine changes in the world.
d. It is used in many situations that is not accurate or complete.
e. It narrows the range of possibilities.

PART- B

1. Explain about informed search strategies in detail

We can also call informed search as Heursitic search. It can be classified as below
 Informed searching or Heursitic search
 Best first search
 Branch and Bound Search
 A* Search
 AO* Search
 Hill Climbing
 Constraint satisfaction
 Means end analysis
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

8
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.
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.

9
 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

Implementation:

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 B

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. ICD

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

10
 Step12:Now M is the FRONT node of the queue which is our goal node. So stop here
and exit. LHM

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.

Step 4: This process will continue until we are getting the goal 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.

11
 Step 5: Go to step 5.

Step 6: Exit.

Implementation:

Let us take the following example for implementing the Branch and Bound algorithm.

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.

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.

AB

Step 2:

Now the stack will be

C FBA

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.

A5B 4C

12
Step 3:

As the top of the stack is D. So calculate neighbours of D.

CFDB

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:

CFD

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

13
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.

14
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.

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.

15
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.

A f A

i.e. OPEN
= CLOSE = (NULL)

Step 2:

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 =

B C

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 = CLO
SE =

16
 G D E A B C

‘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 =

A B C G (O) D E H (O)

D E

Step 5:

Now we have been reached at our goal state. So place F into CLOSE.

(O) (O)

A B C G D E H

i.e. CLOSE =

Step 6:

Success and Exit

AO* Graph:

17
 B C
D E

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.

2. Explain about uninformed search strategies in detail

 Uninformed searching is classified as

 Breadth first search
 Depth first search
 Brute force or blind search
 Greedy 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:

18
 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.

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 BCA

Step 4: Now b is the FRONT node of the queue .So process B and finds the neighbours of B

i.e. D. CDB

Step 5: Now find out the neighbours of C i.e. E

DBE

Step 6: Next find out the neighbours of D as D is the FRONT node of the queue

EFD

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

19
Step 8: Finally F is our goal node which is the FRONT of the queue. So exit.

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.

Step 4: This process will continue until we are getting the goal 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.

20
 Step 6: Exit.

Implementation:

Let us take an example for implementing the above DFS algorithm.

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.A

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.

BFGC

Step 5: Now G is the top node of the stack. Find its neighbour i.e. M

BFMG

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.

BFM

Step 7: Now F is the top node and its neighbours are K and L. so PUSH them on to
the stack.

BKLF

Step 8: Now L is the top node of the stack, which is our goal node.
21
 BKL

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
Disadvantages:

 It is possible that may states keep reoccurring. There is no guarantee of finding

the goal node.
 S ometimes the states may also enter into infinite loops
.

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.

22
 Figure Greedy Search

3. Explain briefly the various 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 independent smaller or easier 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 knowledge base to be used for solving the problem internally consistent?
 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.

Is the problem decomposable?

 Can the problem be broken down to smaller problems to be solved independently?

 Decomposable problem can be solved easily.

(x2 + 3x + sin2x.cos2x)dx

x2dx 3xdx sin2x.cos2xdx

23
 (1  cos2x)cos2xdx

cos2xdx cos4xdx

The next problem is called blocks world.

START GOAL

C B

A B C

Blocks World

Applying the technique of problem decomposition to this simple blocks lead to the
solution to the problem.

ON(B, C) and ON(A, B)

ON(B, C) ON(A, B)

CLEAR(A) ON(A, B)

A
C B
A B C
Decomposable problems can be solved by the Divide and Conquer technique of problem
decomposition.

Can solution steps be ignored or undone?

Consider a mathematical theorem. First we take a lemma and prove it thinking that it is
useful. But later we realize that the lemma is no help. But here there is no problem. The rules
that could have been applied at the outset can still be applied. We can just proceed from the
first place.

24
Example :

The 8 Puzzle. It is a square tray in which are placed 8 square tiles. The remaining 9th
square is uncovered. Each tile has a number in it. A tile that is adjacent to the blank space can
be slid into that space. A game consisting 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.

Start state
2 8 3

1 6 4

7 5

Goal state.

1 2 3

8 4

7 6 5

We can backtrack and undo the first move

Mistakes can be recovered.

The control mechanism must keep track of order in which operations are performed so that
the operations can be undone one at a time if necessary. The control structure for a theorem
prover does not need to record all that information. Consider another problem of playing
chess. Suppose a chess playing program realizes a stupid move after some moves, it cannot
backup and start the game over from that point. It can try a best of the current situation and
go on from there. These problems - theorem proving, 8 puzzle and chess illustrate the
difference between three important classes of problems. Ignorable problems can be solved
using a simple control structure that never backtracks.

 Recoverable problems can be solved using backtracking.

 Irrecoverable problems can be solved by recoverable style methods via planning.
These three definitions make reference to the steps of the solution to a problem and thus may
appear to characterize particular productive systems for solving a problem rather than the
problem itself.

A different formulation of the same problem would lead to the problem being characterized
differently.

The recoverability of a problem plays an important role in determining the complexity of the
control structure necessary for its solution. Ignorable problems can be solved using a simple

25
control structure that never backtracks. such a control structure is easy to implement.
Recoverable problems can be solved by a simple technique called backtracking. It can be
implemented by using push down stack., in which the decisions are recorded in case they
need later to be undone. Irrecoverable problems will need to be solved by a system that
expends a great deal of effort making each decision, since the decision must be final. Some
irrecoverable problems can be solved by recoverable style methods used in a planning
process in which an entire sequence of steps is analysed in advance to discover where it will
lead before the first step is actually taken.

Is the universe predictable?

Consider the 8-Puzzle problem(Certain outcome). Every time we make a move, we

know exactly what will happen. This means it is possible to plan the entire sequence of
moves and be confident of the resulting state. We can use planning to avoid having to undo
actual moves, although it will be still necessary to backtrack past those moves one at a time
during the planning process. Thus a control structure that allows backtracking is necessary.

Consider another game of Playing Bridge( Uncertain outcome). We cannot know

exactly where all the cards are or what the other players will do on their turns. The best is to
investigate several plans and use probabilities of the various outcomes to choose a plan that
has a highest expected probability of leading to a good score.

Generally,

 For certain-outcome problems, planning can used to generate a sequence of operators that
is guaranteed to lead to a solution.
 For uncertain-outcome problems, a sequence of generated operators can only have a good
probability of leading to a solution.
 Plan revision is made as the plan is carried out and the necessary feedback is provided.
Is a good solution absolute or relative?

Consider the problem of answering questions based on a database of simple facts.

1. Marcus was a man.

2. Marcus was a Pompeian.
3. Marcus was born in 40 A.D.
4. All men are mortal.
5. All Pompeian’s died when the volcano erupted in 79 A.D.
6. No mortal lives longer than 150 years.
7. It is now 2004 A.D.
The question asked is “ is Marcus alive”?

Different reasoning paths lead to the answer. It does not matter which path we follow.

Justification

1 Marcus was a man. Axiom 1

26
 4 All men are mortal. Axiom 4

8 Marcus is mortal. 1,4

3 Marcus was born in 40 A.D. axiom 3

7 It is now 1983. axiom 7

9 Marcus age is 1943 years. 3,7

6 No mortal lives longer than 150 years. Axiom 6

10 Marcus is dead. 8,6,9

Or

7 it is now 1983 axiom 7

5 all Pompeian’s died in 79 axiom 5

11 all Pompeian’s are dead now 7,5

2 Marcus was a Pompeian axiom 2

12 Marcus is dead 11,2

The Travelling Salesman Problem

The goal is to find the shortest route that visits each city exactly once.

Generally

 Any-path problems can be solved in a reasonable amount of time using heuristics that
suggest good
paths to explore. If the heuristics are not perfect , the search for a solution may not
be as direct as possible.

 For best-path problems, much more exhaustive search will be performed.

Is the solution a state or a path?

Finding a consistent intepretation

 “The bank president ate a dish of pasta salad with the fork”.
 “bank” refers to a financial situation or to a side of a river?
 “dish” or “pasta salad” was eaten?
 Does “pasta salad” contain pasta, as “dog food” does not contain “dog”?
 Which part of the sentence does “with the fork” modify?
 What if “with vegetables” is there?
No record of the processing is necessary.

The Water Jug Problem : The path that leads to the goal must be reported.
27
A path-solution problem can be reformulated as a state-solution problem by describing a state
as a partial path to a solution. The question is whether that is natural or not.

What is the role of knowledge?

Suppose if we have unlimited computing power available , then knowledge can be

acquired easily to speed up the execution of the program.

Playing Chess

Knowledge is important only to constrain the search for a solution.

Reading Newspaper

Knowledge is required even to be able to recognize a solution.

Does the task require human-interaction?

It is useful for computers to be programmed to solve problems in ways that the majority of
people woul not be able to understand. It is possible when human computer interaction is
good.

There are two types of problems.

 Solitary problem, in which there is no intermediate communication and no demand for an

explanation of the reasoning process.
 Conversational problem, in which intermediate communication is to provide either
additional assistance to the computer or additional information to the user.
Example

Chess

28
 Satisfie
Problem characteristic Reason
d

Is the problem
No One game have Single solution
decomposable?

Can solution steps be In actual game(not in PC) we can’t undo previous

No
ignored or undone? steps

Is the problem universe Problem Universe is not predictable as we  are not
No
predictable? sure about move of other player(second player)

Absolute solution : once you get  one solution you

do need to bother about  other possible solution.

Is a good solution absolute Relative Solution : once you get one solution you
absolute
or relative? have to find another possible solution to check
which solution is best(i.e low cost).

By considering this chess is absolute

Is the solution a state or a path to a state?

– For natural language understanding, some of the

words have  different

Is the solution a state or a interpretations .therefore  sentence may cause

path? ambiguity. To solve the problem we need to find
Path
interpretation only 

, the workings are not necessary (i.e path to

solution is not necessary)

So In chess  winning state(goal state) describe path

to state

What is the role of Lot of knowledge helps to constrain the search for
knowledge? a solution.

Conversational

·         In which there is intermediate

communication between a person and the
Does the task require computer, either to
No
human-interaction? Provide additional assistance to the computer or to
provide additional information to the user, or
both.

In chess  additional assistance  is not required

29
 4. Explain briefly the various production system in detail?

Production system is a mechanism that describes and performs the search process. It consists
of
1. A set of rules.

2. One or more knowledge or database.

3. A control strategy that specifies the order of the rules to be applied.

4. A rule applied.

Requirements of a good control strategy:

1. The first requirement is that it can cause motion.

Consider  the  water  jug  problem. Suppose  we implemented the  simple  control  strategy of
starting  each time at the top of the list of rules and choosing the first applicable one.
2. The second requirement is that it be systematic. The requirement that a control  strategy  be
systematic  corresponds to the  need  for global motion as  well as  for
local motion.

Production System Characteristics:

We have argued that production systems are a good way to describe the operations that can
be performed in a search for a solution to a problem.

1. Can production systems, like problems, be described by a set of characteristics that shed
some light on how they can easily be implemented?

2. If  so,  what relationships are  there  between  problem  types and the types  of  production
system  best suited  to solving the problems?

 There are different classes of the production system:

A monotonic production system: It is a production system in which the application of a rule

never prevents the later application of another rule that could also have been applied at the
time first rule was selected.

A  non-monotonic  production  system:  A  non-monotonic  production  system  is  one  in

which this  is  not  true.

Partially  commutative  production  system:  A  partially  commutative  production  system

is a  production system with the property that if the application of a particular sequence of
rules transforms state x into state y then any permutation  of those rules that is allowable (i.e.
each rules  preconditions are  satisfied  when it  is applied) also
transforms  state  x  into  state  y.  

30
 Monotonic Nonmonotonic

Blocks world
Partially commutative Theorem proving
8 puzzle

Not partially
Chemical synthesis Bridge
commutative

A  commutative  production  system:  A  commutative  production  system  is  a

production  system  that  is  both monotonic and partially commutative.

The significance of these categories of production systems lies in the relationship between the
categories and approximate implementation strategies.

For any problem, there exists an infinite number of production systems that describe
ways to find solutions. Some will be natural or efficient than others. Any problem that can be
solved by any production system can be solved by a commutative one.

There is a relationship between kinds of problems and the kinds of systems that lend
themselves naturally to describe the problem.

Partially  commutative,  monotonic  production  systems  are  useful  for  solving

ignorable problems.

Partially commutative monotonic production systems are important from implementation

standpoint because they can be implemented without the ability to backtrack to previous
states when it is discovered that an incorrect path was followed.

Nonmonotonic partially commutative systems are useful for problems in which changes
occur but can be reversed and in which order of operations is not critical.

Partially commutative production systems are significant from an implementation point of

view because they tend to lead to many duplications of individual states durning the search
process.

5. Write in detail about the constraint satisfaction procedure with example?

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.

31
Algorithm:

 Open all objects that must be assigned values in a complete solution.

 Repeat until all objects assigned valid values.
 Select an object and strengthen as much as possible. The set of
constraints that apply to object.
 If set of constraints is different from previous set then open all objects
that share any of these constraints. Remove selected objects.
 If union of constraints discovered above defines a solution, return
solution.
 If union of constraints discovered above defines a contradiction, return
failure.
 Make a guess in order to proceed. Repeat until a solution is found.
 Select an object with a number assigned value and try strengthening its
constraints.
Consider the problem of solving a puzzle
E+V+O+L=AI, If A = 3 && 5<L<O<V<E<10, Find the value of I?
Many AI problems can be viewed as problems of constraint satisfaction.
As compared with a straightforward search procedure, viewing a problem as one of constraint
satisfaction can reduce substantially the amount of search.
Constraint Satisfaction
 Operates in a space of constraint sets.
 Initial state contains the original constraints given in the problem.
 A goal state is any state that has been constrained “enough”.
Two-step process:
 Constraints are discovered and propagated as far as possible.
 If there is still not a solution, then search begins, adding new constraints.
Two kinds of rules:
 Rules that define valid constraint propagation.
 Rules that suggest guesses when necessary.
Constraints
 The simplest type is the unary constraint, which constraints the values of just one
variable.
 A binary constraint relates two variables.
 Higher-order constraints involve three or more variables. Cryptarithmetic puzzles are an
example:

Cryptarithmetic puzzles

32
•Variables:F, T, U, W, R, O, X1, X2, X3
•Domains: {0,1,2,3,4,5,6,7,8,9}
•Constraints:
–Alldiff(F,T,U,W,R,O)
–O + O = R + 10 · X1
–X1+ W + W = U + 10 ·X2
–X2+ T + T = O + 10 ·X3
–X3= F, T ≠0, F≠0

If 2+2 = 4, F=1, then how many fours?

928+928=1856
867+867=1734
846+846=1692
836+836=1672
765+765=1530
734+734=1468
TWO
TWO
FOUR
938+938=1876

6. What are the problems encountered during hill climbing and what are the ways
available to deal with these problems?

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 n eighbour 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:

33
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:

 Select a state that has not been yet applied to the current state and apply it to produce a new
state.
 Procedure to evaluate a new state.
 If the current state is a goal state, then stop and return success.
 If it is better than the current state, then make it current state and proceed further.
 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 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.

34
 Figure Local Maxima

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 neighbouring have 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

To overcome these problems we can

 Back track to some earlier nodes and try a different direction. This is a good way of
dealing with local maxim.
 Make a big jump an some direction to a new area in the search. This can be done by
applying two more rules of the same rule several times, before testing. This is a
good strategy is dealing with plate and ridges. Hill climbing becomes inefficient in
large problem spaces, and when combinatorial explosion occurs. But it is a useful
when combined with other methods.
Steepest Descent Hill Climbing

 This is a variation of simple hill climbing which considers all the moves from the current
state and selects the best one as the next state.
 Also known as Gradient search

35
Algorithm

 Evaluate the initial state. If it is also a goal state, then return it and quit. Otherwise,
continue with the initial state as the current state.
 Loop until a solution is found or until a complete iteration produces no change to current
state:
 Let SUCC be a state such that any possible successor of the current state will be better
than SUCC
 For each operator that applies to the current state do:
 Apply the operator and generate a new state
 Evaluate the new state. If is is a goal state, then return it and quit. If not, compare it
to SUCC. If it is better, then set SUCC to this state. If it is not better, leave SUCC
alone.
 If the SUCC is better than the current state, then set current state to SYCC,
7. Explain the various problem solving methods. (December 2011)
 A problem in AI is defined as a task to be completed.
 Most of the problems are complex and cannot be solved by direct techniques.
 They can be solved by search strategies.
 Every search process can be viewed as a traversal of directed graph in which
each node represents a problem state.
 Arcs represent the relationship between the states represented by the node it
connects.
 The search process starts with an initial state and ends in a final state leading a
path.
 Five important issues have to be considered.
1. The direction to which to conduct the search.
2. The topology of search process.
3. How each node of the search process will be represented?
4. Selecting applicable rules.
5. Using a heuristic function to guide the search.
 The various methods used to solve problems are
1. Problem graph
2. Matching
a. indexing
3. Heuristic functions.
1. Problem Graph

A simple way to implement the search strategy is tree traversal.

Each node of the tree is expanded by the production rules to generate a set of successor
nodes, each of which can in turn be expanded, continuing until a node representing a solution
is found.

Example: Water Jug Problem- two levels of Breadth first search Tree.

36
Implementing this procedure is

 Simple
 Require little bookkeeping.

Disadvantage
Same path may be generated many times and so processing takes place more than once.
This wastage can be avoided by bookkeeping and elimination of redundant nodes.

Graph search procedure

1. Examine the set of nodes that have been created so far to see if the new node already
exists.
2. If it doesn’t, add it to the graph just as for a tree.
3. If it is already present, then
a. Set the node that is being expanded to point to already existing node
corresponding to its successor, rather than new one. Throw the new node.
b. If we keep track of the best path to each node, then check if the new path is
best or worse.
i. If worse do nothing.
ii. If best record the new path as correct path and propagate the change in
cost to all its successor nodes.

Uses

Graph search procedures are useful in partially commutative production system.

1. Matching
It is the process of applying a rule between the current state and precondition of the
rules to determine some identity between them.

They are done by

1. Indexing
2. Matching with variables.
3. Complex matching
37
 4. Filtering.
1. Indexing
 Use the current state as an index into rules and select the matching ones
immediately.
 There are two problems
i. It requires the use of large number of rules.
ii. It is unobvious that rules preconditions are satisfied by a particular
state.
 Instead of using rules, use the current state as index and select the
matching accordingly.
 In a game of chess, an index can be assigned to each board position. This
scheme can be used while searching.
Disadvantage

Indexing cannot be helpful when the preconditions of rules are stated as

highly level predicates.
Advantage

 Indexing in some from is very efficient in rule based system.

 Matching with variables
 It is a simple kind of non literal match that require extensive search when
patterns contain variables.
 Complex and approximate matching.
Rules should be applied if their preconditions approximately match the current situation

Example: A speech-understanding program

 Filtering
The result of matching is a list of rules whose left side have matched the current state
description, along with variable binding that is generated by matching process.

 Heuristic Function
 A Heuristic function is a function that maps from problem state
descriptions to measures of desirability usually represented as numbers.
 A Heuristic is a technique that improves the efficiency of a search process
possibly by sacrificing claims of completeness.
 The purpose of heuristic function is to guide the search process in the most
profitable directions suggesting which paths to follow first when more that
one is available.
8. Explain Means- Ends Analysis with an example. (December 2011)
 Most of the search strategies either reason forward of backward however, often a
mixture of the two directions is appropriate.
 Such mixed strategy would make it possible to solve the major parts of problem
first and solve the smaller problems that arise when combining them together.

38
  Such a technique is called "Means - Ends Analysis".
 The means -ends analysis process centers around finding the difference between
current state and goal state.
 The problem space of means - ends analysis has an initial state and one or more
goal state, a set of operate with a set of preconditions their application and
difference functions that computes the difference between two state a(i) and s(j).
 A problem is solved using means - ends analysis by
1. Computing the current state s1 to a goal state s2 and computing their difference
D12.

2. Satisfy the preconditions for some recommended operator op is selected, then to

reduce the difference D12.

3. The operator OP is applied if possible. If not the current state is solved a goal is
created and means- ends analysis is applied recursively to reduce the sub goal.

4. If the sub goal is solved state is restored and work resumed on the original
problem.

(The first AI program to use means - ends analysis was the GPS General problem
solver)

Means- ends analysis is useful for many human planning activities.

EXAMPLE

 Problem for household robot: moving desk with 2 things on it from one room to
another.
 Main difference between start and goal state is location.
 Choose PUSH and CARRY
Move desk with 2 things on it to new room

PushCarryWalkPickupPutdownPlace

Move object * *

Move robot *

Clear object *

Get object on object *

Get arm empty * *

Be holding object *

39
THE ROBOT OPERATORS
Operator Preconditions Results
PUSH (obj, loc) at(robot,obj) at(obj, loc)
&large (obj) & at (robot, loc)
&clear (obj) &
arm empty
CARRY (obj, loc) at(robot, obj) &Small at(obj, loc) &at(robot,
(obj) loc)

WALK(loc) none At(robot, loc)

PICKUP(obj) At(robot, obj) Holding(obj)
PUTDOWN(obj) Holding(obj) Not holding (obj)
PLACE(obj1, obj2) At(robot,obj2) & holding on(obj1, obj2)
(obj1)
A robot wants to move a desk with two things from one room to another. The object on top
must be moved.
S B________C G
Start PUSH Goal
If CARRY is chosen first, preconditions cannot be met. This results in two more differences
that must be reduced. The location of the robot and the size of the desk.
If PUSH is chosen, then 3 preconditions has to be met. Two of which produce difference
between the start and goal states.
The robot must be brought to current location be using WALK, the surface can be cleaned by
two uses of PICKUP and PUTDOWN.
S(Start)_____________________________________B__________C
WALK(desk_loc)->PU(obj1)->PD(obj1)->PU(obj2)->PD(obj2) PUSH
Once PUSH is performed, the problem state is close to the goal state. The object must be
placed back on the desk. PLACE can be used to put them there. But it cannot be applied
immediately because the robot will be holding the objects.
S(Start)______________________________B_____ C E________G(Goal)
WALK->PU(obj1)->PD(obj1)->PU(obj2)->PD(obj2) -> PUSH PLACE
Use WALK to get the robot back to the objects, followed by PICKUP and CARRY.
Final plan is
 WALK (start_desk_loc)  PU(obj1)  PD(obj1)  PU(obj2)
WALK(start_desk_loc)  PUSH(desk, goal_loc)  PD(obj2)
 PU(obj1)  CARRY(obj1,goal_loc)  PLACE(obj1,desk) 
PLACE(obj2,desk)  CARRY(obj2, goal_loc)  PU(obj2)  WALK(start)
Disadvantages
 It is not used to solve large and complex problems.
 The number of permutation of differences may get too large.
 Working on one difference may interfere with the plan for reducing another.
 Difference table would be immense for complex problems.

40
 UNIT II REPRESENTATION OF KNOWLEDGE

Game playing - Knowledge representation, Knowledge representation using

Predicate logic, Introduction to predicate calculus, Resolution, Use of predicate
calculus, Knowledge representation using other logic-Structured representation of
knowledge.

PART A

1. Define Predicate Logic. (December 2011)

 Relational predicate logic is a generalization of propositional logic that is
obtained by allowing proposition symbols to have arguments, and they are
then called predicates.
 The concepts and definitions for predicate logic are similar to those for
propositional logic.
 However, in order to establish the convention that all occurrences of a
predicate have the same number of arguments one must start with the
definition of a vocabulary.
2. State the use of predicate calculus. (December 2011)
 This is used for capturing the essence of relationship between the classes.
 We can represent complex facts using this calculus.
 Resolving the rules is easy.
3. Draw the mapping between facts and representations. (June 2013)

4. What is meant by alpha beta pruning

The alpha beta cutoff strategy required the maintenance of two threshold
values, one that represents a lower bound on the value that a maximizing node may
have and another that represents a upper found on a value that minimizes node may be
assigned.

41
 5. What is meant by alpha beta pruning?What is the difference between declarative
knowledge and procedural knowledge? (May 2014)

S.No Declarative Knowledge Procedural Knowledge

1 A declarative representation is A procedural knowledge is one in 6. W

one in which knowledge is which the control information that r
specified, but the level of usage is needed to use the knowledge is i
of that knowledge is unknown. assumed to be embedded within the t
knowledge itself. e

2 Logical assertions are viewed as Logical assertions are viewed as

programs. data.

3 Reasoning paths define the We need to augment viable the

viable execution paths reasoning with an interpreter that
proceeds according to instructions
in the knowledge.

4 Control information exists as Control information controls the

resides in constructs knowledge base.

about iterative deepening. (November 2012, May 2014)

 It is a recursive procedure in which we start searching to a fixed depth in the
game tree. We apply the single evaluation function to the first branch,
applying to it the result of possible moves.
 This will result in the initiation of a new minimax search, this time the depth
getting increased to 2.
 This process gets repeated.
7. Define conflict resolution.
The result of matching process is a list of rules whose background have
matched the current state description along with whatever variable bindings were
generated by the matching process. It is the prerogative of the search method to
decide the order in which the rules are to be applied, but sometimes we can
incorporate the decision making in matching process. This phase is called conflict
resolution.

8. What is the purpose of unification? May-12,Dec-12

It is used for finding substitutions for inference rules, which can make different
logical expression to look identical. It helps to match to logical expressions. Therefore
it is used in many algorithms in first order logic.

9. Differentiate between propositional versus first order predicate logic.

42
 Dec-11

Propositional logic First order logic

Propositional logic is less expensive First order logic is more expressive

and do not reflect individual object’s and can represent individual object
properties explicitly along with all its properties.

Propositional logic cannot represent First order logic can represent

relationship among objects. relationship.

It does not consider generalization of It handles generalization

objects.

It includes sentence letters and logical It has the same connectives as

connectives but not quantifier. propositional logic, but it also has
variables for individual objects,
quantifier, symbols for functions and
symbols for relations.

10. What are the limitations in using propositional logic to represent knowledge
base?
May-11
1) It has limited expressive power
2) It cannot directly represent properties of individuals or relations between
individuals.
3) Generalizations, patterns, regularities cannot easily be represented.
4) Many rules are required to write so as to allow inference.
11. Specify the complexity of expectiminimax.
 The Time complexity is O(bmnm)
Where
b is the branching factor
n is the number of chance nodes for each max or min node
m is the maximum depth.
 This extra cost can make games of chance very difficult to solve.
12. What is unification algorithm?
In propositional logic it is easy to determine that two literals cannot both be true at the
same time. Simply look for L and ~L . In predicate logic, this matching process is more
complicated, since bindings of variables must be considered.

For example man (john) and man(john) is a contradiction while man (john) and
man(Himalayas) is not. Thus in order to determine contradictions we need a matching
procedure that compares two literals and discovers whether there exist a set of substitutions
that makes them identical. There is a recursive procedure that does this matching . It is called
Unification algorithm.

43
In Unification algorithm each literal is represented as a list, where first element is the name of
a predicate and the remaining elements are arguments. The argument may be a single element
(atom) or may be another list.
13. How can you represent the resolution in predicate logic?
1. Convert all the statements of S to clausal form.
2. Negate P and convert the result to clausal form. Add it to the set of clauses
obtained in I.
3. Repeat until a contradiction is found.
14. List the canonical forms of resolution.
 eliminate implications
 move negations down to the atomic formulas
 purge existential quantifiers
 rename variables if necessary
 move the universal quantifiers to the left
 move the disjunctions down to the literals
 eliminate the conjunctions
 rename the variables if necessary
 purge the universal quantifiers
15. State the use of unification.

 Syntactical first-order unification is used in logic programming and programming

language type system implementation, especially in Hindley–Milner based type
inference algorithms.
 Semantic unification is used in SMT solvers and term rewriting algorithms. Higher-
order unification is used in proof assistants, for example Isabelle and Twelf, and
restricted forms of higher-order unification (higher-order pattern unification) are
used in some programming language implementations, such as lambda Prolog, as
higher-order patterns are expressive.

PART B

1. State Representation of facts in predicate logic, Explain resolution and unification

algorithm in detail.

Knowledge can be represented as “symbol structures” that characterize bits of

knowledge about objects, concepts, facts, rules, strategies.

Examples: “red” represents colour red, “car1” represents my car , "red(car1)" represents
fact that my car is red.

Assumptions about KR :

44
  Intelligent Behavior can be achieved by manipulation of symbol structures.
 KR languages are designed to facilitate operations over symbol structures, have
precise syntax and semantics; Syntax tells which expression is legal?
e.g., red1 (car1), red1 car1, car1(red1), red1(car1 & car2) ?; and Semantic tells
what an expression means ?

e.g., property “dark red” applies to my car.

 Make Inferences, draw new conclusions from existing facts.

To satisfy these assumptions about KR, we need formal notation that allows automated
inference and problem solving. One popular choice is use of logic.

Logic

Logic is concerned with the truth of statements about the world. Generally each
statement is either TRUE or FALSE. Logic includes: Syntax, Semantics and Inference
Procedure.

1. Syntax:

Specifies the symbols in the language about how they can be combined to form
sentences. The facts about the world are represented as sentences in logic.

2. Semantic:

Specifies how to assign a truth value to a sentence based on its meaning in the world.
It Specifies what facts a sentence refers to. A fact is a claim about the world, and it may be
TRUE or FALSE.

3. Inference Procedure:

Specifies methods for computing new sentences from the existing sentences.

Logic as a KR Language

Logic is a language for reasoning, a collection of rules used while doing logical
reasoning. Logic is studied as KR languages in artificial intelligence. Logic is a formal
system in which the formulas or sentences have true or false values.

Problem of designing KR language is a tradeoff between that which is

(a) Expressive enough to represent important objects and relations in a problem domain.

(b) Efficient enough in reasoning and answering questions about implicit information in a
reasonable amount of time.

Logics are of different types: Propositional logic, Predicate logic, temporal logic, Modal
logic, Description logic etc;

They represent things and allow more or less efficient inference.

45
Propositional logic and Predicate logic are fundamental to all logic.

Propositional Logic is the study of statements and their connectivity.

Predicate Logic is the study of individuals and their properties.

Logic Representation

Logic can be used to represent simple facts.

The facts are claims about the world that are True or False.

To build a Logic-based representation:

1. User defines a set of primitive symbols and the associated semantics.

2. Logic defines ways of putting symbols together so that user can define legal
sentences in the language that represent TRUE facts.

3. Logic defines ways of inferring new sentences from existing ones.

4. Sentences - either TRUE or false but not both are called propositions.

5. A declarative sentence expresses a statement with a proposition as content;

example:

the declarative "snow is white" expresses that snow is white; further, "snow is
white" expresses that snow is white is TRUE.

Resolution and Unification algorithm

In propositional logic it is easy to determine that two literals cannot both be true at the
same time. Simply look for L and ~L . In predicate logic, this matching process is more
complicated, since bindings of variables must be considered.

For example man (john) and man(john) is a contradiction while man (john) and
man(Himalayas) is not. Thus in order to determine contradictions we need a matching
procedure that compares two literals and discovers whether there exist a set of substitutions
that makes them identical. There is a recursive procedure that does this matching . It is called
Unification algorithm.

In Unification algorithm each literal is represented as a list, where first element is the name of
a predicate and the remaining elements are arguments. The argument may be a single element
(atom) or may be another list. For example we can have literals as

(tryassassinate Marcus Caesar)

(tryassassinate Marcus (ruler of Rome))

To unify two literals, first check if their first elements re same. If so proceed. Otherwise they
cannot be unified. For example the literals

46
(try assassinate Marcus Caesar)

(hate Marcus Caesar)

Cannot be unified. The unification algorithm recursively matches pairs of elements, one pair
at a time. The matching rules are :

i) Different constants, functions or predicates can not match, whereas identical ones can.

ii) A variable can match another variable, any constant or a function or predicate expression,
subject to the condition that the function or [predicate expression must not contain any
instance of the variable being matched (otherwise it will lead to infinite recursion).

iii) The substitution must be consistent. Substituting y for x now and then z for x later is
inconsistent. (a substitution y for x written as y/x)

The Unification algorithm is listed below as a procedure UNIFY (L1, L2). It returns a list
representing the composition of the substitutions that were performed during the match. An
empty list NIL indicates that a match was found without any substitutions. If the list contains
a single value F, it indicates that the unification procedure failed.

UNIFY (L1, L2)

1. if L1 or L2 is an atom part of same thing do

(a) if L1 or L2 are identical then return NIL

(b) else if L1 is a variable then do

(i) if L1 occurs in L2 then return F else return (L2/L1)

© else if L2 is a variable then do

(i) if L2 occurs in L1 then return F else return (L1/L2)

else return F.

2. If length (L!) is not equal to length (L2) then return F.

3. Set SUBST to NIL

(at the end of this procedure , SUBST will contain all the substitutions used to unify L1 and
L2).

4. For I = 1 to number of elements in L1 do

i) call UNIFY with the i th element of L1 and I’th element of L2, putting the result in S

ii) if S = F then return F

iii) if S is not equal to NIL then do

47
(A) apply S to the remainder of both L1 and L2

(B) SUBST := APPEND (S, SUBST) return SUBST.

Resolution yields a complete inference algorithm when coupled with any complete search
algorithm. Resolution makes use of the inference rules. Resolution performs deductive
inference. Resolution uses proof by contradiction. One can perform Resolution from a
Knowledge Base. A Knowledge Base is a collection of facts or one can even call it a database
with all facts.

Resolution basically works by using the principle of proof by contradiction. To find the
conclusion we should negate the conclusion. Then the resolution rule is applied to the
resulting clauses.

Each clause that contains complementary literals is resolved to produce a 2new clause, which
can be added to the set of facts (if it is not already present). This process continues until one
of the two things happen. There are no new clauses that can be added. An application of the
resolution rule derives the empty clause An empty clause shows that the negation of the
conclusion is a complete contradiction, hence the negation of the conclusion is invalid or
false or the assertion is completely valid or true.

Steps

Steps for Resolution

1. Convert the given statements in Predicate/Propositional Logic

2. Convert these statements into Conjunctive Normal Form

3. Negate the Conclusion (Proof by Contradiction)

4. Resolve using a Resolution Tree (Unification)

Steps to Convert to CNF (Conjunctive Normal Form)

1. Every sentence in Propositional Logic is logically equivalent to a conjunction of

disjunctions of literals.

A sentence expressed as a conjunction of disjunctions of literals is said to be in Conjunctive

normal Form or CNF.

1. Eliminate implication ‘→’

2. a → b = ~a v b

3. ~ (a ^ b) = ~ a v ~ b …………DeMorgan’sLaw

4. ~ (a v b) = ~ a ^ ~ b ………... DeMorgan’sLaw

5. ~ (~a) = a

48
Eliminate Existential Quantifier ‘∃’

To eliminate an independent Existential Quantifier, replace the variable by a Skolemconstant.

This process is called as Skolemization.

Example: ∃y: President (y)

Here ‘y’ is an independent quantifier so we can replace ‘y’ by any name (say – George Bush).

So, ∃y: President (y) becomes President (George Bush).

To eliminate a dependent Existential Quantifier we replace its variable by SkolemFunction

that accepts the value of ‘x’ and returns the corresponding value of ‘y.’

Example: ∀x : ∃y : father_of(x, y)

Here ‘y’ is dependent on ‘x’, so we replace ‘y’ by S(x).

So, ∀x : ∃y : father_of(x, y) becomes ∀x : ∃y : father_of(x, S(x)).

Eliminate Universal Quantifier ‘∀’

To eliminate the Universal Quantifier, drop the prefix in PRENEX NORMAL FORM i.e. just
drop ∀and the sentence then becomes in PRENEX NORMAL FORM.

Eliminate AND ‘^’

a ^ b splits the entire clause into two separate clauses i.e. a and b

(a v b) ^ c splits the entire clause into two separate clauses a v b and c

(a ^ b) v c splits the clause into two clauses i.e. a v c and b v c

To eliminate ‘^’ break the clause into two, if you cannot break the clause, distribute the OR
‘v’ and then break the clause.

Problem Statement:

1. Ravi likes all kind of food.

2. Apples and chicken are food

3. Anything anyone eats and is not killed is food

4. Ajay eats peanuts and is still alive

5. Rita eats everything that Ajay eats

Prove by resolution that Ravi likes peanuts using resolution.

Step 1: Converting the given statements into Predicate/Propositional Logic

49
 i. ∀x : food(x) → likes (Ravi, x)

ii. food (Apple) ^ food (chicken)

iii. ∀a : ∀b: eats (a, b) ^ ~killed (a) → food (b)

iv. eats (Ajay, Peanuts) ^ alive (Ajay)

v. ∀c : eats (Ajay, c) → eats (Rita, c)

vi. ∀d : alive(d) → ~killed (d)

vii. ∀e: ~killed(e) → alive(e)

Conclusion: likes (Ravi, Peanuts)

Step 2: Convert into CNF

i. ~food(x) v likes (Ravi, x)

ii. Food (apple)

iii. Food (chicken)

iv. ~ eats (a, b) v killed (a) v food (b)

v. Eats (Ajay, Peanuts)

vi. Alive (Ajay)

vii. ~eats (Ajay, c) V eats (Rita, c)

viii. ~alive (d) v ~ killed (d)

ix. Killed (e) v alive (e)

Conclusion: likes (Ravi, Peanuts)

Negate the conclusion

~ likes (Ravi, Peanuts)

50
 Uses of Resolution in Today’s World

 Used widely in AI.

 Helps in the development of computer programs to automate reasoning and theorem
proving

2. Explain the various Issues in knowledge representation.

The fundamental goal of Knowledge Representation is to facilitate inference (conclusions)

from knowledge.

The issues that arise while using KR techniques are many. Some of these are

1. Important Attributes: Any attribute of objects so basic that they occur in almost every
problem domain?
2. Relationship among attributes: Any important relationship that exists among object
attributes?
3. Choosing Granularity : At what level of detail should the knowledge be represented ?
4. Set of objects : How sets of objects be represented ?
5. Finding Right structure : Given a large amount of knowledge stored, how can relevant
parts be accessed ?
1. Important Attributes:

There are attributes that are of general significance.

There are two attributes "instance" and "isa", that are of general importance. These
attributes are important because they support property inheritance.

2. Relationship among Attributes : The attributes to describe objects are themselves

entities they represent.

The relationship between the attributes of an object, independent of specific knowledge they
encode, may hold properties like:

51
Inverses, existence in an isa hierarchy, techniques for reasoning about values and single
valued attributes.

1. Inverses : This is about consistency check, while a value is added to one attribute.

The entities are related to each other in many different ways.

There are two ways of realizing this:

a. first, represent two relationships in a single representation; e.g., a logical representation,

team(Pee-Wee-Reese, Brooklyn–Dodgers), that can be interpreted as a statement about Pee-
Wee-Reese or Brooklyn–Dodger.

b. second, use attributes that focus on a single entity but use them in pairs, one the inverse of
the other; for e.g., one, team = Brooklyn– Dodgers , and the other, team = Pee-Wee-Reese, . .
..

This second approach is followed in semantic net and frame-based systems, accompanied by
a knowledge acquisition tool that guarantees the consistency of inverse slot by checking, each
time a value is added to one attribute then the corresponding value is added to the inverse.

2. Existence in an "isa" hierarchy :

This is about generalization-specialization, like, classes of objects and specialized subsets of

those classes. There are attributes and specialization of attributes.

Example: the attribute "height" is a specialization of general attribute "physical-size" which

is, in turn, a specialization of "physical-attribute".

These generalization-specialization relationships for attributes are important because they

support inheritance.

3. Techniques for reasoning about values :

This is about reasoning values of attributes not given explicitly. Several kinds of information
are used in reasoning, like, height : must be in a unit of length, age : of person can not be
greater than the age of person's parents.

The values are often specified when a knowledge base is created.

4. Single valued attributes :

This is about a specific attribute that is guaranteed to take a unique value.

Example : A baseball player can at time have only a single height and be a member of only
one team. KR systems take different approaches to provide support for single valued
attributes.

3. Choosing Granularity

52
 
What level should the knowledge be represented and what are the primitives ?

Should there be a small number or should there be a large number of low-level
primitives or High-level facts.
 High-level facts may not be adequate for inference while Low-level primitives may
require a lot of storage.
Example of Granularity :

John spotted Sue. This could be represented as Spotted (agent(John), object (Sue))

4. Set of Objects

Certain properties of objects that is true as member of a set but not as individual.

This is done in different ways :

1. In logical representation through the use of universal quantifier, and

2. in hierarchical structure where node represent sets, the inheritance propagate set level
assertion down to individual.

5. Finding Right Structure

 Access to right structure for describing a particular situation.

 It requires, selecting an initial structure and then revising the choice.
 While doing so, it is necessary to solve following problems :
1. How to perform an initial selection of the most appropriate structure.
2. How to fill in appropriate details from the current situations.
3. How to find a better structure if the one chosen initially turns out not to be
appropriate.
4. What to do if none of the available structures is appropriate.
5. When to create and remember a new structure.
3. Explain in detail about min-max procedure and alpha beta cutoffs. (May 2012,
November 2012, May 2014)

ALPHA-BETA pruning is a method that reduces the number of nodes explored in

Minimax strategy. It reduces the time required for the search and it must be restricted so that
no time is to be wasted searching moves that are obviously bad for the current player.

The exact implementation of alpha-beta keeps track of the best move for each side as it
moves throughout the tree.

We proceed in the same (preorder) way as for the minimax algorithm. For the MIN

nodes, the score computed starts with +infinity and decreases with time.

For MAX nodes, scores computed starts with –infinity and increase with time.

53
The efficiency of the Alpha-Beta procedure depends on the order in which successors of a
node are examined. If we were lucky, at a MIN node we would always consider the nodes in
order from low to high score and at a MAX node the nodes in order from high to low score.
In general it can be shown that in the most favorable circumstances the alpha-beta search
opens as many leaves as minimax on a game tree with double its depth.

Alpha-Beta algorithm: The algorithm maintains two values, alpha and beta, which
represents the minimum score that the maximizing player is assured of and the maximum
score that the minimizing player is assured of respectively. Initially alpha is negative infinity
and beta is positive infinity. As the recursion progresses the "window" becomes smaller.
When beta becomes less than alpha, it means that the current position cannot be the result of
best play by both players and hence need not be explored further.

Pseudocode for the alpha-beta algorithm.

evaluate (node, alpha, beta)

if node is a leaf

return the heuristic value of node

if node is a minimizing node

for each child of node

beta = min (beta, evaluate (child, alpha, beta))

if beta <= alpha

return beta

return beta

if node is a maximizing node

for each child of node

alpha = max (alpha, evaluate (child, alpha, beta))

if beta <= alpha

return alpha

return alpha.

Min Max Algorithm

The Min-Max algorithm is applied in two player games, such as tic-tac-toe, checkers,
chess, go, and so on.

54
There are two players involved, MAX and MIN. A search tree is generated, depth-first,
starting with the current game position upto the end game position. Then, the final game
position is evaluated from MAX’s point of view, as shown in Figure 1. Afterwards, the inner
node values of the tree are filled bottom-up with the evaluated values. The nodes that belong
to the MAX player receive the maximum value of it’s children. The nodes for the MIN player
will select the minimun value of it’s children.

The values represent how good a game move is. So the MAX player will try to select the
move with highest value in the end. But the MIN player also has something to say about it
and he will try to select the moves that are better to him, thus minimizing MAX’s outcome.

Algorithm

MinMax (GamePosition game) {

return MaxMove (game);

MaxMove (GamePosition game) {

if (GameEnded(game)) {

return EvalGameState(game);

else {

best_move <- {};

moves <- GenerateMoves(game);

ForEach moves {

move <- MinMove(ApplyMove(game));

if (Value(move) > Value(best_move)) {

best_move <- move;

return best_move;

MinMove (GamePosition game) {

55
best_move <- {};

moves <- GenerateMoves(game);

ForEach moves {

move <- MaxMove(ApplyMove(game));

if (Value(move) > Value(best_move)) {

best_move <- move;

return best_move;

}Optimization

1. price
2. Limit the depth of the tree.
Speed up the algorithm

This all means that sometimes the search can be aborted because we find out that the
search subtree won’t lead us to any viable answer. This optimization is known as alpha-beta
cutoffs.

The algorithm Have two values passed around the tree nodes:

The alpha value which holds the best MAX value found;

The beta value which holds the best MIN value found.

At MAX level, before evaluating each child path, compare the returned value with of the
previous path with the beta value. If the value is greater than it abort the search for the current
node;

At MIN level, before evaluating each child path, compare the returned value with of the
previous path with the alpha value. If the value is lesser than it abort the search for the current
node.

4. Write an algorithm for converting to causal form. (June 2013)

To convert the axioms into conjunctive normal form (clause form)

1. Eliminate implications
2. Move negations down to the atomic formulas
3. Purge existential quantifiers
4. Rename variables if necessary
5. Move the universal quantifiers to the left

56
 6. Move the disjunctions down to the literals
7. Eliminate the conjunctions
8. Rename the variables if necessary
9. Purge the universal quantifiers
Example

Consider the sentence

“All music lovers who enjoy Bach either dislike Wagner or think that anyone who dislikes
any composer is a philistine''.

Use enjoy() for enjoying a composer, and for dislike.

"x[musiclover(x)  enjoy(x,Bach) 
    dislike(x,Wagner) ( "y[$z[dislike(y,z)] =>think-philistine(x,y)])]

Conversion

Step 1: Filter the expression to remove symbols.

"x [ (musiclover(x) enjoy(x,Bach)) => dislike(x,Wagner)  ( "y[ $z [dislike(y,z)]=> think-

philistine(x,y)])].

Step 2: Filter using the following relationships:

(  P) <=> P;
( a b ) <=>  a   b;
( a  b ) <=>  a   b; 
"x P(x) <=> $x  P(x); and 
$x P(x) <=>"x  P(x).

Now the expression becomes

"x[  musiclover(x)   enjoy(x,Bach)  dislike(x,Wagner)  "y[ "z[

dislike(y,z)]  think-philistine(x,y)]].

Step 3: Standardize variables so that each quantifier binds a unique variable.

"x Pred1(x) "x Pred2(x) 

becomes 
"x Pred1(x)  "y Pred2(y).

Step 4: Step 3 allows us to move all the quantifiers to the left in Step 4.

The expression becomes

"x"y"z[  musiclover(x)  enjoy(x,Bach) 

    dislike(x,Wagner)  dislike(y,z)  think-philistine(x,y)].

57
This is called the prenex normal form.

Step 5: Eliminate existential quantifiers, by arguing that if $y Composer(y), then if we could

actually find an Object S1 to replace the Variable x. So this gets replaced simply by
Composer(S1).

Now, if existential quantifiers exist within the scope of universal quantifiers, we can't use
merely an object, but rather a function that returns an object. The function will depend on the
univeral quantifier.

"x$y tutor-of(y,x) 
gets replaced by 
"x tutor-of(S2(x),x).

This process is called Skolemization, and the S2 is a Skolem function.

Step 6:  Any variable left must be universally quantified out on the left.

The expression becomes:

musiclover(x)   enjoy(x,Bach)      dislike(x,Wagner)   dislike(y,z)  think-

philistine(x,y).

Step 7: Convert everything into a conjunction of disjunctions using the associative,

commutative, distributive laws.

The form is

(a  b  c  d  ...)   ( p  q  ...)  ...

Step 8: Call each conjunct a separate clause. In order for the entire wff to be true, each clause
must be true separately.

Step 9: Standard apart the variables in the set of clauses generated in 7 and 8. This requires
renaming the variables so that no two clauses make reference to the same variable. All
variables are implicitly universally quantified to the left.

"x P(x)  Q(x) <=> "xP(x)  "xQ(x) <=> "xP(x)   "yQ(y)

This completes the Procedure.

After application to a set of wffs, we end up with a set of clauses each of which is a
disjunction of literals.

5. Explain the various approaches to knowledge representation. (June 2013)

A good knowledge representation enables fast and accurate access to knowledge and
understanding of the content.

A knowledge representation system should have following properties.

58
 1. Representational Adequacy: The ability to represent all kinds of knowledge that are
needed in that domain.
2. Inferential Adequacy: The ability to manipulate the representational structures to
derive new structure corresponding to new knowledge inferred from old.
3. Inferential Efficiency: The ability to incorporate additional information into the
knowledge structure that can be used to focus the attention of the inference
mechanisms in the most promising direction.
4. Acquisitional Efficiency: The ability to acquire new knowledge using automatic
methods wherever possible rather than reliance on human intervention.
Knowledge Representation Schemes

There are four types of Knowledge representation: Relational, Inheritable, Inferential,

and Declarative/Procedural.

1. Relational Knowledge:

 Provides a framework to compare two objects based on equivalent attributes.

 Any instance in which two different objects are compared is a relational type of
knowledge.
2. Inheritable Knowledge

 Is obtained from associated objects.

 It prescribes a structure in which new objects are created which may inherit all or a
subset of attributes from existing objects.
3. Inferential Knowledge

 Is inferred from objects through relations among objects.

 e.g., a word alone is a simple syntax, but with the help of other words in phrase the
reader may infer more from a word; this inference within linguistic is called
semantics.
4. Declarative Knowledge

 A statement in which knowledge is specified, but the use to which that knowledge is
to be put is not given.
 e.g. laws, people's name; these are facts which can stand alone, not dependent on
other knowledge;

Procedural Knowledge

 A representation in which the control information, to use the knowledge, is embedded

in the knowledge itself.
 e.g. computer programs, directions, and recipes; these indicate specific use or
implementation.

59
 UNIT III KNOWLEDGE INFERENCE
PART A

Knowledge representation -Production based system, Frame based system.

Inference - Backward chaining, Forward chaining, Rule value approach, Fuzzy
reasoning - Certainty factors, Bayesian Theory-Bayesian Network-Dempster - Shafer
theory.

1. What is a production based system. (December 2011)

 A production based system is one
 That makes use of the production rules.
 Advantages of production rules:
 They are modular,
 Each rule defines a small and independent piece of knowledge.
 New rules may be added and old ones deleted
 Rules are usually independently of other rules.
2. State Bayesian Probability. (December 2011)
 Bayesian probability is one of the most popular interpretations of the concept
of probability.
 The Bayesian interpretation of probability can be seen as an extension of logic
that enables reasoning with uncertain statements.
 To evaluate the probability of a hypothesis, the Bayesian probability specifies
some prior probability, which is then updated in the light of new relevant data.
 The Bayesian interpretation provides a standard set of procedures and
formulae to perform this calculation.
3. What is non-monotonic inference? (May 2012)
Non monotonic inference is one that uses axioms and rules that are expanded so that
complete information can be got from incomplete data.

This system preserves the property that, at any instant of time, a statement is either
believed to be true or false.

4. What is meant by rule based system. (November 2012)

 The rule based system is one that uses several rules as evidences and relates it to a
single hypothesis.
 They use the hypothesis and evidences collected by various inferences and
function.

5. When the knowledge is uncertain. How will you handle?

(June 2013)

1. No monotonic logic
2. Probabilistic reasoning

60
 3. Fuzzy logic
4. Truth values.

6. What is a Frame? (June 2013)

a. A frame is a data structure with typical knowledge about a particular object or
concept.
b. Frames were first proposed by Marvin Minsky in the 1970s.
c. Each frame has its own name and a set of attributes associated with it. Name,
weight, height and age are slots in the frame Person.
d. Model, processor, memory and price are slots in the frame Computer. Each
attribute or slot has a value attached to it.
e. Frames provide a natural way for the structured and concise representation of
knowledge.
7. Define certainty factor. (November 2013)
A certainty factor CF(h,e) is defined in terms of two components.

i. MB(h,e) is a measure of belief in hypothesis h given the

evidence.
1. MB measures the extent to which the evidence
supports the hypothesis else it is zero it the evidence
fails to support the hypothesis.
ii. MD(h,e) is the measure of disbelief in hypothesis h given
the evidence e.
1. MD measures the extent to which the evidence
supports the negation else it is zero if the evidence
fails to support the hypothesis.
iii. These two factors define certainty as
1. CF(h,e)=MB(h,e)-MD(h,e)
8. When will you opt for a fuzzy system? (May 2014)
Fuzzy system is used when

1. The input information is noisy.

2. To solve complex problems.
3. Mathematical calculations are involved.
9. Define Baye’s theorem. (May 2014)
A formula for determining conditional probability named after 18th-century British
mathematician Thomas Bayes. The theorem provides a way to revise existing predictions
or theories given new or additional evidence. In finance, Bayes' Theorem can be used to
rate the risk of lending money to potential borrowers.
Formula:

Also called "Bayes' Rule."

61
 10. State the use of inference engine. (May 2014)
 An inference engine interprets and evaluates the facts in the knowledge base in
order to provide an answer
 The inference engine enables the expert system to draw deductions from the rules
in the KB
11. Define fuzzy logic.
FL is a problem-solving control system methodology that lends itself to
implementation in systems ranging from simple, small, embedded micro-controllers to large,
networked, multi-channel PC or workstation-based data acquisition and control systems. It
can be implemented in hardware, software, or a combination of both. FL provides a simple
way to arrive at a definite conclusion based upon vague, ambiguous, imprecise, noisy, or
missing input information. FL's approach to control problems mimics how a person would
make decisions, only much faster.

12. Define ontological engineering.

Ontology engineering in computer science and information science is a field which
studies the methods and methodologies for building ontologies: formal representations of a
set of concepts within a domain and the relationships between those concepts. A large-scale
representation of abstract concepts such as actions, time, physical objects and beliefs would
be an example of ontological engineering.
13. List down the applications of Bayesian network.
 Turbo code generation.
 Build medical diagnosis system.
 Implementation of spam filters.
14. List the advantages of production rules.
Advantages of production rules:

 They are modular,

 Each rule defines a small and independent piece of knowledge.
 New rules may be added and old ones deleted
 Rules are usually independently of other rules.
15. Define certainty factor. (November 2013)
A certainty factor CF(h,e) is defined in terms of two components.

i. MB(h,e) is a measure of belief in hypothesis h given the

evidence.
1. MB measures the extent to which the evidence
supports the hypothesis else it is zero it the evidence
fails to support the hypothesis.
ii. MD(h,e) is the measure of disbelief in hypothesis h given
the evidence e.

62
 1. MD measures the extent to which the evidence
supports the negation else it is zero if the evidence
fails to support the hypothesis.
iii. These two factors define certainty as
1. CF(h,e)=MB(h,e)-MD(h,e)

PART B

1. Explain about forward chaining and backward chaining with proper example.
(June 2013, November 2013)
FORWARD CHAINING

Forward chaining working from the facts to a conclusion. Sometimes called the data driven
approach. To chain forward, match data in working memory against 'conditions‘ of rules in
the rule-base. Starts with the facts, and sees what rules apply (and hence what should be
done) given the facts.

WORKING

Facts are held in a working memory. Condition-action rules represent actions to take when
specified facts occur in working memory. Typically the actions involve adding or deleting
facts from working memory.

STEPS

•To chain forward, match data in working memory against 'conditions' of rules in the
rule base.

•When one of them fires, this is liable to produce more data.

•So the cycle continues up to conclusion.

Example

63
 •Here are two rules:

•If corn is grown on poor soil, then it will get blackfly.

•If soil hasn't enough nitrogen, then it is poor soil.

Forward chaining algorithm

Repeat

Collect the rule whose condition matches a fact in WM.

Do actions indicated by the rule.

(add facts to WM or delete facts from WM)

Until problem is solved or no condition match

Apply on the Example 2 extended (adding 2 more rules and 1 fact)

Rule R1 : IF hot AND smoky THEN ADD fire

Rule R2 : IF alarm_beeps THEN ADD smoky

Rule R3 : If fire THEN ADD switch_on_sprinklers

Rule R4 : IF dry THEN ADD switch_on_humidifier

Rule R5 : IF sprinklers_on THEN DELETE dry

Fact F1 : alarm_beeps [Given]

Fact F2 : hot [Given]

Fact F2 : Dry [Given]

Now, two rules can fire (R2 and R4)

Rule R4 ADD humidifier is on [from F2]

ADD smoky [from F1]

ADD fire [from F2 by R1]

ADD switch_on_sprinklers [by R3]

Rule R2

[followed by

sequence of

64
 actions]

DELEATE dry, ie

humidifier is off a conflict !

[by R5 ]

BACKWARD CHAINING

Backward chaining: working from the conclusion to the facts. Sometimes called the
goal-driven approach. Starts with something to find out, and looks for rules that will help in
answering it goal driven.

Steps in BC

•To chain backward, match a goal in working memory against 'conclusions' of rules in
the rule-base.

•When one of them fires, this is liable to produce more goals.

•So the cycle continues

Example

•Same rules:

•If corn is grown on poor soil, then it will get blackfly.

•If soil hasn't enough nitrogen, then it is poor soil.

■ Backward chaining algorithm

Prove goal G

If G is in the initial facts , it is proven.

Otherwise, find a rule which can be used to conclude G, and

try to prove each of that rule's conditions.

Encoding of rules

Rule R1 : IF hot AND smoky THEN fire

Rule R2 : IF alarm_beeps THEN smoky

Rule R3 : If fire THEN switch_on_sprinklers

Fact F1 : hot [Given]

Fact F2 : alarm_beeps [Given]

65
 Goal : Should I switch sprinklers on?

Example 1

Rule R1 : IF hot AND smoky THEN fire

Rule R2 : IF alarm_beeps THEN smoky

Rule R3 : IF fire THEN switch_on_sprinklers

Fact F1 : alarm_beeps [Given]

Fact F2 : hot [Given]

Example 2

Rule R1 : IF hot AND smoky THEN ADD fire

Rule R2 : IF alarm_beeps THEN ADD smoky

Rule R3 : IF fire THEN ADD switch_on_sprinklers

Fact F1 : alarm_beeps [Given]

Fact F2 : hot [Given]

Example 3 : A typical Forward Chaining

Rule R1 : IF hot AND smoky THEN ADD fire

Rule R2 : IF alarm_beeps THEN ADD smoky

Rule R3 : If fire THEN ADD switch_on_sprinklers

Fact F1 : alarm_beeps [Given]

66
 Fact F2 : hot [Given]

Fact F4 : smoky [from F1 by R2]

Fact F2 : fire [from F2, F4 by R1]

Fact F6 : switch_on_sprinklers [from F2 by R3]

Example 4 : A typical Backward Chaining

Rule R1 : IF hot AND smoky THEN fire

Rule R2 : IF alarm_beeps THEN smoky

Rule R3 : If _fire THEN switch_on_sprinklers

Fact F1 : hot [Given]

Fact F2 : alarm_beeps [Given]

Goal : Should I switch sprinklers on?

2. Explain in detail about production and frame based system.

PRODUCTION BASED SYSTEM

A Knowledge representation formalism consists of collections of condition-action

rules(Production Rules or Operators), a database which is modified in accordance with
the rules, and a Production System Interpreter which controls the operation of the rules
i.e The 'control mechanism' of a Production System, determining the order in which
Production Rules are fired.A system that uses this form of knowledge representation is
called a production system.

A production system consists of rules and factors. Knowledge is encoded in a

declarative from which comprises of a set of rules of the form

Situation ------------ Action

SITUATION that implies ACTION.

Example:-

IF the initial state is a goal state THEN quit.

The major components of an AI production system are

 A global database
 A set of production rules and
 A control system
The goal database is the central data structure used by an AI production system. The
production rules operate on the global database. Each rule has a precondition that is

67
either satisfied or not by the database. If the precondition is satisfied, the rule can be
applied. Application of the rule changes the database. The control system chooses which
applicable rule should be applied and ceases computation when a termination condition
on the database is satisfied. If several rules are to fire at the same time, the control
system resolves the conflicts.

Four classes of production systems:-

 A monotonic production system

 A non monotonic production system
 A partially commutative production system
 A commutative production system.
Advantages of production systems:-

 Production systems provide an excellent tool for structuring AI programs.

 Production Systems are highly modular because the individual rules can be
added, removed or modified independently.
 The production rules are expressed in a natural form, so the statements contained
in the knowledge base should the a recording of an expert thinking out loud.
Disadvantages of Production Systems:-

 One important disadvantage is the fact that it may be very difficult analyse the
flow of control within a production system because the individual rules don’t call
each other.
 Production systems describe the operations that can be performed in a search for a
solution to the problem.
 They can be classified as follows.
Monotonic production system :-

 A system in which the application of a rule never prevents the later application of
another rule, that could have also been applied at the time the first rule was
selected.
Partially commutative production system:-

 A production system in which the application of a particular sequence of rules

transforms state X into state Y, then any permutation of those rules that is
allowable also transforms state x into state Y.
Theorem proving falls under monotonic partially communicative system. Blocks
world and 8 puzzle problems like chemical analysis and synthesis come under
monotonic, not partially commutative systems. Playing the game of bridge comes under
non monotonic , not partially commutative system.

Partially commutative , monotonic production systems are useful for solving

ignorable problems. Monotonic partially commutative systems are useful for problems in
which changes occur but can be reversed and in which the order of operation is not
critical (ex: 8 puzzle problem).

68
 Production systems that are not partially commutative are useful for many problems
in which irreversible changes occur, such as chemical analysis.

FRAME BASED SYSTEM.

A frame is a data structure with typical knowledge about a particular object or

concept. Frames were first proposed by Marvin Minsky in the 1970s.

Each frame has its own name and a set of attributes associated with it. Name, weight, height
and age are slots in the frame Person.

Model, processor, memory and price are slots in the frame Computer. Each attribute or slot
has a value attached to it.

Frames provide a natural way for the structured and concise representation of knowledge.

A frame provides a means of organising knowledge in slots to describe various attributes and
characteristics of the object.

Frames are an application of object-oriented programming for expert systems.

EXAMPLE

BOARDING PASS FRAME

Object-oriented programming is a programming method that uses objects as a basis for

analysis, design and implementation.

In object-oriented programming, an object is defined as a concept, abstraction or

thing with crisp boundaries and meaning for the problem at hand. All objects have identity
and are clearly distinguishable. Michael Black, Audi 5000 Turbo, IBM Aptiva S35 are
examples of objects.

An object combines both data structure and its behaviour in a single entity.

69
When an object is created in an object-oriented programming language, we first assign a
name to the object, and then determine a set of attributes to describe the object’s
characteristics, and at last write procedures to specify the object’s behaviour.

A knowledge engineer refers to an object as a frame (the term, which has become the
AI jargon).Frames as a knowledge representation technique

The concept of a frame is defined by a collection of slots. Each slot describes a

particular attribute or operation of the frame.

Slots are used to store values. A slot may contain a default value or a pointer to another
frame, a set of rules or procedure by which the slot value is obtained.

Typical information included in a slot

Frame name.Relationship of the frame to the other frames.

The frame IBM Aptiva S35 might be a member of the class Computer, which in turn
might belong to the class Hardware.

Slot value.

A slot value can be symbolic, numeric or Boolean. For example, the slot Name has
symbolic values, and the slot Age numeric values. Slot values can be assigned when the
frame is created or during a session with the expert system.

Default slot value. The default value is taken to be true when no evidence to the contrary
has been found. For example, a car frame might have four wheels and a chair frame four legs
as default values in the corresponding slots.

Range of the slot value. The range of the slot value determines whether a particular object
complies with the stereotype requirements defined by the frame. For example, the cost of a
computer might be specified between $750 and $1500.

Procedural information. A slot can have a procedure attached to it, which is executed if the
slot value is changed or needed.

Class and instances

The word frame often has a vague meaning. The frame may refer to a particular
object, for example the computer IBM Aptiva S35, or to a group of similar objects. To be
more precise, we will use the instance-frame when referring to a particular object, and the
class-frame when referring to a group of similar objects.

A class-frame describes a group of objects with common attributes. Animal, person, car and
computer are all class-frames.

70
Each frame “knows” its class.

EXAMPLE : COMPUTER CLASS

INSTANCES OF A CLASS

Class inheritance in frame-based systems

Frame-based systems support class inheritance.

The fundamental idea of inheritance is that attributes of the class-frame represent things that
are typically true for all objects in the class. However, slots in the instance-frames can be
filled with actual data uniquely specified for each instance.

Relationships among objects

1. Generalisation denotes a-kind-of or is-a relationship between superclass and its

subclasses. For example, a car is a vehicle, or in other words, Car represents a
subclass of the more general superclass Vehicle. Each subclass inherits all features of
the superclass

71
 2. Aggregation is a-part-of or part-whole relationship in which several subclasses
representing components are associated with a superclass representing a whole. For
example, an engine is a part of a car.

3. Association describes some semantic relationship between different classes which are
unrelated otherwise. For example, Mr Black owns a house, a car and a computer.
Such classes as House, Car and Computer are mutually independent, but they are
linked with the frame Mr Black through the semantic association.

Methods and demons

Expert systems are required not only to store the knowledge but also to validate and
manipulate this knowledge. To add actions to our frames, we need methods and demons.

A method is a procedure associated with a frame attribute that is executed whenever

requested.

We write a method for a specific attribute to determine the attribute’s value or execute a
series of actions when the attribute’s value changes.

Most frame-based expert systems use two types of methods:

WHEN CHANGED and WHEN NEEDED.

72
3. Explain the need of fuzzy set and fuzzy logic with example. (May-12)

What is Fuzzy Set ?

 The word "fuzzy" means "vagueness". Fuzziness occurs when the boundary of a piece of
information is not clear-cut.
 Fuzzy sets have been introduced by Lotfi A. Zadeh (1965) as an extension of the classical
notion of set.
 Classical set theory allows the membership of the elements in the set in binary terms, a
bivalent condition - an element either belongs or does not belong to the set.
 Fuzzy set theory permits the gradual assessment of the membership of elements in a set,
described with the aid of a membership function valued in the real unit interval [0, 1].
Fuzzy Set Theory

Fuzzy set theory is an extension of classical set theory where elements have varying
degrees of membership. A logic based on the two truth values, True and False, is sometimes
inadequate when describing human reasoning. Fuzzy logic uses the whole interval between
0 (false) and 1 (true) to describe human reasoning.

 Fuzzy logic is derived from fuzzy set theory dealing with reasoning that is approximate
rather than precisely deduced from classical predicate logic.
 Fuzzy logic is capable of handling inherently imprecise concepts.
 Fuzzy set theory defines Fuzzy Operators on Fuzzy Sets.
Crisp and Non-Crisp Set

 The characteristic function μA(x) which has values 0 ('false') and 1 ('true'') only are
crisp sets.
 For Non-crisp sets the characteristic function μA(x) can be defined.
The characteristic function μA(x) for the crisp set is generalized for the Non-crisp
sets.

 This generalized characteristic function μA(x) is called membership function.

 Such Non-crisp sets are called Fuzzy Sets.
 Crisp set theory is not capable of representing descriptions and classifications in
many cases, In fact, Crisp set does not provide adequate representation for most
cases.

Representation of Crisp and Non-Crisp Set

Example:

Classify students for a basketball team

This example explains the grade of truth value.

 tall students qualify and not tall students do not qualify

73
  if students 1.8 m tall are to be qualified, then should we exclude a student who is
1/10" less? Or should we exclude a student who is 1" shorter?
 Non-Crisp Representation to represent the notion of a tall person

Fig. 1 Set Representation – Degree or grade of truth

 A student of height 1.79m would belong to both tall and not tall sets with a particular
degree of membership.
 As the height increases the membership grade within the tall set would increase whilst
the membership grade within the not-tall set would decrease.
Capturing Uncertainty

Instead of avoiding or ignoring uncertainty, Lotfi Zadeh introduced Fuzzy Set theory that
captures uncertainty.

 A fuzzy set is described by a membership function μA (x) of A. This membership

function associates to each element xσ Є X a number as μA (xσ ) in the closed unit
interval [0, 1].
 The number μA (xσ ) represents the degree of membership of xσ in A.
 The notation used for membership function μA (x) of a fuzzy set A is
Α : Χ Є [0, 1]

 Each membership function maps elements of a given universal base set X , which is
itself a crisp set, into real numbers in [0, 1] .
Example:

74
 Fig. 2 Membership function of a Crisp set C and Fuzzy set F

 In the case of Crisp Sets the members of a set are : either out of the set, with
membership of degree " 0 ", or in the set, with membership of degree " 1 ",
 Therefore, Crisp Sets Є Fuzzy Sets
In other words, Crisp Sets are Special cases of Fuzzy Sets.

Examples of Crisp and Non-Crisp Set

Example 1: Set of prime numbers ( a crisp set)

 If we consider space X consisting of natural numbers ≤ 12 ie X = {1, 2, 3, 4, 5,

6, 7, 8, 9, 10, 11, 12}
 Then, the set of prime numbers could be described as follows.
PRIME = {x contained in X | x is a prime number} = {2, 3, 5, 6, 7, 11}

Fuzzy Set

A Fuzzy Set is any set that allows its members to have different degree of
membership, called membership function, in the interval [0 , 1].

Definition of Fuzzy set

A fuzzy set A, defined in the universal space X, is a function defined in X which

assumes values in the range [0, 1].

A fuzzy set A is written as a set of pairs {x, A(x)} as

A = {{x , A(x)}} , x in the set X

where x is an element of the universal space X, and

A(x) is the value of the function A for this element.

The value A(x) is the membership grade of the element x in a fuzzy set A.

Example : Set SMALL in set X consisting of natural numbers ≤ to 12.

Assume: SMALL(1) = 1, SMALL(2) = 1, SMALL(3) = 0.9, SMALL(4) = 0.6,

SMALL(5) = 0.4, SMALL(6) = 0.3, SMALL(7) = 0.2, SMALL(8) = 0.1,

SMALL(u) = 0 for u >= 9.

Then, following the notations described in the definition above :

Set SMALL = {{1, 1 }, {2, 1 }, {3, 0.9}, {4, 0.6}, {5, 0.4}, {6, 0.3}, {7, 0.2},
{8, 0.1}, {9, 0 }, {10, 0 }, {11, 0}, {12, 0}}

Note that a fuzzy set can be defined precisely by associating with each x , its grade of
membership in SMALL.

75
• Definition of Universal Space

Originally the universal space for fuzzy sets in fuzzy logic was defined only on the
integers. Now, the universal space for fuzzy sets and fuzzy relations is defined with
three numbers.

The first two numbers specify the start and end of the universal space, and the third
argument specifies the increment between elements.

This gives the user more flexibility in choosing the universal space.

Example : The fuzzy set of numbers, defined in the universal space

X = { xi } = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12} is presented as SetOption [FuzzySet,

UniversalSpace → {1, 12, 1}]

Fuzzy Membership

 A fuzzy set A defined in the universal space X is a function defined in X which

assumes values in the range [0, 1].
 A fuzzy set A is written as a set of pairs {x, A(x)}.
A = {{x , A(x)}} , x in the set X

where x is an element of the universal space X, and

A(x) is the value of the function A for this element.

The value A(x) is the degree of membership of the element x in a fuzzy set A. The
Graphic Interpretation of fuzzy membership for the fuzzy sets :

• Graphic Interpretation of Fuzzy Sets SMALL

The fuzzy set SMALL of small numbers, defined in the universal space

X = { xi } = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12} is presented as

SetOption [Fuzzy Set, Universal Space → {1, 12, 1}]

The Set SMALL in set X is : SMALL = FuzzySet {{1, 1 }, {2, 1 }, {3, 0.9}, {4, 0.6}, {5, 0.4},
{6, 0.3}, {7, 0.2}, {8, 0.1}, {9, 0 }, {10, 0 }, {11, 0}, {12, 0}}

Therefore SetSmall is represented as

SetSmall = FuzzySet [{{1,1},{2,1}, {3,0.9}, {4,0.6}, {5,0.4},{6,0.3}, {7,0.2}, {8, 0.1},

{9, 0}, {10, 0}, {11, 0}, {12, 0}} , UniversalSpace → {1, 12, 1}]

Fuzzy Plot [ SMALL, AxesLable → {"X", "SMALL"}]

76
 Graphic Interpretation of Fuzzy Sets EMPTY

 An empty set is a set that contains only elements with a grade of membership equal to
0.
Example: Let EMPTY be a set of people, in Minnesota, older than 120.

 The Empty set is also called the Null set.

 The fuzzy set EMPTY , defined in the universal space X = { xi } = {1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12} is presented as SetOption [FuzzySet, UniversalSpace → {1, 12, 1}]
 The Set EMPTY in set X is : EMPTY = FuzzySet {{1, 0}, {2, 0}, {3, 0}, {4, 0}, {5,
0}, {6, 0}, {7, 0}, {8, 0}, {9, 0}, {10, 0}, {11, 0}, {12, 0}}
 Therefore SetEmpty is represented as SetEmpty = FuzzySet [{{1,0},{2,0}, {3,0},
{4,0}, {5,0},{6,0}, {7,0}, {8, 0}, {9, 0}, {10, 0}, {11, 0}, {12, 0}} , UniversalSpace →
{1, 12, 1}]
 FuzzyPlot [ EMPTY, AxesLable ={"X", " UNIVERSAL SPACE "}]

4. Explain how Bayesian network is used for representing knowledge in an uncertain

domain.

 Bayesian networks (BN) are a network-based framework for representing and

analyzing models involving uncertainty;
 BN are different from other knowledge-based systems tools because uncertainty is
handled in mathematically rigorous yet efficient and simple way.
 BN are different from other probabilistic analysis tools because of network
representation of problems, use of Bayesian statistics, and the synergy between these.
Definition of a Bayesian Network

Knowledge structure:

 variables are nodes

 arcs represent probabilistic dependence between variables
 conditional probabilities encode the strength of the dependencies
Computational architecture:

 computes posterior probabilities given evidence about some nodes

 exploits probabilistic independence for efficient computation

77
  A Bayesian network is a representation of the joint distribution over all the variables
represented by nodes in the graph. Let the variables be X(1), ..., X(n).
 Let parents(A) be the parents of the node A. Then the joint distribution for X(1)
through X(n) is represented as the product of the probability distributions
P(Xi|Parents(Xi)) for i = 1 to n. If X has no parents, its probability distribution is
said to be unconditional, otherwise it is conditional.

By the chaining rule of probability, the joint probability of all the nodes in the graph above
is:

P(C, S, R, W) = P(C) * P(S|C) * P(R|C) * P(W|S,R)

W=Wet Grass, C=Cloudy, R=Rain, S=Sprinkler

Example: P(W∩-R∩S∩C)

= P(W|S,-R)*P(-R|C)*P(S|C)*P(C)

= 0.9*0.2*0.1*0.5 = 0.009

What is the probability of wet grass on a given day - P(W)?

P(W) = P(W|SR) * P(S) * P(R) +

P(W|S-R) * P(S) * P(-R) +

P(W|-SR) * P(-S) * P(R) +

P(W|-S-R) * P(-S) * P(-R)

78
Here P(S) = P(S|C) * P(C) + P(S|-C) * P(-C)

P(R) = P(R|C) * P(C) + P(R|-C) * P(-C)

P(W)= 0.5985

Advantages of Bayesian Approach

 Bayesian networks can readily handle incomplete data sets.

 Bayesian networks allow one to learn about causal relationships
 Bayesian networks readily facilitate use of prior knowledge.

5. Explain in detail about Ruled based approach and Dempster –Shafer theory.

Ruled based approach

Rule-Based system architecture consists a set of rules, a set of facts, and an inference
engine.

Types of Rules

Three types of rules are mostly used in the Rule-based production systems.

 Knowledge Declarative Rules :

These rules state all the facts and relationships about a problem.

Example :

IF inflation rate declines

THEN the price of gold goes down.

These rules are a part of the knowledge base.

 Inference Procedural Rules

These rules advise on how to solve a problem, while certain facts are known.

Example :

IF the data needed is not in the system

THEN request it from the user.

These rules are part of the inference engine.

 Meta rules
These are rules for making rules. Meta-rules reason about which rules should be
considered for firing.

Example :

79
 IF the rules which do not mention the current goal in their premise,

AND there are rules which do mention the current goal in their premise,

THEN the former rule should be used in preference to the latter.

Meta-rules direct reasoning rather than actually performing reasoning.

Meta-rules specify which rules should be considered and in which order they should be
invoked.

FACTS

They represent the real world information

Inference Engine

The inference engine uses one of several available forms of inferencing. By inferencing
means the method used in a knowledge-based system to process the stored knowledge
and supplied data to produce correct conclusions.

Example

How old are you?

Subtract the year you were born in from 2014.
The answer will either be exactly right,
Or one year short.
Dempster/Shafer theory

 The Dempster-Shafer theory, also known as the theory of belief functions, is a

generalization of the Bayesian theory of subjective probability.
 The Bayesian theory requires probabilities for each question of interest, belief
functions allow us to base degrees of belief for one question on probabilities for a
related question.
 These degrees of belief may or may not have the mathematical properties of
probabilities; how much they differ from probabilities will depend on how closely the
two questions are related.
 The Dempster-Shafer theory owes its name to work by A. P. Dempster (1968) and
Glenn Shafer (1976), but the kind of reasoning the theory uses can be found as far
back as the seventeenth century.
 The theory came to the attention of AI researchers in the early 1980s, when they were
trying to adapt probability theory to expert systems.
 Dempster-Shafer degrees of belief resemble the certainty factors in MYCIN, and this
resemblance suggested that they might combine the rigor of probability theory with
the flexibility of rule-based systems.
 Subsequent work has made clear that the management of uncertainty inherently
requires more structure than is available in simple rule-based systems, but the
Dempster-Shafer theory remains attractive because of its relative flexibility.

80
 The Dempster-Shafer theory is based on two ideas: the idea of obtaining degrees of
belief for one question from subjective probabilities for a related question, and
Dempster's rule for combining such degrees of belief when they are based on
independent items of evidence.

81
 UNIT IV PLANNING AND MACHINE LEARNING

Basic plan generation systems - Strips -Advanced plan generation systems – K strips
-Strategic explanations -Why, Why not and how explanations. Learning- Machine
learning, adaptive Learning.

PART A

1. What are the various planning systems? (November 2012)

Other planning techniques are

1. Triangle tables – provides a way to record the goals that each operator is
expected to satisfy as well as the goals that must be true for it to execute
correctly. If something unexpected happens during the execution of a plan, the
table provides information required to patch the plan.
2. Meta planning – a technique for reasoning not just the problem being solved
but also planning process itself.
3. Macro operators – allow a planner to build new operators that represent
commonly used sequences of operators.
4. Case based planning – Reuse old plans to make new ones.
2. Define plan. (June 2013)
Plan is a way of decomposing a problem into some parts and the various
techniques can be applied on it to handle the various interactions as they are
detected during the problem solving process.
3. What do you mean by goal stack planning? (June 2013)
i. One of the earliest techniques is planning using goal stack.
ii. Problem solver uses single stack that contains
1. sub goals and operators both
2. Sub goals are solved linearly and then finally the conjoined sub
goal is solved.
iii. Plans generated by this method will contain
1. Complete sequence of operations for solving one goal followed
by complete sequence of operations for the next etc.
iv. Problem solver also relies on
1. A database that describes the current situation.
2. Set of operators with precondition, add and delete lists.
4. Define machine learning. (June 2013, November 2013)
Machine learning is a type of artificial intelligence (AI) that provides
computers with the ability to learn without being explicitly programmed.
Machine learning focuses on the development of computer programs that can
teach themselves to grow and change when exposed to new data. 
5. What is planning? (November 2013)

82
 Planning refers to the use of methods that focus on ways of decomposing
the original problem into appropriate subparts and on ways of recording and
handling interaction among the subparts as they are detected during the problem
solving process.

6. State the components of a planning system. (May 2014)

1. Choosing rules to apply.
2. Applying rules.
3. Detecting a solution.
4. Detecting dead ends.
5. Repairing an almost correct solution.
7. Define reactive system.
 A reactive system is one that has access to a knowledge base of some
sort that describes the actions to be taken under different
circumstances.
 It does not have the power of anticipation.
8. How is the complexity of learning measured?
The complexity of learning is a function of 3 factors
1. Error tolerance (h).
2. The number of binary features present in the example(t)
3. The size of the rule necessary to make the discrimination.(f).
9. What is explanation based learning? May-10
Explanation based learning is a method for extracting rules from individual
observations through an explanation.

General rule follows logically from the background of the Cavemen’s usual
cooking process.

10. List the various issues that affect the design of a learning element.
There are several factors affecting the performance. They are,

 Types of training provided.

 The form and extent of any initial background knowledge.
 The type of feedback provided.
 The learning algorithms used.
11. List the Features of ideal planner

 The planner should be able to represent the states, goals, and actions
 The planner should be able to add new actions at any time.
 The planner should be able to use divide and conquer method for solving very
big problems
12. What is Q learning?
Q-learning uses temporal differences to estimate the value of Q*(s,a).

83
 In Q-learning, the agent maintains a table of Q[S,A], where S is the set of states
and A is the set of actions. Q[s,a] represents its current estimate of Q*(s,a).

An experience s,a,r,s' provides one data point for the value of Q(s,a). The data point is
that the agent received the future value of r+ γV(s'), where V(s') =maxa' Q(s',a'); this
is the actual current reward plus the discounted estimated future value. This new data
point is called a return. The agent can use the temporal difference equation to update
its estimate for Q(s,a):

Q[s,a] ←Q[s,a] + α(r+ γmaxa' Q[s',a'] - Q[s,a])

or, equivalently,

Q[s,a] ←(1-α) Q[s,a] + α(r+ γmaxa' Q[s',a']).

This assumes that α is fixed; if α is varying, there will be a different count for each
state-action pair and the algorithm would also have to keep track of this count.

13. Explain the concept of learning with example.

Learning involves
 Task
The behavior or task that is being improved

 Data
The experiences that are used to improve performance in the task

 Measure of improvement
How the improvement is measured - for example, new skills that
were not present initially, increasing accuracy in prediction, or improved
speed

14. Differentiate between offline, online and active learning.

 In offline learning, all of the training examples are available to an agent before it
needs to act.
 In online learning, training examples arrive as the agent is acting. An agent that
learns online requires some representation of its previously seen examples before it
has seen all of its examples. As new examples are observed, the agent must update its
representation. Typically, an agent never sees all of its examples.
 Active learning is a form of online learning in which the agent acts to acquire useful
examples from which to learn. In active learning, the agent reasons about which
examples would be useful to learn from and acts to collect these examples.
15. Distinguish between supervised learning and unsupervised learning.
Dec-11, May- 13
Supervised learning involves learning a function from example of its input and
outputs where as unsupervised learning involves learning patterns in the input when
no specific output values are supplied.

84
 In supervised learning the environment is fully observable where as in
unsupervised learning the environment is partially observable.

PART B

1. Briefly explain the advanced plan generation systems. (December 2011)

Nonlinear Planning using Constraint Posting

 Idea of constraint posting is to build up a plan by incrementally

 Hypothesizing operators,
 Partial ordering between operators and
 Binding of variables within operators
 At any given time in planning process, a solution is a partially ordered, partially
instantiated set of operators.
 To generate actual plan, convert the partial order into any of a number of total orders.
Main steps involved in non linear plan generation are:

 Step addition - Creating new operator (step) for a plan.

 Promotion - Constraining one operator to come before another in final plan
 Declobbering - Placing operator Op2 between two operators Op1 and Op3
such that Op2 reasserts some pre conditions of Op3 that was negated by
Op1
 Simple Establishment Assigning a value to a variable, in order to ensure
the pre conditions of some step.
Algorithm:

 Initialize S to be set of propositions in the goal state.

 Remove some unachieved proposition P from S.
 Achieve P by using step addition, promotion, declobbering, simple establishment.
 Review all the steps in the plan, including any new steps introduced by step addition
to see if any of their preconditions are unachieved.
 Add to S the new set of unachieved preconditions.
 If S = φ, complete the plan by converting the partial order of steps into a total order
and instantiate any variables as necessary and exit.
 Otherwise go to step 2.
Hierarchical planning

 Main problem in strips-like planning (as well as in other planning

frameworks): complexity.
 One reason for complexity: no structure
 No distinction between important and unimportant properties
 No distinction between important and unimportant operators.
 This observation gives raise to two different ways of abstraction in planning:
i. Abstraction of situations and
ii. Abstraction of operators.

85
Abstraction of Situations - abstrips (Abstraction-Based strips)

Main idea: introduce weights for each literal and consider only the most important ones in
first loop, then refine by considering also literals with second highest weight.

In blocksworld, for instance, properties with weights:

Property Weight

On 4

Clear 3

Holds 2

Ontable 2

Handempty 1

Higher weights indicate more important properties. Means here concretely: first consider only
the property On, in the second loop the properties On and Clear and so on. Use the abstract
plan for a refinement of the more detailed plans. In the last loop all properties have to be
considered.

Algorithm

Assume non-linear planning algorithm NLP with input: partial nonlinear plan & output: total
well-formed conflict-free non-linear plan.

NLPAS (Non-Linear Planning with Abstraction of Situations) calls NLP initially considering
only the most important properties. This plan is refined by taking into account less important
properties.

Input: non-linear plan and index (initially the highest weight)

begin

if Index=0 then return(P)

else

disregard all preconditions of operators,

whose weights are smaller than Index

select P’:=NLP(P) ;;; arbitrary choice → backtrack point

endif

86
return(NLPAS(P’,Index-1))

end

Problems with abstraction

Abstraction certainly plays an important role in human problem solving. Selection of useful
order is not possible.

Also maintenance of visibility is not done

Abstraction of Operators

Operators of type PICKUP is divided into three parts:

 POSITION for positioning the robot’s hand.

 CATCH for grasping it.
 LIFT to lift it.
Pseudocode

input: initial non-linear plan output: non-linear plan

begin

while nodes unsolved or unexpanded in P do

select unsolved or unexpanded node N from P

arbitrary choice → backtrack point

if N unsolved

then

select action A, which has N

in add list ;;; arbitrary choice → backtrack point

insert A as operator node immediately

before N in P add achieve-nodes for the preconditions of A

in parallel immediately before A

else

replace N by its plot and supplement preconditions

end if

resolve interactions

87
 criticise plan

end do

return(P)

end

2. Explain the components of a planning system with suitable examples.

(November 2012)

Planning refers to the use of methods that focus on ways of decomposing the original
problem into appropriate subparts and on ways of recording and handling interaction among
the subparts as they are detected during the problem solving process.

Components Of Planning System

The important components of planning are

1. Choosing the best rule to apply next based on the best variable available heuristic
information.
2. Applying the chosen rule to compute the new problem state that arises from its
application.
3. Detecting when a solution has been found.
4. Detecting dead ends so that they can be abandoned and the system’s effort directed in
correct direction.
5. Repairing an almost correct solution.
1. Choosing rules to apply:

 First isolate a set of differences between the desired goal state and the current state.
 Detect rules that are relevant to reduce the differences.
If several rules are found, a variety of heuristic information can be exploited to choose
among them.

This technique is based on the means end analysis method.

2. Applying rules:

Applying the rules is easy.

Each rule specifies the problem state that would result from its application.

We must be able to deal with rules that specify only a small part of the complete problem
state.

Different ways to do this are

1. Describe, for each action, each of the changes it makes the state description.

88
A state was described by a set of predicates representing the facts that were true in that state.
Each state is represented as a predicate.

The manipulation of the state description is done using a resolution theorem prover.

3. Detecting 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.

Any of the corresponding reasoning mechanisms could be used to discover when a solution
has been found.

4. Detecting dead ends

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 mechanisms that can be used to detect a solution can often be used for
detecting a dead end.

If the search process is reasoning forward 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 initial state cannot be reached or little progress is being
made.

In reasoning backward, each goal is decomposed into sub goals. Each of them may
lead to a set of additional sub goals. Sometimes it is easy to detect that there is now way that
the entire sub goals in a given set can be satisfied at once. Other paths can be pruned because
they lead nowhere.

5.Repairing an almost correct solution

Solve the sub problems separately and then combine the solution to yield a correct
solution. But it leads to wasted effort.

The other way is to look at the situations that result when the sequence of operations
corresponding to the proposed solution is executed and to compare that situation to the
desired goal. The difference between the initial state and goal state is small. Now the problem
solving can be called again and asked to find a way of eliminating a new difference. The first
solution can then be combined with second one to form a solution to the original problem.

When information as possible is available, complete the specification in such a way

that no conflicts arise. This approach is called least commitment strategy.

89
 It can be applied in a variety of ways.

 To defer deciding on the order in which operations can be performed.

 Choose one order in which to satisfy a set of preconditions, we could leave the
order unspecified until the very end. Then we could look at the effects of each of
the sub solutions to determine the dependencies that exist among them. At that
point, an ordering can be chosen.

3. Explain in detail about the various machine learning methods.

(June 2013)

The various types of learning are

1. Role learning
2. Learning by taking advice.
3. Learning by parameter adjustment.
4. Learning with macro-operators.
5. Chunking
6. Explanation based learning
7. Clustering
8. Correcting mistakes
9. Recording cases.
10. Managing multiple models.
11. Back propagation.
12. Reinforcement learning.
13. Genetic algorithms
1. Role learning

Rote learning is the basic learning activity. It is also called memorization because
the knowledge, without any modification is, simply copied into the knowledge base. As
computed values are stored, this technique can save a significant amount of time.
Rote learning technique can also be used in complex learning systems provided
sophisticated techniques are employed to use the stored values faster and there is a
generalization to keep the number of stored information down to a manageable level.
Checkers-playing program, for example, uses this technique to learn the board positions it
evaluates in its look-ahead search.
Capabilities are

1. Organized storage of information.

2. Generalization
2. Learning by taking advice
This is a simple form of learning. Suppose a programmer writes a set of instructions
to instruct the computer what to do, the programmer is a teacher and the computer is a
student. Once learned (i.e. programmed), the system will be in a position to do new things.

90
 The advice may come from many sources: human experts, internet to name a few.
This type of learning requires more inference than rote learning. The knowledge must be
transformed into an operational form before stored in the knowledge base. Moreover the
reliability of the source of knowledge should be considered.
The system should ensure that the new knowledge is conflicting with the existing knowledge.
FOO (First Operational Operationaliser), for example, is a learning system which is used to
learn the game of Hearts. It converts the advice which is in the form of principles, problems,
and methods into effective executable (LISP) procedures (or knowledge). Now this
knowledge is ready to use.
3. Learning by parameter adjustment
Here the learning system relies on evaluation procedure that combines information
from several sources into a single summary static. For example, the factors such as demand
and production capacity may be combined into a single score to indicate the chance for
increase of production. But it is difficult to know a priori how much weight should be
attached to each factor.
The correct weight can be found by taking some estimate of the correct settings and then
allow the program modify its settings based on its experience. This type of learning systems
is useful when little knowledge is available. In game programs, for example, the factors such
as piece advantage and mobility are combined into a single score to decide whether a
particular board position is desirable. This single score is nothing but a knowledge which the
program gathered by means of calculation.
4. Learning with Macro-operators
Sequence of actions that can be treated as a whole are called macro-operators. Once a
problem is solved, the learning component takes the computed plan and stores it as a macro-
operator. The preconditions are the initial conditions of the problem just solved, and its post
conditions correspond to the goal just achieved.
The problem solver efficiently uses the knowledge base it gained from its previous
experiences. By generalizing macro-operators the problem solver can even solve different
problems. Generalization is done by replacing all the constants in the macro-operators with
variables.The STRIPS, for example, is a planning algorithm that employed macro-operators
in it’s learning phase. It builds a macro operator MACROP, which contains preconditions,
post-conditions and the sequence of actions. The macro operator will be used in the future
operation.
5. Chunking
Chunking is similar to learning with macro-operators. Generally, it is used by
problem solver systems that make use of production systems.
A production system consists of a set of rules that are in if-then form. That is given a
particular situation, what are the actions to be performed. For example, if it is raining then
take umbrella.
Production system also contains knowledge base, control strategy and a rule applier. To solve
a problem, a system will compare the present situation with the left hand side of the rules. If
there is a match then the system will perform the actions described in the right hand side of
the corresponding rule.

91
Problem solvers solve problems by applying the rules. Some of these rules may be more
useful than others and the results are stored as a chunk. Chunking can be used to learn general
search control knowledge. Several chunks may encode a single macro-operator and one
chunk may participate in a number of macro sequences. Chunks learned in the beginning of
problem solving, may be used in the later stage. The system keeps the chunk to use it in
solving other problems.
Soar is a general cognitive architecture for developing intelligent systems. Soar
requires knowledge to solve various problems. It acquires knowledge using chunking
mechanism. The system learns reflexively when impasses have been resolved. An impasse
arises when the system does not have sufficient knowledge. Consequently, Soar chooses a
new problem space (set of states and the operators that manipulate the states) in a bid to
resolve the impasse. While resolving the impasse, the individual steps of the task plan are
grouped into larger steps known as chunks. The chunks decrease the problem space search
and so increase the efficiency of performing the task.
In Soar, the knowledge is stored in long-term memory. Soar uses the chunking mechanism to
create productions that are stored in long-term memory. A chunk is nothing but a large
production that does the work of an entire sequence of smaller ones. The productions have a
set of conditions or patterns to be matched to working memory which consists of current
goals, problem spaces, states and operators and a set of actions to perform when the
production fires. Chunks are generalized before storing. When the same impasse occurs
again, the chunks so collected can be used to resolve it.
6. Explanation based learning
An Explanation-based Learning (EBL) system accepts an example (i.e. a training
example) and explains what it learns from the example. The EBL system takes only the
relevant aspects of the training. This explanation is translated into particular form that a
problem solving program can understand. The explanation is generalized so that it can be
used to solve other problems.
PRODIGY is a system that integrates problem solving, planning, and learning methods in a
single architecture. It was originally conceived by Jaime Carbonell and Steven Minton, as an
AI system to test and develop ideas on the role that machine learning plays in planning and
problem solving. PRODIGY uses the EBL to acquire control rules.

The EBL module uses the results from the problem-solving trace (ie. Steps in solving
problems) that were generated by the central problem solver (a search engine that searches
over a problem space). It constructs explanations using an axiomatized theory that describes
both the domain and the architecture of the problem solver. The results are then translated as
control rules and added to the knowledge base. The control knowledge that contains control
rules is used to guide the search process effectively.

7. Clustering
Discovery is a restricted form of learning. The knowledge acquisition is done without
getting any assistance from a teacher. Discovery Learning is an inquiry-based learning
method.

92
In discovery learning, the learner uses his own experience and prior knowledge to discover
the truths that are to be learned. The learner constructs his own knowledge by experimenting
with a domain, and inferring rules from the results of these experiments. In addition to
domain information the learner need some support in choosing and interpreting the
information to build his knowledge base.
A cluster is a collection of objects which are similar in some way. Clustering groups data
items into similarity classes. The properties of these classes can then be used to understand
problem characteristics or to find similar groups of data items. Clustering can be defined as
the process of reducing a large set of unlabeled data to manageable piles consisting of similar
items. The similarity measures depend on the assumptions and desired usage one brings to
the data.
Clustering begins by doing feature extraction on data items and measure the values of the
chosen feature set. Then the clustering model selects and compares two sets of data items and
outputs the similarity measure between them. Clustering algorithms that use particular
similarity measures as subroutines are employed to produce clusters.
The clustering algorithms are generally classified as Exclusive Clustering, Overlapping
Clustering, Hierarchical Clustering and Probabilistic Clustering. The selection of clustering
algorithms depends on various criteria such as time and space complexity. The results are
checked to see if they meet the standard otherwise some or all of the above steps have to be
repeated.
Some of the applications of clustering are data compression, hypothesis generation and
hypothesis testing. The conceptual clustering system accepts a set of object descriptions in
the form of events, observations, facts and then produces a classification scheme over the
observations.
COBWEB is an incremental conceptual clustering system. It incrementally adds the objects
into a classification tree. The attractive feature of incremental systems is that the knowledge
is updated with each new observation. In COBWEB system, learning is incremental and the
knowledge it learned in the form of classification trees increase the inference abilities.
8. Correcting mistakes
While learning new things, there is a possibility that the learning system may make
mistakes. Like human beings, learning system can correct itself by identifying reasons for its
failure, isolate it, explain how the particular assumption causes failure, and modifies its
knowledge base. For example, while playing chess a learning system may make a wrong
move and ends up with failure. Now the learning system thinks of the reasons for the failure
and corrects its knowledge base. Therefore when it plays again it will not repeat the same
mistake.

In his work, Active Learning with Multiple Views, Ion Muslea has used this technique to
label the data. He develops a technique known as Co-EMT which is a combination of two
techniques: Co-testing and Co-EM. The Co-testing method interacts with the user to label the
data. If it does any mistake in labeling, it learns from the mistakes and improves. After
learning, the system labels the unlabelled data extracted from a source efficiently. The
labeled data constitutes what is called knowledge.

93
9. Recording cases.
A program that learns by recording cases generally makes use of consistency
heuristic. According to consistency heuristic, a property of something can be guessed by
finding the most similar cases from a given set of cases. For example, a computer is given the
images of different types of insects, birds, and animals. If the computer is asked to identify a
living thing which is not in the recorded list, it will compare the given image with already
recorded ones, and will at least tell whether the given image is insect, bird or animal.Learning
by recoding cases technique is mainly used in natural language learning tasks.

During the training phase, a set of cases that describe ambiguity resolution episodes for a
particular problem in text analysis is collected. Each case contains a set of features or
attribute-value pairs that encode the context in which the ambiguity was encountered.

Moreover, each case is annotated with solution features that explain how the ambiguity was
resolved in the current example. The cases which are created are then stored in a case base.
Once the training is over, the system can use the case base to resolve ambiguities in new
sentences. This way, the system acquires the linguistic knowledge.

10. Managing multiple models.

Learning can be done by managing a set of general models and a set of specific
models kept in a version space. A version space is nothing but a representation that is used to
get relevant information from a set of learning examples.

A version space description consists of two trees which are complement to each other: one
represents general model and the other represents specific model. Both positive and negative
examples are used to get the two set of models converge on one just-right model. That is,
positive examples generalize specific models and negative examples specialize general
models. Ultimately, a correct model that matches only the observed positive examples is
obtained. Query By Committee (QBC) is an algorithm which implements this technique in
order to acquire knowledge.

11. Back propagation.

Backpropagation is a kind of neural network. A Neural Network (or artificial neural
network) is a collection of interconnected processing elements or nodes. The nodes are
termed simulated neurons as they attempt to imitate the functions of biological neurons. The
nodes are connected together via links. We can compare this with axon-synapse-
dendrite connections in the human brain.

HOW BACKPROPAGATION WORKS?

Initially, a weight is assigned at random to each link in order to determine the strength
of one node’s influence on the other. When the sum of input values reaches a threshold value,
the node will produce the output 1 or 0 otherwise. By adjusting theweights the desired output
can be obtained. This training process makes the network learn. The network, in other words,
acquires knowledge in much the same way human brains acquire namely learning from

94
experience. Backpropagation is one of the powerful artificial neural network technique which
is used acquire knowledge automatically.

Backpropagation method is the basis for training a supervised neural network. The output is a
real value which lies between 0 and 1 based on the sigmoid function. The formula for the
output is,
Output = 1 / (1+e-sum).

As the sum increases, the output approaches 1. As the sum decreases, the output approaches
0.

A Multilayer Network

A multilayer network is a kind of neural network which consists of one or more layers

of nodes between the input and the output nodes. The input nodes pass values to the hidden
layer, which in turn passes to the output layer. A network with a layer of input units, a layer
of hidden units and a layer of output units is a two-layer network. A network with two layers
of hidden units is a three-layer network, and so on. The multilayer network is the basis
for back propagation network.

As the name implies, there is a backward pass of error to each internal node within the
network, which is then used to calculate weight gradients for that node. The network learns
by alternately propagating forward the activations and propagating backward the errors as
and when they occur.

Back propagation network can deal with various types of data. It also has the ability to model
a complex decision system. If the problem domain involves large amount of data, the network
will require having more input units or hidden units. Consequently this will increase the
complexity of the model and also increase its computational complexity. Moreover it takes
more time in solving complex problems. In order to overcome this problem multi-back
propagation network is used.

In the beginning of the training process, weights are assigned to the connections at random.
The training process is iterative. The training cases are given to the network iteratively. The

95
actual outputs are compared with desired outputs and the errors are calculated. The errors are
then propagated back to the network in order to adjust the weights. The training process
repeats till the desired outputs are obtained or the error reaches an acceptable level.

TYPES OF BACKPROPAGATION NETWORKS

Static back-propagation is one kind of back propagation networks that produces a

mapping of a static input to a static output. These networks can solve static classification
problems such as optical character recognition (OCR). Recurrent Back propagation is another
kind of type used for fixed-point learning. NeuroSolutions, for example, is software that has
this ability. In recurrent bac kpropagation, activations are fed forward until a fixed value is
achieved. There after the error is computed and propagated backwards. The difference
between static and recurrent backpropagation is that the mapping is instantaneous in static
back-propagation while it is not in the case of latter type. Moreover training a network using
fixed-point learning is more difficult than with static backpropagation.

APPLICATION OF BACKPROPAGATION

The researchers, Thomas Riga, Angelo Cangelosi (University of Plymouth) and

Alberto Greco (University of Genoa), have developed a model for grounding of symbol. It
assumes the brain as a symbol system and explains cognition as a manipulation of symbols
governed by rules.

The symbol grounding problem is nothing but connecting meaning to the symbols or images
of objects received from input stimuli. The model uses two learning algorithms: Kohonen
Self-Organizing Feature Map and backpropagation algorithm.

To perform the tasks, it has two modules and a retina for input. The first module
uses Kohonen Self-Organizing Feature Map and categorizes the images projected on the
retina. It expresses the intrinsic order of the stimulus set in a bi-dimensional matrix known as
activation matrix.

The second module relates visual input, symbolic input and output stimuli. Now it
uses backpropagation algorithm for learning. The error in the output is computed with respect
to the training set and is sent back to input unit and the weight distribution is corrected in
order to get the correct output. In this process, the symbols which are grounded constitute the
knowledge. The knowledge so acquired is then used to generate and describe new meanings
for the symbols.

12. Reinforcement learning.

Reinforcement learning is one of the most active research areas in Artificial
Intelligence. Reinforcement learning is training by rewards and punishments. Here we train a
computer as if we train a dog. If the dog obeys and acts according to our instructions we
encourage it by giving biscuits or we punish it by beating or by scolding. Similarly, if the
system works well then the teacher gives positive value (i.e. reward) or the teacher gives

96
negative value (i.e. punishment). The learning system which gets the punishment has to
improve itself. Thus it is a trial and error process. The reinforcement learning
algorithms selectively retain the outputs that maximize the received reward over time. To
accumulate a lot of rewards, the learning system must prefer the best experienced actions;
however, it has to try new actions in order to discover better action selections for the future.

TEMPORAL DIFFERENCE LEARNING

Temporal difference learning is a central idea to reinforcement learning. It is based

on Monte Carlo methods and dynamic programming. It is an unsupervised technique.
Temporal difference learning methods can learn directly from raw experience without a
model of the environment’s dynamics. Examples are learning to play games, robot control,
elevator control, network routing and animal learning.

APPLICATIONS OF REINFORCEMENT LEARNING

Personalization Travel Support System is one software that is designed to provide

travelling information as per the user’s interests. It applies the reinforcement learning to
analyze and learn customer behaviors and list out the products that the customers wish to buy.
If the system selects the right item that the customer wish to buy then it is given reward by
assigning a particular value for the state that a user selects to perform and if the system
selects an item which the user does not wish to buy then it is given the penalty. This way the
system learns the personal interests. In this process, the system acquires the knowledge of the
user behavior and interest which makes it decide which information should be given to a
particular user. This result in greater customer satisfaction and increase in the success rate of
product promotion.

13. Genetic algorithms

Genetic algorithms are based on biological evolution. Genetic algorithms can be used
to solve a wide variety of problems. Given a problem a genetic algorithm generates a set of
possible solutions and evaluates each in order to decide which solutions are fit for
reproduction. If a particular solution is more fit then it will have more chances to generate
new solutions. Finally we can find a real solution.

Genetic algorithms are so powerful that they can exhibit more efficiency if programmed
perfectly. Applications include learning Robot behavior, molecular structure optimization,
automated design of mechatronic systems, and electronic circuit design.

A Genetic Approach Methodology for Knowledge Acquisition for Intelligent

Diagnosis is a project done by CEEP (Center for Engineering Education and Practice),
University of Michigan-Dearborn, to develop diagnosis knowledge for robot arm movements.
In this project, a robot was selected for analysis. The wrong arm movements of robot were
observed and collected as failure data. Then the project applied genetic algorithm to extract
knowledge from the failure data. The acquisition of knowledge was automatic.
The knowledge extracted is then stored in the knowledge base to make use of in the

97
intelligent diagnosis system. The knowledge so stored is known as diagnosis knowledge as it
s used to detect what went wrong and to decide the course of action in order to make the
robot perfect.

4. Explain adaptive learning with an example. (November 2013)

Adaptive learning

Adaptive learning is an educational method which uses computers as interactive

teaching devices, and to orchestrate the allocation of human and mediated resources
according to the unique needs of each learner. Computers adapt the presentation of
educational material according to students' learning needs, as indicated by their responses to
questions, tasks and experiences. The technology encompasses aspects derived from various
fields of study including computer science, education, psychology, and brain science.

Adaptive learning has been partially driven by a realization that tailored learning
cannot be achieved on large-scale using traditional, non-adaptive approaches. Adaptive
learning systems endeavor to transform the learner from passive receptor of information to
collaborator in the educational process. Adaptive learning systems' primary application is in
education, but another popular application is business training. They have been designed as
desktop computer applications, web applications, and are now being introduced into overall
curricula.

Adaptive learning has been implemented in several kinds of educational systems such
as adaptive educational hypermedia, intelligent tutoring systems, Computerized adaptive
testing, and computer-based pedagogical agents, among others.

Technology and methodology

Adaptive learning systems have traditionally been divided into separate components
or 'models'. While different model groups have been presented, most systems include some
or all of the following models (occasionally with different names):
 Expert model - The model with the information which is to be taught
 Student model - The model which tracks and learns about the student
 Instructional model - The model which actually conveys the information
 Instructional environment - The user interface for interacting with the system
Expert model
The expert model stores information about the material which is being taught. This
can be as simple as the solutions for the question set but it can also include lessons and
tutorials and, in more sophisticated systems, even expert methodologies to illustrate
approaches to the questions.
Adaptive learning systems which do not include an expert model will typically incorporate
these functions in the instructional model.

98
Student model
Student model algorithms have been a rich research area over the past twenty years.
The simplest means of determining a student's skill level is the method employed in CAT
(Computerized adaptive testing). In CAT, the subject is presented with questions that are
selected based on their level of difficulty in relation to the presumed skill level of the subject.
As the test proceeds, the computer adjusts the subject's score based on their answers,
continuously fine-tuning the score by selecting questions from a narrower range of difficulty.
An algorithm for a CAT-style assessment is simple to implement. A large pool of
questions is amassed and rated according to difficulty, through expert analysis,
experimentation, or a combination of the two. The computer then performs what is essentially
a binary search, always giving the subject a question which is half way between what the
computer has already determined to be the subject's maximum and minimum possible skill
levels. These levels are then adjusted to the level of the difficulty of the question, reassigning
the minimum if the subject answered correctly, and the maximum if the subject answered
incorrectly. Obviously, a certain margin for error has to be built in to allow for scenarios
where the subject's answer is not indicative of their true skill level but simply coincidental.
Asking multiple questions from one level of difficulty greatly reduces the probability of a
misleading answer, and allowing the range to grow beyond the assumed skill level can
compensate for possible misevaluations.
Richer student model algorithms look to determine causality and provide a more
extensive diagnosis of student's weaknesses by linking 'concepts' to questions and defining
strengths and weaknesses in terms of concepts rather than simple 'levels' of ability. Because
multiple concepts can influence a single question, questions have to be linked to all relevant
concepts. For example, a matrix can list binary values (or even scores) for the intersection of
every concept and every question. Then, conditional probability values have to be calculated
to reflect the likelihood that a student who is weak in a particular concept will fail to correctly
answer a particular question. A student takes a test, the probabilities of weakness in all
concepts conditional on incorrect answers in all questions can be calculated using Baye’s
Law (these adaptive learning methods are often called bayesian algorithms).
A further extension of identifying weaknesses in terms of concepts is to program the student
model to analyze incorrect answers. This is especially applicable for multiple choice
questions. Consider the following example:
Q. Simplify: 

a) Can't be simplified

b) 

c) ...

d) ...

Clearly, a student who answers (b) is adding the exponents and failing to grasp the concept of
like terms. In this case, the incorrect answer provides additional insight beyond the simple
fact that it is incorrect.

99
Instructional model
The instructional model generally looks to incorporate the best educational tools that
technology has to offer (such as multimedia presentations) with expert teacher advice for
presentation methods. The level of sophistication of the instructional model depends greatly
on the level of sophistication of the student model. In a CAT-style student model, the
instructional model will simply rank lessons in correspondence with the ranks for the
question pool. When the student's level has been satisfactorily determined, the instructional
model provides the appropriate lesson. The more advanced student models which assess
based on concepts need an instructional model which organizes its lessons by concept as
well. The instructional model can be designed to analyze the collection of weaknesses and
tailor a lesson plan accordingly.
When the incorrect answers are being evaluated by the student model, some systems
look to provide feedback to the actual questions in the form of 'hints'. As the student makes
mistakes, useful suggestions pop up such as "look carefully at the sign of the number". This
too can fall in the domain of the instructional model, with generic concept-based hints being
offered based on concept weaknesses, or the hints can be question-specific in which case the
student, instructional, and expert models all overlap.
Implementation
Classroom Implementation
Adaptive learning that is implemented in the classroom environment using
information technology is often referred to as an Intelligent Tutoring System or an Adaptive
Learning System. Intelligent Tutoring Systems operate on three basic principles:
Systems need to be able to dynamically adapt to the skills and abilities of a student.

Environments utilize cognitive modeling to provide feedback to the student while assessing
student abilities and adapting the curriculum based upon past student performance.

Inductive logic programming (ILP) is a way to bring together inductive learning and logic
programming to an Adaptive Learning System. Systems using ILP are able to create
hypothesis from examples demonstrated to it by the programmer or educator and then use
those experiences to develop new knowledge to guide the student down paths to correct
answers.

Systems must have the ability to be flexible and allow for easy addition of new content.

Cost of developing new Adaptive Learning Systems is often prohibitive to educational

institutions so re-usability is essential.

School districts have specific curriculum that the system needs to utilize to be effective for
the district. Algorithms and cognitive models should be broad enough to teach mathematics,
science, and language.

Systems need to also adapt to the skill level of the educators.

100
Many educators and domain experts are not skilled in programming or simply do not have
enough time to demonstrate complex examples to the system so it should adapt to the abilities
of educators.

Distance learning implementation

Adaptive Learning systems can be implemented on the Internet for use in distance
learning and group collaboration applications.
The field of distance learning is now incorporating aspects of adaptive learning. Initial
systems without adaptive learning were able to provide automated feedback to students who
are presented questions from a preselected question bank. That approach however lacks the
guidance which teachers in the classroom can provide. Current trends in distance learning call
for the use of adaptive learning to implement intelligent dynamic behavior in the learning
environment.
During the time a student spends learning a new concept they are tested on their
abilities and databases track their progress using one of the models. The latest generation of
distance learning systems take into account the students' answers and adapt themselves to the
student's cognitive abilities using a concept called 'cognitive scaffolding'. Cognitive
scaffolding is the ability of an automated learning system to create a cognitive path of
assessment from lowest to highest based on the demonstrated cognitive abilities. A current
successful implementation of adaptive learning in web-based distance learning is the Maple
engine of WebLearn by RMIT university. WebLearn is advanced enough that it can provide
assessment of questions posed to students even if those questions have no unique answer like
those in the Mathematics field.
Group collaboration is also a hot field in the adaptive learning research area. Group
collaboration is a key field in Web 2.0 which extends the functionality of distance learning.
Adaptive learning can be incorporated to facilitate collaboration within distance learning
environments like forums or resource sharing services. Some examples of how adaptive
learning can help with collaboration include:
 Automated grouping of users with the same interests.
 Personalization of links to information sources based on the user's
stated interests or the user's surfing habits.
Game Design Implementation
In 2014, an educational researcher concluded a multi-year study of adaptive learning
for educational game design. The research developed and validated the ALGAE (Adaptive
Learning GAme dEsign) model, a comprehensive adaptive learning model based on game
design theories and practices, instructional strategies, and adaptive models. The research
extended previous researching in game design, instructional strategies, and adaptive learning,
combining those three components into a single complex model.
The study resulted in the development of an adaptive educational game design model
to serve as a guide for game designers, instructional designers, and educators with the goal of
increasing learning outcomes. Survey participants validated the value of the ALGAE model
and provided specific insights on the model's construction, use, benefits, and challenges. The
current ALGAE model is based on these insights. The model now serves as a guideline for
the design and development of educational computer games.

101
5. Solve the following by goal stack planning. (June 2013)

B C B

A C D Goal A D

Goal stack planning

 One of the earliest techniques is planning using goal stack.

 Problem solver uses single stack that contains
o sub goals and operators both
o sub goals are solved linearly and then finally the conjoined sub goal is solved.
 Plans generated by this method will contain
o complete sequence of operations for solving one goal followed by complete
sequence of operations for the next etc.
 Problem solver also relies on
o A database that describes the current situation.
o Set of operators with precondition, add and delete lists.
Algorithm

 Let us assume that the goal to be satisfied is:

GOAL = G1  G2  … Gn

 Sub-goals G1, G2, … Gn are stacked with compound goal G1  G2  … Gn at the

bottom.
 Top G1
G2

Gn

 Bottom G1  G2  …  G4
 At each step of problem solving process, the top goal on the stack is pursued.
 Find an operator that satisfies sub goal G1 (makes it true) and replace G1 by the
operator.
o If more than one operator satisfies the sub goal then apply some heuristic to
choose one.
 In order to execute the top most operation, its preconditions are added onto the stack.
o Once preconditions of an operator are satisfied, then we are guaranteed that
operator can be applied to produce a new state.
o New state is obtained by using ADD and DELETE lists of an operator to the
existing database.
102
  Problem solver keeps track of operators applied.
o This process is continued till the goal stack is empty and problem solver
returns the plan of the problem.

Example

 Logical representation of Initial and Goal states:

o Initial State: ON(B, A)  ONT(C)  ONT(A)  ONT(D)  CL(B) CL(C) 
CL(D)  AE
o Goal State: ON(C, A)  ON(B, D)  ONT(A)  ONT(D)  CL(C)  CL(B)
 AE

 Following sub-goals in goal state are also true in initial state.

 ONT(A)  ONT(D)  CL(C)  CL(B)  AE
 Represent for the sake of simplicity - TSUBG.
 Only sub-goals ON(C, A) & ON(B, D) are to be satisfied and finally make sure that
TSUBG remains true.
 Either start solving first ON(C, A) or ON(B, D). Let us solve first ON(C, A).
Goal Stack:

ON(C, A)

ON(B, D)

ON(C, A)  ON(B, D)  TSUBG

 To solve ON(C, A), operation S(C, A) could only be applied.

 So replace ON(C, A) with S(C, A) in goal stack.
Goal Stack:

 S (C, A)
 ON(B, D)
 ON(C, A)  ON(B, D)  TSUBG
 S(C, A) can be applied if its preconditions are true. So add its preconditions on the
stack.
Goal Stack:

CL(A)

103
HOLD(C) Preconditions of STACK

CL(A)  HOLD(C)

S (C, A) Operator

ON(B, D)

ON(C, A)  ON(B, D)  TSUBG

 Next check if CL(A) is true in State_0.

 Since it is not true in State_0, only operator that could make it true is US(B, A).
 So replace CL(A) with US(B, A) and add its preconditions.
Goal Stack: ON(B, A)

CL(B) Preconditions of UNSTACK

AE

ON(B, A)  CL(B)  AE

US(B, A) Operator

HOLD(C)

CL(A) )  HOLD(C)

S (C, A) Operator

ON(B, D)

ON(C, A)  ON(B, D)  TSUBG


ON(B, A), CL(B) and AE are all true in initial state, so pop these along with its
compound goal.
 Next pop top operator US(B, A) and produce new state by using its ADD and
DELETE lists.
 Add US(B, A) in a queue of sequence ofoperators.
SQUEUE = US (B, A)

State_1:

ONT(A) ONT(C)  ONT(D)  HOLD(B) CL(A)  CL(C) CL(D)

Goal Stack:

HOLD(C)

CL(A) )  HOLD(C)

S (C, A) Operator
104
 ON(B, D)

ON(C, A)  ON(B, D)  TSUBG

 To satisfy the goal HOLD(C), two operators can be used e.g., PU(C ) or US(C,
X), where X could be any block. Let us choose PU(C ) and proceed further.
 Repeat the process. Change in states is shown below.
State_1:

ONT(A) ONT(C)  ONT(D)  HOLD(B) CL(A)  CL(C) CL(D)

SQUEUE = US (B, A)

 Next operator to be popped of is S(B, D). So

State_2:

ONT(A) ONT(C)  ONT(D)  ON(B, D) CL(A)  CL(C) CL(B)AE

SQUEUE = US (B, A), S(B, D)

State_3:

ONT(A) HOLD(C)  ONT(D)  ON(B, D) CL(A) CL(B)

SQUEUE = US (B, A), S(B, D), PU(C )

State_4:

ONT(A) ON(C, A)  ONT(D)  ON(B, D) CL(C) CL(B) AE

SQUEUE = US (B, A), S(B, D), PU(C ), S(C, A)

 The Goal stack method is not efficient for difficult problems such as Sussman
anomaly problem.
 It fails to find good solution.
 Consider the Sussman anomaly problem

Initial State: ON(C, A)  ONT(A)  ONT(B)

Goal State: ON(A, B)  ON(B, C)

 Remove CL and AE predicates for the sake of simplicity.

 To satisfy ON(A, B), following operators are applied

105
  US(C, A) , PD(C), PU(A) and S(A, B)

State_1: ON(B, A)  ONT(C)

 To satisfy ON(B, C), following operators are applied

US(A, B) , PD(A), PU(B) and S(B, C)

State_2: ON(B, C)  ONT(A)

 Finally satisfy combined goal ON(A, B)  ON(B, C).

 Combined goal fails as while satisfying ON(B, C), we have undone ON(A, B).
 Difference in goal and current state is ON(A, B).
 Operations required are PU(A) and S(A, B)

 The complete plan for solution is as follows:

1. US(C, A)

2. PD (C)

106
 3. PU(A)

4. S(A, B)

5. US(A, B)

6. PD(A)

7. PU(B)

8. S(B, C)

9. PU(A)

10. S(A, B)

 Although this plan will achieve the desired goal, but it is not efficient.
 In order to get efficient plan, either repair this plan or use some other method.
 Repairing is done by looking at places where operations are done and undone
immediately, such as S(A, B) and US(A, B).
 By removing them, we get
1. US(C, A)

2. PD (C)

3. PU(B)

4. S(B, C)

5. PU(A)

6. S(A, B)

6. Solve the blocks world problem using strips. How does it act as a planning system?
( May 2014)
Block World Problem

 Block world problem assumptions

o Square blocks of same size
o Blocks can be stacked one upon another.
o Flat surface (table) on which blocks can be placed.
o Robot arm that can manipulate the blocks. It can hold only one block at a time.
 In block world problem, the state is described by a set of predicates representing the
facts that were true in that state.
 One must describe for every action, each of the changes it makes to the state
description.
 In addition, some statements that everything else remains unchanged is also
necessary.
Actions (Operations) done by Robot

 UNSTACK (X, Y) : [US (X, Y)]

107
 o Pick up X from its current position on block Y. The arm must be empty and X
has no block on top of it.
 STACK (X, Y): [S (X, Y)]
o Place block X on block Y. Arm must holding X and the top of Y is clear.
 PICKUP (X): [PU (X) ]
o Pick up X from the table and hold it. Initially the arm must be empty and top
of X is clear.
 PUTDOWN (X): [PD (X)]
o Put block X down on the table. The arm must have been holding block X.
 Predicates used to describe the state
o ON(X, Y) - Block X on block Y.
o ONT(X) - Block X on the table.
o CL(X) - Top of X clear.
o HOLD(X) - Robot-Arm holding X.
o AE - Robot-arm empty.
 Logical statements true in this block world.
o Holding X means, arm is not empty
 ( X) HOLD (X)  ~ AE
o X is on a table means that X is not on the top of any block
 ( X) ONT (X)  ~ ( Y) ON (X, Y)
o Any block with no block on has clear top
 ( X) (~ ( Y) ON (Y,X))  CL (X)
 Effect of Unstack operation
 The effect of US(X, Y) is described by the following axiom
 [CL(X, State)  ON(X, Y, State)] 
 [HOLD(X, DO(US (X, Y), State))  CL(Y, DO(US(X, Y),
State)) ]
o DO is a function that generates a new state as a result of given action and a
state.
 For each operator, set of rules (called frame axioms) are defined where the
components of the state are
o affected by an operator
 If US(A, B) is executed in state S0, then we can infer that HOLD (A,
S1)  CLEAR (B, S1) holds true, where S1 is new state after Unstack
operation is executed.
o not affected by an operator
 If US(A, B) is executed in state S0, B in S1 is still on the table but we can’t derive it.
So frame rule stating this fact is defined as ONT(Z, S)  ONT(Z, DO(US (A, B), S))
 Advantage of this approach is that
o simple mechanism of resolution can perform all the operations that are
required on the state descriptions.
 Disadvantage is that
o number of axioms becomes very large for complex problem such as COLOR
of block also does not change.
o So we have to specify rule for each attribute.
 COLOR(X, red, S) 
 COLOR(X, red, DO(US(Y, Z), s))

108
  To handle complex problem domain, there is a need of mechanism that does not
require large number of explicit frame axioms.

7. Explain the steps in designing a learning system. (May 2014)

To solve problems computers require intelligence. Learning is central

to intelligence. As intelligence requires knowledge, it is necessary for the computers to
acquire knowledge. Machine learning serves this purpose.
Machine learning refers to a system capable of acquiring and integrating the knowledge
automatically. The capability of the systems to learn from experience, training, analytical
observation, and other means, results in a system that can continuously self-improve and
thereby exhibit efficiency and effectiveness.
A machine learning system usually starts with some knowledge and a
corresponding knowledge organization so that it can interpret, analyze, and test the
knowledge acquired.

The figure shown above is a typical learning system model. It consists of the following
components.
1. Learning element.
2. Knowledge base.
3. Performance element.
4. Feedback element.
5. Standard system.
1. Learning element
It receives and processes the input obtained from a person ( i.e. a teacher), from
reference material like magazines, journals, etc, or from the environment at large.
2. Knowledge base
This is somewhat similar to the database. Initially it may contain some basic
knowledge. Thereafter it receives more knowledge which may be new and so be
added as it is or it may replace the existing knowledge.
3. Performance element
It uses the updated knowledge base to perform some tasks or solves some problems
and produces the corresponding output.
4. Feedback element

109
 It is receiving the two inputs, one from learning element and one from standard (or
idealized) system. This is to identify the differences between the two inputs. The feedback is
used to determine what should be done in order to produce the correct output.
5. Standard system
It is a trained person or a computer program that is able to produce the correct
output. In order to check whether the machine learning system has learned well, the same
input is given to the standard system. The outputs of standard system and that of performance
element are given as inputs to the feedback element for the comparison. Standard system is
also called idealized system.
The sequence of operations described above may be repeated until the system gets the desired
perfection.
There are several factors affecting the performance.
They are,
 Types of training provided.
 The form and extent of any initial background knowledge.
 The type of feedback provided.
 The learning algorithms used.
Training is the process of making the system able to learn. It may consist of randomly
selected examples that include a variety of facts and details including irrelevant data. The
learning techniques can be characterized as a search through a space of possible hypotheses
or solutions. Background knowledge can be used to make learning more efficient by reducing
the search space. The feedback may be a simple yes or no type of evaluation or it may
contain useful information describing why a particular action was good or bad. If the
feedback is always reliable and carries useful information, the learning process will be faster
and the resultant knowledge will be correct.
The success of machine learning system also depends on the algorithms. These
algorithms control the search to find and build the knowledge structures. The algorithms
should extract useful information from training examples. 

110
 UNIT V EXPERT SYSTEMS

Expert systems - Architecture of expert systems, Roles of expert systems -

Knowledge Acquisition –Meta knowledge, Heuristics. Typical expert systems -
MYCIN, DART, XOON, Expert systems shells.

PART A
1. What are expert systems?
An expert system is an interactive computer-based decision tool that uses both facts
and heuristics to solve difficult decision making problems, based on knowledge acquired
from an expert.

An expert system is a model and associated procedure that exhibits, within a specific
domain, a degree of expertise in problem solving that is comparable to that of a human
expert.

An expert system compared with traditional computer:

Inference engine + Knowledge = Expert system

( Algorithm + Data structures = Program in traditional computer )

First expert system, called DENDRAL, was developed in the early 70's at Stanford
University.

2. What is the difference between conventional system and expert system?

Conventional System Expert System

Knowledge and processing are combined in Knowledge database and the processing
one programme. mechanism (inference) are two different
components.

Programme does not make errors (only The ES programme may be make a
mistake.
Programming error).

Usually it will not explain why the data Explanation is part of an ES component
needs to be input or how the decision is
achieved.

System is operational only when fully An ES can operate with small amount of
developed. rules.

111
Step by step execution according to fixed Execution done logically and heuristically.
algorithms is necessary.

Needs complete and full information. Can operate with sufficient or insufficient
information.

Manipulates a large and effective database. Manipulates a big and effective database.

Referencing and use of data. Referencing and use of Knowledge.

Main objective is efficiency. Main objective is effectiveness.

Easily operated with quantitative data. Easily operated with qualitative data

3. Define heuristics
Heuristic is a rule of thumb that probably leads to a solution. Heuristics play a major
role in search strategies because of exponential nature of the most problems. Heuristics help
to reduce the number of alternatives from an exponential number to a polynomial number.

4. Define Meta Knowledge and its use. (December 2011, November 2013)
 Meta knowledge includes the ability to evaluate the knowledge available, the
additional knowledge required and the systematic implied by the present rules.
Jacques Pitrat definition:

 Meta-knowledge is knowledge about knowledge, rather than knowledge from a

specific domain such as mathematics, medicine or geology.
 According to this definition, meta-knowledge is at the heart of the learning process,
which consists in transforming information into knowledge:
o By attributing values to knowledge from other domains: truth, usefulness,
importance, knowledge priority, competence of an individual towards a
knowledge object, etc.
o By describing « intellectual acts », processes that facilitate knowledge
processing in other domains: memorization, understanding, application,
analysis, synthesis, evaluation, etc.
o by representing strategies to acquire, process and use knowledge from other
domains : memorization techniques, heuristic principles for problem solving,
project management strategies, etc.
5. List the major uses of expert system.

The Expert System is used to:

 Help experts in their routine to improve productivity.

 Help experts in some of their more complex and difficult tasks so that the problem
can be managed effectively.
 Help an expert to obtain information needed by other experts who have forgotten
about it or who are too busy to search for it.
6. What is the use of expert system tools?
112
 Expert system tool is extremely expensive and powerful. The advantage of using this
tool is that it provides a variety in knowledge representation techniques such as rules and
frames.

7. Who is fit to be called an expert?

Anyone can be called an expert as long as that person has a vast knowledge of the
particular field and has practical experience in a certain domain. However, the person is
restricted to his or her own domain. For example, being an IT expert does not mean that the
person is an expert in all IT domains but she may be an expert in intelligence systems or an
expert in only the development of an intelligence agent.

8. Name the programming methods supported by expert system tools?

An ES can be developed using a symbolic language such as LISP or PROLOG, or a

conventional higher-level language such as FORTRAN, C and PASCAL.

LISP

All ES developed in the early days used LISP, or tools written using the LISP
language.

PROLOG

The on-going research of artificial intelligent has given birth to the programming
language PROLOG. PROLOG is the acronym for 'Programming in Logic'. A programme
using PROLOG can be assumed to be a knowledge database that stores facts and rules.

9. Name some expect systems developed for voice recognition and their uses.

Hearsay was an early attempt at solving voice recognition through an expert systems
approach. Hearsay and all interpretation systems are essentially pattern recognition systems
—looking for patterns in noisy data. In the case of Hearsay recognizing phonemes in an audio
stream. Other early examples were analyzing sonar data to detect Russian submarines. These
kinds of systems proved much more amenable to a neural network AI solution than a rule-
based approach.

10. Name some Expert systems used in Medicine

CADUCEUS and MYCIN were medical diagnosis systems. The user describes their

symptoms to the computer as they would to a doctor and the computer returns a medical
diagnosis.

11. How is Expert system used to study organic molecules?

Dendral was a tool to study hypothesis formation in the identification of organic

molecules. The general problem it solved—designing a solution given a set of constraints—

113
was one of the most successful areas for early expert systems applied to business domains
such as sales people configuring Dec Vax computers and mortgage loan application
development.

SMH.PAL is an expert system for the assessment of students with multiple disabilities.

12. How is expert system used to monitor dam safety and landslides?

Mistral  is an expert system for the monitoring of dam safety developed in the 90's by
Ismes (Italy). It gets data from an automatic monitoring system and performs a diagnosis of
the state of the dam. Its first copy, installed in 1992 on the Ridracoli Dam (Italy), is still
operational 24/7/365. It has been installed on several dams in Italy and abroad (e.g. Itaipu
Damin Brazil), as well as on landslides under the name of Eydenet, and on monuments under
the name of Kaleidos. Mistral is a registered trade mark of CESI.

13. Distinguish between Ignorance and uncertainty?

Ignorance does not know something that is knowable. Uncertainty is where something
is not knowable: it is inherent in the situation. Ignorance can come from (a) the limited
knowledge of human expert; (b) inexact data; or (c) incomplete data (which forces a
premature decision).

14. When is an Expert System Appropriate?

An expert system is appropriate.

 Need justifies cost and effort

 Human expertise not always available
 Problem requires symbolic reasoning
 Problem domain is well structured
 Traditional computing methods fail
 Cooperative and articulate experts exist
 Problem is not too large

15. Name any three universities and mention the expert system tools developed there?

1. Stanford University - MYCIN

2. University of Texas at Austin – TMYCIN - Tiny Mycin

3. National Taiwan University - EJAUNDICE

114
 PART B

1. Draw the schematic of expert system and explain. (December 2011, November
2012, June 2013)
Expert systems have a number of major system components and interface with
individuals who interact with the system in various roles.

I. Components and Interfaces

1. Knowledge base :

Expert system is built around a knowledge base module. Expert system

contains a formal representation of the information provided by the domain expert.
This information may be in the form of problem-solving rules, procedures, or data
intrinsic to the domain. To incorporate these information into the system, it is
necessary to make use of one or more knowledge representation methods.
Transferring knowledge from the human expert to a computer is often the most
difficult part of building an expert system.

The knowledge acquired from the human expert must be encoded in such a
way that it remains a faithful representation of what the expert knows, and it can be
manipulated by a computer.

Three common methods of knowledge representation evolved over the years are

1. IF-THEN rules.
2. Semantic networks.
3. Frames.

115
1. IF-THEN rules

Human experts usually tend to think along :

condition action or Situation conclusion

Rules "if-then" are predominant form of encoding knowledge in expert systems.

These are of the form :

If a1 , a2 , . . . . . , an Then b1 , b2 , . . . . . , bn where

each ai is a condition or situation, and each bi is an action or a conclusion.

2.Semantic Networks

In this scheme, knowledge is represented in terms of objects and relationships

between objects. The objects are denoted as nodes of a graph. The relationship between two
objects are denoted as a link between the corresponding two nodes. The most common form
of semantic networks uses the links between nodes to represent IS-A and HAS relationships
between objects.

Example of Semantic Network

The Fig. below shows a car IS-A vehicle; a vehicle HAS wheels. This kind of
relationship establishes an inheritance hierarchy in the network, with the objects lower down
in the network inheriting properties from the objects higher up.

116
3. Frames

In this technique, knowledge is decomposed into highly modular pieces called frames,
which are generalized record structures. Knowledge consist of concepts, situations, attributes
of concepts, relationships between concepts, and procedures to handle relationships as well as
attribute values.

Each concept may be represented as a separate frame.

The attributes, the relationships between concepts, and the procedures are allotted to slots in a
frame.

The contents of a slot may be of any data type - numbers, strings, functions or procedures
and so on.

The frames may be linked to other frames, providing the same kind of inheritance as that
provided by a semantic network.

A frame-based representation is ideally suited for objected-oriented programming techniques.

Example : Frame-based Representation of Knowledge.

Two frames, their slots and the slots filled with data type are shown.

117
2. Working storage

Working memory refers to task-specific data for a problem. The contents of the
working memory, changes with each problem situation. Consequently, it is the most dynamic
component of an expert system, assuming that it is kept current.

o Every problem in a domain has some unique data associated with it.
o Data may consist of the set of conditions leading to the problem, its parameters and so
on.
o Data specific to the problem needs to be input by the user at the time of using, means
consulting the expert system. The Working memory is related to user interface
Fig. below shows how Working memory is closely related to user interface of the expert
system.

3. Inference engine :

The code at the core of the system which derives recommendations from the
knowledge base and problem specific data in working storage;

The inference engine is a generic control mechanism for navigating through and
manipulating knowledge and deduce results in an organized manner. The inference engine's
generic control mechanism applies the axiomatic (self-evident) knowledge present in the
knowledge base to the task-specific data to arrive at some conclusion.

o Inference engine the other key component of all expert systems.

o Just a knowledge base alone is not of much use if there are no facilities for navigating
through and manipulating the knowledge to deduce
o something from knowledge base.A knowledge base is usually very large, it is
necessary to have inferencing mechanisms that search through the database and
deduce results in an organized manner.

118
The Forward chaining, Backward chaining and Tree searches are some of the techniques used
for drawing inferences from the knowledge base.

4. User interface :

The code that controls the dialog between the user and the system.

II. Roles of Individuals who interact with the system

1. Domain expert : The individuals who currently are experts in solving the
problems; here the system is intended to solve;

2. Knowledge engineer : The individual who encodes the expert's knowledge in a

declarative form that can be used by the expert system;

3. User : The individual who will be consulting with the system to get advice which
would have been provided by the expert.

III. Expert System Shells

Many expert systems are built with products called expert system shells. A shell is a
piece of software which contains the user interface, a format for declarative knowledge in the
knowledge base, and an inference engine. The knowledge and system engineers use these
shells in making expert systems.

1. Knowledge engineer: uses the shell to build a system for a particular problem
domain.

2. System engineer: builds the user interface, designs the declarative format of the
knowledge base, and implements the inference engine.

Depending on the size of the system, the knowledge engineer and the system engineer might
be the same person.

2. Write short notes on DART and MYCIN (May 2012, November 2012, June
2013, November 2013, May 2014)

i) DART

It is an artificial intelligence based expert system used for computer system fault
diagnosis. This system is an automated consultant that advises IBM field service personnel on
the diagnosis of faults occurring in computer installations.

The consultant identifies specific system components (both hardware and software) likely to
be responsible for an observed fault and offers a brief explanation of the major factors and
evidence supporting these indictments.
The consultant, called DART, was constructed using HMYCIN, and is part of a larger
research effort investigating automated diagnosis of machine faults.
Motivation and Scope of Effort

119
A typical, large-scale computer installation is composed of numerous subsystems including
CPUs, primary and secondary storage, peripherals, and supervisory software. Each of these
subsystems, in turn, consists of a richly connected set of both hardware and software
components such as disk drives, controllers, CPUs, memory modules, and access methods.
Generally, each individual component has an associated set of diagnostic aids designed to test
its own specific integrity. However, very few maintenance tools and established diagnostic
strategies aimed at identifying faults on the system or subsystem level. As a result,
identification of single or multiple faults from systemic manifestations remains a difficult
task. The non-specialist field service engineer is trained to use the existing component-
specific tools and, as a result, is often unable to attack the failure at the systemic level. Expert
assistance is then required, increasing both the time and cost required to determine and repair
the fault. The design of DART reflects the expert's ability to take a systemic viewpoint on
problems and to use that viewpoint to indict specific components, thus making more effective
use of the existing maintenance capabilities.
For our initial design, the concentration was on problems occurring within the
teleprocessing (TP) subsystems for the IBM 370-class computers. This subsystem includes
various network controllers, terminals, remote-job entry facilities, modems, and several
software access methods. In addition to these well-defined components there are numerous
available test points the program can use during diagnosis. The focus was on handling two of
the most frequent TP problems, (1) when a user is unable to log on to the system from a
remote terminal, and (2) when the system operator is unable to initialize the TP network
itself. In a new system configuration, these two problems constitute a significant percentage
of service calls received.

Interviews with field-service experts made it apparent that much of their problem-solving
expertise is derived from their knowledge of several well-defined communications protocols.
Often composed of simple request-acknowledge sequences, these protocols represent the
transactions between components that are required to perform various TP tasks. Although
based on information found in reference manuals it is significant that these protocols are not
explicitly detailed anywhere in the standard maintenance documentation. Knowledge of the
basic contents of these protocols and their common sequence forms the basis of a diagnostic
strategy: use the available tracing facilities to capture the actual communications occurring in
the network, and analyze this data to determine which link in the protocol chain has broken.
This procedure is sufficient to identify specific faulty components in the network.

The DART Consultation

During a DART consultation session, the field engineer focuses on a particular

computer system that is experiencing a problem. Many installations arc composed of
numerous CPUs that partially share peripherals, thus, the term "system" is taken to mean a
single CPU-complex and its attached peripherals. Within each such system, the user
describes one or more problems by indicating a failure symptom, currently using a list of
keywords. Using this description, the consultant makes an initial guess about which of the
major subsystems might be involved in the fault. The user is then given the opportunity to
select which of these implicated subsystems are to be pursued and in which order.

120
Each subsystem serves as a focal point for tests and findings associated with that segment of
the diagnostic activity. These subsystems currently correspond to various input/output
facilities (e.g.. DISK. TAPE. TP) or the CPU-complex itself, lor each selected subsystem, the
user is asked to identify one or more logical pathways which might be involved in the
situation. Kach of these logical pathways correspond to a line of communication between a
peripheral and an application program. On the basis of this information and details of the
basic composition of the network, the appropriate communications protocol can be selected.
The user is also asked to indicate which diagnostic tools (e.g., traces, dumps, logic probes)
arc available for examining each logical pathway.

Once the logical pathway and protocol have been determined, descriptions arc gathered of the
often multiple physical pathways that actually implement the logical pathway, it is on this
level that diagnostic test results arc presented and actual component indictments occur. For
DART to be useful at this level, the field engineer must be familiar with the diagnostic
equipment and software testing and tracing facilities which can be requested, and, of course,
must also have access to information about the specific system hardware and softwate
configuration of the installation. Finally, at the end of the consultation session, DART
summarizes its findings and recommends additional tests and procedures to follow. Figure
below depicts the major steps of the diagnostic process outlined above.

Report Findings and Recommendations

for the Session
Determine Protocol Violations and
Make Specific Component Indictments
for each Physical Pathway
Select Appropriate Protocol and
Available Diagnostic Tools
for each logical Pathway

Identify Logical Pathways

for each Subsystem

Infer Suspected CPU-I/O Subsystems

for each Problem

Gather Initial Symptom and

Configuration Information
for each System

The DART Diagnostic Inference Process

121
 After DART has indicated the components most likely to be at fault, the responsibility
for performing a detailed determination and repair of the actual component failures (i.e..
microcode bugs, integrated circuit failure, etc.) would then shift to the appropriate
maintenance groups for those components.

Result

The current DART knowledge base consists of 300 EMYCIN parameters and 190
production rules and was constructed over a period of 8 months. During this period 5
specialists were interviewed about different aspects of the diagnostic process and the
knowledge base reflects their composite expertise. Much of the requested diagnostic data is
already in a machine-readable form on the subject computer system. However, as the
transcript shows, this information must currently be entered by the user. This interaction
forms a substantial fraction of me users input. Indeed, wc estimate 30 to 60 percent of the
current interaction between program and user will be eliminated when this on-line data is
exploited.

It is clear that the communications protocols form the crux of the expertise, both for
the human specialist and for DART. Although the experts were able to easily articulate these
protocols, their translation into the production rule formalism was tedious. A protocol
represents an expected sequence of transactions between components. However, in order to
identify specific faulty components, the production rules capture only the possible deviations
from this expected sequence. Thus each protocol yields a substantial number of rules which
only indirectly reflect the original sequence and tend to produce rather opaque explanations
of the diagnostic reasoning. Furthermore, for any lengthy protocol, ensuring the completeness
of the resulting rule set becomes a significant problem. In a collateral effort, wc are
investigating the use of explicit representations of these protocols with general diagnostic
rules which will hypothesize deviations directly from the protocols.

ii. MYCIN.
MYCIN was developed at Stanford university by Shortliffe (1970’s). It was an
expert system for diagnosing blood diseases. It was a precursor to today’s expert
systems and acts as an ideal case study.

Features

MYCIN is a rule based ES using backward chaining.

A typical rule

IF stain of organism is gram negative

AND morphology is rod

AND is anaerobic

THEN suggestive that class is enterobacteriaceae

122
 The rules were actualyl stored as LISP expressions.

Inexact reasoning was employed using certainty factors. This is a number on the scale -1 to 1.
-1 being definitely false +1 definitely true. MYCIN used meta-rules to redirect search at
stages.

An example of such a meta rule is

IF infection is pelvic abscess

AND rules mention in premise E

AND rules mention in premise gram pos rods

THEN evidence should use rules for E before rules for gram pos rods

The above is guiding the search by saying which rules to try first.

To make the system acceptable to users (doctors) MYCIN incorporated Natural Language i.e
Users entered information in English and were answered in English. A spelling checker made
intelligent corrections to the input. Both WHY, WHY NOT and HOW explanation facilities
were provided. Several suggestions were often offered listed with some priority, allowing the
doctor some freedom of selection.

The Decision Process

There are four key questions in the process of deciding on treatment:
1. Does the patient have a significant infection?
2. What are the organism(s) involved?
3. What drugs might be work to treat the infection?
4. What is the best choice of drug or combination of drugs to treat the infection?
MYCIN Components
There are 5 basic components present in mycin. They are
1. KNOWLEDGE BASE: facts and knowledge about the domain
2. DYNAMIC PATIENT DATABASE: information about a particular case
3. CONSULTATION MODULE: asks questions, gives advice on a particular case
4. EXPLANATION MODULE: answers questions and justifies advice
5. KNOWLEDGE ACQUISITION PROGRAM: adds new rules and changes existing
rules
Phases of MYCIN

MYCIN has two phases in its approach, a diagnosis and a prescription phase.

In the first phase the nature of the infection and the organism causing the infection are
determined (hopefully). The prescription phase then indicates the drugs for the treatment
taking into account any side effects they may have on the patient.

MYCIN consists of about 500 rules. It is a backward chaining rule based system using
a depth-first search strategy.

123
Evaluation

In order to assess the performance of MYCIN a set of ten case histories of meningitis were
selected. The results of MYCIN were compared with medics of various degrees of expertise
from Stanford. They were also compared with evaluations from experts outside of Stanford
(the evaluators). They were to classify the prescriptions according to the categories:

1. Equivalent identical or equivalent recommendation.

2. Acceptable alternative different but also considered correct.
3. Unacceptable considered incorrect or inappropriate.
The evaluations were carried out in a double blind situation.

The results are

Prescriber %Acceptable No. Wrong

MYCIN 70 0

Prior Rx 70 0

Faculty 1 50 1

Faculty 2 50 1

Fellow 50 1

Faculty 4 50 0

Faculty 3 40 0

Resident 30 1

124
 Faculty5 30 0

Student 10 3

3. Explain the pitfalls in using an expert system. (May 2012)

Several weaknesses concerning the use of ES.

(a) Not Widely Used

ES is not widely used in business firms or organisations. Due to limited usage, firms
are still in doubt about the capability and, most definitely, the high cost involved in
investing in an ES.

(b) Difficult to Use

Using an ES is very difficult and learning and mastering it requires a long time. This
discourages managers from using ES. In one aspect, developing an ES that is user-
friendly is the biggest challenge for ES developer.

(c) Limited Scope

This is the most obvious weakness in an ES; its scope is very limited to its field only. In
the development aspect, the ES built is best developed because of its high accuracy.
However, usage-wise decision makers face constantly changing problems which involve
different fields that are interrelated.

(d) Probable Decision Error

The main source of the knowledge is experts. Humans make mistakes. If the experts input
wrong information into the Expert system, this will have a negative impact on the results
produced.

(e) Difficult to Maintain

The information in ES must be constantly updated to solve new problems. Every new
problem that occurs needs new knowledge and expertise. This means that there must be
an on-going relationship between the domain experts and the ES developer. This situation
requires the domain experts update the source of knowledge and expertise to suit the
current situation.

(f) Costly Development

The cost to consult a group of experts is not cheap, what if ES was built traditionally
without involving the use of an Expert System shell? On the other hand, programming

125
 cost is high because the artificial intelligence technique is difficult to master and needs a
very skilful programmer.

(g) Legal and Ethical Dilemma

We must be responsible for our actions and decisions. An expert has to take responsibility
for the information he or she provides. . The difficult question here is who should
shoulder the responsibility if a decision suggested by ES results in a negative outcome.

4. Explain XCON. (November 2012)

This system is an Expert configurer.

Stages of Expert System building

There are 6 stages in expert system building. They are

1. Identification

2. Conceptualization

3. Formalization

4. Implmentation

5. Testing and evaluation.

6. Maintainence

1. Identification: this phase is used to identify the problems, data, goals, company,
people…

• DEC, Digital Equipment Corporation Large computer manufacturer, started in

1957
• Catalogue has 40.000 different parts
• Buyer (with Sales Rep) sends order, typically 100 parts
• Delivery and assembly by DEC personnel
• Too often, part collection does not allow installation
• Too often, installed computer does not meet requirements
• Remedy: Completely assemble and test system in factory
• Automate configuration problem attempts with procedural languages were
unsuccessful
• XS approach started around 1980
2. Conceptualization: this phase is used to characterize different kinds of concepts and
relations
3. Formalization:it is used to express character of search
• Configuration engineers could talk well to Knowledge Engineers of the CSDG
• Could explain in what stage which component should be configured how

126
 • This was expressed in production rules
IF c1, c2 c3 THEN a1, a2, a3
• Configuration stage was explicitly represented as data: current goal or context
• Changing contexts moved configuration process through all stages
4. Implementation:
Build the system in executable form
• Language: OPS5 (similar to CLIPS)
• Conflict Resolution: MEA (extends Lex / Specificity)
• Means-Ends Analysis: order by recency of first condition
IF c1, c2 THEN .. is now different from IF c2, c1 THEN
• Contexts are treated as special by putting them first
• End-task is unspecific, thus executed last
• Use MEA + Spec to concentrate on subtasks:
• IF g1, x, y THEN assert barify // Signal necessity of subtask
• IF barify, a THEN p, q // Two rules perform the task
• IF barify, b THEN r, s // of barification per se
• IF barify THEN retract barify // Termination when ready
5. Testing and Evaluation :Does it do what we wanted?
• Field test after 1 year, production after another year
• Accuracy over 95%:No more pre-assembly was necessary!
• The installed configurations are optimized:Buyers are happier because they see
better productsLess retraining of staff on product changes:Quicker change of
production
• Net return to Digital is estimated to 40M$ per year.
6. Maintenance
Adapt to changing environment or requirements
Biological components

1. Users of XCON

• Sales: Use for quotations and ensure technical validity of orders (XSEL)
• Manufacturing: Check installability of order, guide assembly instructions and
diagnostics
• Field service: Easy assembly at customer’s site (XFL)
• Development: Anticipate integration problems for new products
The system found more use than what it was designed for

2. Key Roles

• Champion: Vision, belief in technology, influence on sponsor.

• Sponsor: Has interest in problem and control over resources: money & people
• Manager: Keeps parties together, goals realistic
• Knowledge Engineer: AI and Knowledge
• Software Engineer: Manage traditional parts, versions, architecture

127
 • Experts: Provide domain knowledge
• Users: Feedback on fit in business process

Data set of XCON

Five data bases:

1. Components: Holds 30k parts, 25 to 125 attributes

2. Container: Extra data on backplanes and cabinets, size and slots.
3. Diagrams: 1200 ASCII arts for use in output diagrams
4. Questions: Motivate and ask data from user, eg. Do you already have …?
5. Configuration: Consultation data like input and diagrams
Impact of RIME/XCON on later developments

• Use of Elicitation Tools has become common practice

• Such tools can bridge the gap between knowledge and their implementation as rule
(cp. Façade templates)
• Knowledge structure model can be made explicit in the tool and guide the knowledge
elicitation by providing the expert with questions in a context
• Success stimulated XS applications

5. Explain in detail about expert system shells. (June 2013)

An Expert system shell is a software development environment. It contains the basic
components of expert systems. A shell is associated with a prescribed method for building
applications by configuring and instantiating these components.

Shell components and description

128
 The generic components of a shell : the knowledge acquisition, the knowledge Base,
the reasoning, the explanation and the user interface are shown below. The knowledge base
and reasoning engine are the core components.

1. Knowledge Base

A store of factual and heuristic knowledge. Expert system tool provides one or
more knowledge representation schemes for expressing knowledge about the
application domain. Some tools use both Frames (objects) and IF-THEN rules. In
PROLOG the knowledge is represented as logical statements.

2. Reasoning Engine

Inference mechanisms for manipulating the symbolic information and

knowledge in the knowledge base form a line of reasoning in solving a problem. The
inference mechanism can range from simple modus ponens backward chaining of IF-
THEN rules to Case-Based reasoning.

3. Knowledge Acquisition subsystem

A subsystem to help experts in build knowledge bases. However, collecting

knowledge, needed to solve problems and build the knowledge base, is the biggest
bottleneck in building expert systems.

4. Explanation subsystem

A subsystem that explains the system's actions. The explanation can range from how
the final or intermediate solutions were arrived at justifying the need for additional
data.

5. User Interface

A means of communication with the user. The user interface is generally not a
part of the expert system technology. It was not given much attention in the past.
However, the user interface can make a critical difference in the perceived utility of
an Expert system.

6. . Explanation

129
 Most expert systems have explanation facilities that allow the user to ask
questions - why and how it reached some conclusion.

The questions are answered by referring to the system goals, the rules being used, and
existing problem solving. The rules typically reflect empirical, or "compiled"
knowledge. They are codes of an expert's rules of thumb, not the expert's deeper
understanding.

Example :

Dialog with an expert system designed to give advice on car problems.

System Does the car start ?

User No.

System Does the engine turn over ?

User Yes.

System Do you smell gas ?

User Yes

System Recommendation - Wait 5 minutes and try again.

User Why ?

System I used the rule:

If not start, and engine_turn_over, and smell_gas

Then recommendation is 'Wait 5 minutes and try again'

User

Note : The rule gives the correct advice for a flooded car, and knows the questions to
be ask to determine if the car is flooded, but it does not contain the knowledge of
what a flooded car is and why waiting will help.

Types of Explanation

There are four types of explanations commonly used in expert systems.

1. Rule traces reports on the progress of a consultation;

2. Explanation of how the system reached to the given conclusion;

3. Explanation of why the system did not give any conclusion.

4. Explanation of why the system is asking a question.

130
 6. What is knowledge acquisition? Explain in detail. (November 2013)
Knowledge acquisition includes the elicitation, collection, analysis, modeling and
validation of knowledge.

Issues in Knowledge Acquisition

The important issues in knowledge acquisition are:

■ knowledge is in the head of experts

■ Experts have vast amounts of knowledge

■ Experts have a lot of tacit knowledge

o They do not know all that they know and use

o Tacit knowledge is hard (impossible) to describe
■ Experts are very busy and valuable people

■ One expert does not know everything

■ Knowledge has a "shelf life"

Techniques for Knowledge Acquisition

The techniques for acquiring, analyzing and modeling knowledge are :

1. Protocol-generation techniques.
2. Protocol analysis techniques.
3. Hierarchy generation techniques.
4. Matrix-based techniques.
5. Sorting techniques.
6. Limited-information and constrained-processing tasks.
7. Diagram-based techniques.
■ Protocol-generation techniques

Include many types of interviews (unstructured, semi-structured and structured),

reporting and observational techniques.

■ Protocol analysis techniques

Used with transcripts of interviews or text-based information to identify basic

knowledge objects within a protocol, such as goals, decisions, relationships and attributes.
These act as a bridge between the use of protocol-based techniques and knowledge modeling
techniques.

Hierarchy-generation techniques

131
 Involve creation, reviewing and modification of hierarchical knowledge. Hierarchy-
generation techniques, such as laddering, are used to build taxonomies or other hierarchical
structures such as goal trees and decision networks. The Ladders are of various forms like
concept ladder, attribute ladder, composition ladders.

■ Matrix-based techniques

Involve the construction and filling-in a 2-D matrix (grid, table), indicating such
things, as may be, for example, between concepts and properties (attributes and values) or
between problems and solutions or between tasks and resources, etc. The elements within

the matrix can contain: symbols (ticks, crosses, question marks ) , colors , numbers , text.

Sorting techniques

Used for capturing the way people compare and order concepts; it may reveal
knowledge about classes, properties and priorities.

■ Limited-information and constrained-processing tasks

Techniques that either limit the time and/or information available to the expert when
performing tasks. For example, a twenty-questions technique provides an efficient way of
accessing the key information in a domain in a prioritized order.

■ Diagram-based techniques

Include generation and use of concept maps, state transition networks, event diagrams
and process maps. These are particularly important in capturing the "what, how, when, who
and why" of tasks and events.

7. Explain the characteristics, role and advantages of expert system? ( NOV 2013)

Expert system operates as an interactive system that responds to questions, asks for
clarifications, makes recommendations and generally aids the decision-making process.

Expert systems have many Characteristics:

1. Operates as an interactive system

This means an expert system:

 Responds to questions
 Asks for clarifications
 Makes recommendations
 Aids the decision-making process.
2. Tools have ability to sift (filter) knowledge

 Storage and retrieval of knowledge

 Mechanisms to expand and update knowledge base on a continuing basis.

132
3. Make logical inferences based on knowledge stored

 Simple reasoning mechanisms is used.

 Knowledge base must have means of exploiting the knowledge stored, else it is
useless; e.g., learning all the words in a language, without knowing how to combine
those words to form a meaningful sentence.
4. Ability to Explain Reasoning

 Remembers logical chain of reasoning; therefore user may ask

o for explanation of a recommendation
o factors considered in recommendation
 Enhances user confidence in recommendation and acceptance of expert system
5. Domain-Specific

o A particular system caters a narrow area of specialization; e.g., a medical expert

system cannot be used to find faults in an electrical circuit.
o Quality of advice offered by an expert system is dependent on the amount of
knowledge stored.
6. Capability to assign Confidence Values

o Can deliver quantitative information

o Can interpret qualitatively derived values
o Can address imprecise and incomplete data through assignment of confidence values.
7. Applications

o Best suited for those dealing with expert heuristics for solving problems.
o Not a suitable choice for those problems that can be solved using purely numerical
techniques.
8. Cost-Effective alternative to Human Expert

o Expert systems have become increasingly popular because of their specialization,

albeit in a narrow field.
o Encoding and storing the domain-specific knowledge is economic process due to
small size.
o Specialists in many areas are rare and the cost of consulting them is high; an expert
system of those areas can be useful and cost-effective alternative in the long run.
ROLE OF EXPERT SYSTEMS

The Expert systems have found their way into most areas of knowledge work. The
applications of expert systems technology have widely proliferated to industrial and
commercial problems, and even helping NASA to plan the maintenance of a space shuttle for
its next flight. The main applications are

1. Diagnosis and Troubleshooting of Devices and Systems

133
Medical diagnosis was one of the first knowledge areas to which Expert system technology
was applied in 1976. However, the diagnosis of engineering systems quickly surpassed
medical diagnosis.

2. Planning and Scheduling

The Expert system's commercial potential in planning and scheduling has been recognized as
very large. Examples are airlines scheduling their flights, personnel, and gates; the
manufacturing process planning and job scheduling;

3. Configuration of Manufactured Objects from sub-assemblies

Configuration problems are synthesized from a given set of elements related by a set of
constraints. The Expert systems have been very useful to find solutions. For example,
modular home building and manufacturing involving complex engineering design.

4. Financial Decision Making

The financial services are the vigorous user of expert system techniques. Advisory programs
have been created to assist bankers in determining whether to make loans to businesses and
individuals. Insurance companies to assess the risk presented by the customer and to
determine a price for the insurance. ES are used in typical applications in the financial
markets / foreign exchange trading.

5. Knowledge Publishing

This is relatively new, but also potentially explosive area. Here the primary function of the
Expert system is to deliver knowledge that is relevant to the user's problem. The two most
widely known Expert systems are : one, an advisor on appropriate grammatical usage in a
text; and the other, is a tax advisor on tax strategy, tactics, and individual tax policy.

6. Process Monitoring and Control

Here Expert system does analysis of real-time data from physical devices, looking for
anomalies, predicting trends, controlling optimality and failure correction. Examples of real-
time systems that actively monitor processes are found in the steel making and oil refining
industries.

7. Design and Manufacturing

Here the Expert systems assist in the design of physical devices and processes, ranging from
high-level conceptual design of abstract entities all the way to factory floor configuration of
manufacturing processes.

ES usage provides many advantages. Some of the advantages are:

(a) Consistency

134
One of the advantages of an ES is that the results given are consistent. This might be due to
the fact that there are no elements such as exhaustion and emotions as experienced by
humans.

(b) Hazardous Working Environment

Through an ES, we can avoid exposing ourselves to a toxic or radioactive environment. An

ES can take over the place of an expert to handle problems in a high-risk area such as a
nuclear power plant.

(c) Ability to Solve Complex and Difficult Problems

A very difficult problem encountered by an organisation, if not taken seriously, can cause an
adverse impact such as losses or cancellation of a business deal. Sometimes, the problems
need to be attended to quickly. The problems can become more complicated when individuals
or experts involved in solving them are absent or cannot be contacted. Thus, an ES serves as
an alternative to experts.

(d) Combination of Knowledge and Expertise from Various Sources

One of the important components in an ES is the knowledge base. This component contains
the accumulated knowledge and acquired or transferred expertise from many experts. Thus,
an ES is sometimes more superior than an expert because its knowledge and expertise have
come from many sources.

(e) Training Tool for Trainees

An ES can be used by trainees to learn about the knowledge-based system. Trainee who uses
an ES would be able to observe how an expert solves a problem.

Several weaknesses concerning the use of ES.

(a) Not Widely Used

ES is not widely used in business firms or organisations. Due to limited usage, firms are still
in doubt about the capability and, most definitely, the high cost involved in investing in an
ES.

(b) Difficult to Use

Using an ES is very difficult and learning and mastering it requires a long time. This
discourages managers from using ES. In one aspect, developing an ES that is user-friendly is
the biggest challenge for ES developer.

(c) Limited Scope

This is the most obvious weakness in an ES; its scope is very limited to its field only. In the
development aspect, the ES built is best developed because of its high accuracy. However,

135
usage-wise decision makers face constantly changing problems which involve different fields
that are interrelated.

(d) Probable Decision Error

The main source of the knowledge is experts. Humans make mistakes. If the experts input
wrong information into the Expert system, this will have a negative impact on the results
produced.

(e) Difficult to Maintain

The information in ES must be constantly updated to solve new problems. Every new
problem that occurs needs new knowledge and expertise. This means that there must be an
on-going relationship between the domain experts and the ES developer. This situation
requires the domain experts update the source of knowledge and expertise to suit the current
situation.

(f) Costly Development

The cost to consult a group of experts is not cheap, what if ES was built traditionally without
involving the use of an Expert System shell? On the other hand, programming cost is high
because the artificial intelligence technique is difficult to master and needs a very skilful
programmer.

(g) Legal and Ethical Dilemma

We must be responsible for our actions and decisions. An expert has to take responsibility for
the information he or she provides. . The difficult question here is who should shoulder the
responsibility if a decision suggested by ES results in a negative outcome.

136
 Industry/ Practical Connectivity

Industry Connectivity

 The student can work as a software engineer in industry, which is working for
companies like Amazon to shopping list recommendation engines or Facebook
analyzing and processing big data.
 The students can also work as a hardware engineer developing electronic parking
assistants or home assistant robots.

Latest developments in AI

 It can be used in biological research such as detection of diseases.

 The Stanford university team builds Shazam, an expert system for detecting
earthquakes.
 Gabriel is A New Artificial Intelligence Named after the Messenger Angel that acts as
a personal cognitive assistant and whispers instructions into human’s ear.

137
138
139
140
141
142
143
 Anna University
B.E./B.Tech. DEGREE EXAMINATION, APRIL/MAY 2011
Sixth Semester
Computer Science and Engineering
CS 2351 — ARTIFICIAL INTELLIGENCE
(Regulation 2008)
PART A — (10 × 2 = 20 marks)
1. List down the characteristics of intelligent agent.
2. What do you mean by local maxima with respect to search technique?
3. What factors determine the selection of forward or backward reasoning
approach for an AI problem?
4. What are the limitations in using propositional logic to represent the
knowledge base?
5. Define partial order planner?
6. What are the differences and similarities between problem solving and
planning?
7. List down two applications of temporal probabilistic models.
8. Define Dempster-Shafer theory.
9. Explain the concept of learning from example.
10. How statistical learning method differs from reinforcementlearning method?
PART B — (5 × 16 = 80 marks)
11. (a) Explain in detail on the characteristics and applications of learning agents.
Or
(b) Explain AO* algorithm with an example.
12. (a) Explain unification algorithm used for reasoning under predicate logic with an
example.
Or
(b) Describe in detail the steps involved in the knowledge Engineering process.
13. (a) Explain the concept of planning with state space search using suitable examples.
Or
(b) Explain the use of planning graphs in providing better heuristic estimate with suitable
examples.
14. (a) Explain the method of handling approximate inference in Bayesian Networks.
Or
(b) Explain the use of Hidden Markov Models in Speech Recognition.
15. (a) Explain the concept of learning using decision trees and neural network
approach.
Or
(b) Write short notes on :
(i) Statistical learning. (8)
(ii) Explanation based learning. (8)

144
 Anna University
B.E./B.Tech. DEGREE EXAMINATION, NOVEMBER/DECEMBER 2011.
Sixth Semester
Computer Science and Engineering
CS 2351 — ARTIFICIAL INTELLIGENCE
(Common to Seventh Semester – Electronics and Instrumentation Engineering)
(Regulation 2008)

PART A — (10 × 2 = 20 marks)

1. What is a rational agent?
2. State the significance of using heuristic functions?
3. Distinguish between predicate and propositional logic.
4. What factors justify whether the reasoning is to be done in forward or backward
reasoning?
5. Distinguish between state space search and plan space search.
6. Define partial order planning.
7. List two applications of Hidden Markov model.
8. What are the logics used in reasoning with uncertain information?
9. Define Inductive learning.
10. Distinguish between supervised learning and unsupervised learning.
PART B — (5 × 16 = 80 marks)
11. (a) Explain AO* algorithm with a suitable example. State the limitations in the
algorithm.
Or
(b) Explain the constraint satisfaction procedure to solve the crypt arithmetic problem.
CROSS
+ ROADS
DANGER
12. (a) Consider the following facts
Team India
Team Australia
Final match between India and Australia
India scored 350 runs Australia score 350 runs India lost 5 wickets Australia lost 7
wickets
The team which scored the maximum runs wins
If the scores are same then the team which lost minimum wickets wins the match.
Represent the facts in predicate, convert to clause form and prove by resolution
"India wins the match".
Or
(b) Analyse the missionaries and Cannibals problem which is stated as follows. 3
missionaries and 3 cannibals are on one side of the river along with a boat that can
hold one or two people. Find a way to get everyone to the other side, without leaving
a group of missionaries in one place outnumbered by the cannibals in that place.
(i) Formulate a problem precisely making only those distinctions necessary to

145
ensure a valid solution. Draw a diagram of the complete state space.
(ii) Design appropriate search algorithm for it.
13. (a) Explain the concept of planning with state space search. How is it different from
partial order planning?
Or
(b) What are planning graphs? Explain the methods of planning and acting in the real
world.
14. (a) Explain the concept of Bayesian network in representing knowledge in an
uncertain domain.
Or
(b) Write short notes on : (i) Temporal models
(ii) Probabilistic Reasoning.
15. (a) Explain in detail learning from observation and explanation based learning.
Or
(b) Explain in detail statistical learning methods and reinforcement learning.

146
 B.E/B.TECH DEGREE EXAMINATION,APRIL/MAY 2010

Sixth semester (Regulation 2004)

CS 1351 – ARTIFICIAL INTELLIGENCE

Time: three hours maximum: 100 marks

Answer ALL questions PART A- (10 x 2 = 20 marks)

1.Define al rational agent

2.How will you measure the problem solving performance?

3. State the reasons when the hill climbing often gets stuck

4.What is the constraint satisfaction problem?

5. Differentiate between prepositional versus first order logic

6.Define ontological engineering

7.What is explanation- based learning?

8.State the advantages of inductive logic programming.

9. Give the component steps of communication.

10. What are machine translation systems?

PART B – (5 x 16 = 80 marks)

11.(a) Discuss in detail the structure of agents with suitable diagrams?

(or)

(i) Explain the following uninformed search strategies

(iii) Iterative deepening depth-first search

(jjj) Bidirectional search

1 (a) Explain the A* search and give the proof of optimality of A*

(Or)

147
 (b) Describe the Min-Max Algorithm and Alpha –beta Pruning

13. (a) (i)Describe the general process of knowledge engineering

(ii)Discuss the syntax and semantics of first order logic

(Or)
(b) Describe the forward chaining and backward chaining algorithm with
suitable Example

14.(a)(i) Describe the decision tree learning algorithm

(ii)Explain the relevance-based learning

(Or)
(b) Discuss active and passive reinforcement learning with suitable example

15.(a) (i)Describe the semantic interpretation

(ii)Illustrate the grammar induction with suitable example (Or)

(b) Discuss on information retrieval systems and information

148