Code: AC20/AT21 Subject: ARTIFICIAL INTELLIGENCE & NEURAL NETWORKS Time: 3 Hours Max. Marks: 100 NOTE: There are 9 Questions in all.Question 1 is compulsory and carries 20 marks. Answer to Q. 1. must be

kettlecatelbowcornerAI and Robotics

Nov 7, 2013 (3 years and 8 months ago)

52 views


Code: AC20/AT21

Subject: ARTIFICIAL INTELLIGENCE & NEURAL
NETWORKS

Time: 3 Hours



Max.
Marks: 100



NOTE: There are 9 Questions in all.




Question 1 is compulsory and carries 20 marks. Answer to Q. 1. must be
written in the space provided for it in the answer book supplied and nowhere
else.




Out of the remaining EIGHT Questions answer any FIVE Questions. Each
question carries 16

marks.




Any required data not explicitly given, may be suitably assumed and stated.





Q.1


Choose the correct or best alternative in the
following:


(2x10)




a.


Which of the following ver
bs can be represented as MTRANS using
Schank’s CD representation







(A)


walk


(B)

tell

(C)



give


(D)


run




b.


Non monotonic reasoning
deals with



(A)


complete information


(B)


incomplete information

(C)


an always growing fact base


(D)


none of the above






c.


In the alpha
-
beta minimax search, alpha is _____________ on the values that
a


_
_________ node maybe ultimately assigned.





(A)


a lower bound, maximizing


(B)


an upper bound, minimizing

(C)


a lower bound, minimizing


(D)


an upper bound, maximizing





d.


In connectionism, _____________ is a search technique and the learning
techniques are _______________ and ___________.



(A)


Parall
el relaxation, version spaces, A*.



(B)


Parallel relaxation, explanation
-
based learning, and discovery.

(C)


Parallel relaxation, backpropogation, reinforcement learning.

(D)


Parallel relaxation, hill climbing, reinforcement learning.






e.


The ID3 algorithm constructs decision trees by choosing




(A)


Attributes that will yield more information



(B)


Attributes that will yield less information

(C)


Attributes that will


not yie
ld any information



(D)


None of the above





f.


Generalisation hierarchy does not support property inheritance



(A)

True


(B)


False





g.


Resolution works on statements that ar
e in single canonical form



(A)


True


(B)


False





h.


Natural deduction doesn’t use human theorem proving



(A)


True


(B)


False





i.



Unification is a process to represent substitutions during pattern matching



(A)


True


(B)


False





j.


Bayesian networks uses more of local representation which describe clusters






(A)


True


(B)


False







Answer any FIVE Questions out of EIGHT Questions.

Each question carries 16 marks.







Q.2


a.


Discuss the significance of State space search in solving an AI
problem.



(8)






b.


What are the various search strategies known? Differentiate among
them.


(8)






Q.3


a.


Symbolize and validate the following argument.



All law
-
abiding citiz
ens pay their taxes.



Mr. Shyam pays his taxes.



Therefore, Mr. Shyam is a law
-
abiding
citizen.


(6)





b.


Differentiate between AO* and A*
algorithm.



(10)





Q.4


a.


Let STACK(X,Y) and PICKUP(X) be two of the operators for
manipulating blocks on a table using one arm of


a robot. Formulate the
precondition, delete and add lists for these operators. Thes
e operators are
to be used by a simple planner using a goal stack.



Suppose the goal to be achieved is ON(C,A) and ONTABLE(A) from the
initial state ONTABLE(A) and ONTABLE(C). Explain the details of plan
generation where the plan is PICKU
P(C) followed by
STACK(C,A).


(10)








b.


Explain Alpha
-
cut off and Beta
-
cut off with
examples.


(6)





Q.5


a.


Discuss the various problems being faced by Expert
System.


(6)






b.


Consider the following set of propositions





patient has spots




patient has measles



patient has high fever



patient has been inoculated against measles



patient has an allergy






(i) Create a network that defines the casual connections among these
nodes.





(ii) Make a Bayesian network by constructing the necessary conditional
probability matrix.


(10)





Q.6


a.


Describe in general , claus
es, facts, goals and rules of PROLOG. Elaborate
them with respect to the knowledge given below.





“A house has a roof, a door and a window. A door has a knob. If an
entity is a house, then it has a door. My house is


a house.”


(8)





b.


Write a PROLOG program to find the maximum number from a given list
of numbers.


(8)






Q.7


a.


Describe the candidate elimination algorithm in narrowing the version
space using an

example.




(8)





b.


Explain the various components of a planning
system.


(8)





Q.8


Write short notes on

(i)


Non

linear planning

(ii)


Non Monotonic Reasoning

(iii)


Characteristics of a production system

(iv)


Activation functions of a Neuron.


(4

4)






Q.9


a.


Consider the following sentence





John punched Bill.

(i)


Show how it would be represen
ted in case grammar.

(ii)


Show how it would be represented using conceptual dependency
graph.

(iii)


Discuss the advantages and disadvantages of these
representations.


(10)





b.


Describe the different strategies that
can be used for knowledge
Acquisition.


(6)