21
SECTION A : True / False TOTAL 10 MARKS Answer TRUE or FALSE for the following and justify your answer by stating the reason for each answer. There are 5 questions in this section. Each question carries 2 marks. 1. Depth first search will perform best if the goal state lies in upper left-hand side of the tree. 2. Alpha-beta pruning is a procedure to reduce the amount of computation and searching during minimax. 3. A recursive procedure breaks a large and difficult problem into smaller problems. 4. The concept of frames in AI is similar to that of files in data organization. AI Final Sem. 1 2012/2013 Page: 1

SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Embed Size (px)

DESCRIPTION

SCSJ3553 - Artificial Intelligence Final Exam paper - UTM punye

Citation preview

Page 1: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

SECTION A : True / False TOTAL 10 MARKS

Answer TRUE or FALSE for the following and justify your answer by stating the reason for each answer. There are 5 questions in this section. Each question carries 2 marks.

1. Depth first search will perform best if the goal state lies in upper left-hand side of the tree.

2. Alpha-beta pruning is a procedure to reduce the amount of computation and searching

during minimax.

3. A recursive procedure breaks a large and difficult problem into smaller problems.

4. The concept of frames in AI is similar to that of files in data organization.

5. One of applications of machine learning to large databases is web mining such as Google.

AI Final Sem. 1 2012/2013 Page: 1

Page 2: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

SECTION B : Short Explanation Questions TOTAL 10 MARKS Answer each question in the space provided. There are 5 questions in this section. Each question carries 2 marks.

1. A heuristic evaluation function f(n) may be calculated using the formula f(n) = g(n) +

h(n). In the 8-puzzle game h(n) can be:

a) _____________________________________________________________________

b) _____________________________________________________________________

2. What are the 2 rules for alpha beta pruning to stop the searching?

3. Explain the difference between forward and backward chaining. How do we choose

between forward and backward chaining?

4. State the similarities and differences between a semantic network and a conceptual graph.

5. In crime domain, what machine learning applications can be applied to assist in the

detection of criminals?

AI Final Sem. 1 2012/2013 Page: 2

Page 3: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

D

S

A

H B F Q

I F G

PF

SECTION C : Structured Questions TOTAL 80 MARKS

Answer all questions in the space provided. There are 4 questions in this section. Use extra paper(s) if necessary. Each question carries a total of 20 marks.

Question 1 (a) – Uninformed Search Total 10 marks

Based on Figure: 1, answer questions (i) to (iii).

Figure: 1

i) Perform a Breadth-First search on the above tree. List the nodes on OPEN and CLOSED

for each iteration. S = Start, and F = Goal. [ 3 marks ]

Iteration OPEN CLOSED

ii) The Breadth-First search algorithm can be implemented in a

___________________________________ order or with a __________. [1.5 marks]

AI Final Sem. 1 2012/2013 Page: 3

Page 4: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

iii) Perform a Depth-First search on the above tree. List the nodes on OPEN and

CLOSED for each iteration. S = Start, and F = Goal. [ 3 marks ]

Iteration OPEN CLOSED

iv) The Depth-First Search algorithm can be implemented in a

___________________________ order or with a ________. [1.5 marks]

v) In Sudoku game application which search technique is better to be applied (Best first

or depth first search ? ). Give a short reason why (BFS/DFS) is more suitable for

Sudoku.

[ 1 marks]

AI Final Sem. 1 2012/2013 Page: 4

Page 5: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Question 1(b) – Informed Search Total 10 marks

Figure: 2 shows a tree where each node is assigned a heuristic value. Based on this tree,

answer questions (i) to (iii).

Figure: 2

i) Apply the Best-First search from the start state A to reach to the goal state, K. Show

all the steps you will perform by filling in the table below. [4 marks]

Node Evaluated OPEN CLOSED

ii) What is the solution path for the Best-First search performed on this tree? [1 mark]

AI Final Sem. 1 2012/2013 Page: 5

A, 3

B, 3

D, 2

N, 0

H, 1

O, 0

E, 3

C, 2

F, 1

P, 2

J, 1I, 2

Q, 2

L, 1K, 0

G, 2

S, 3R, 2

M, 2

Page 6: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

iii) What are the evaluation function that is the f(n) values for nodes A, B, G, I and Q? Show how

you obtain each value. [5 marks]

AI Final Sem. 1 2012/2013 Page: 6

Page 7: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

5 7 4 56

3

8

67 5

9

6

6

9

Question 2a – Minimax Search Total 10 marks

Refer to Figure: 3 to answer this question.

Figure: 3

Apply the minimax algorithm to the game tree in Figure: 3 above, where it is the

MAXIMIZER’s turn to play. The values estimated by the evaluation function are indicated at

the leaf nodes. Assume that the search always visits children left-to-right.

Compute the backed-up values computed by the minimax algorithm. Show your answer by

writing values at the appropriate nodes in the above tree. [7m]

a. Indicate the proper move of the maximizer by marking the line with bold indicator

( ) to the root’s outgoing edges. [3m]

AI Final Sem. 1 2012/2013 Page: 7

MAX

Page 8: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Question 2b – Alpha-Beta Pruning Total 10 marks

Refer to Figure 3 above, answer the following questions

a. Using alpha-beta pruning (and standard left-to-right evaluation of nodes), how many of the

leaves get evaluated? [4m]

b. Indicate all parts of the tree that are cut off. [3m]

c. if the standard right-to-left evaluation of nodes is applied to the search tree in Figure 3

above:

i. are the minimax value computed at the root will be changed. (just write down yes or

no) [1.5m]

ii. are the number of nodes pruned will be changed. (just write down yes or no)[1.5m]

AI Final Sem. 1 2012/2013 Page: 8

Page 9: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Question 3 (a) – 15 marks

A production system maintains a set of rules about the characteristics of fruits as follows:

Rule 1: IF Shape = long ANDColour = green OR yellow

THEN Fruit = banana

Rule 2: IF Shape = round OR oblong ANDDiameter > 4 inches

THEN Fruitclass = vine

Rule 3: IF Shape = round ANDDiameter < 4 inches

THEN Fruitclass = tree

Rule 4: IF Seedcount = 1THEN Seedclass = stonefruit

Rule 5: IF Seedcount > 1THEN Seedclass = multiple

Rule 6: IF Fruitclass = vine ANDColour = green

THEN Fruit = watermelon

Rule 7: IF Fruitclass = vine ANDSurface = smooth ANDColour = yellow

THEN Fruit = honeydew

Rule 8: IF Fruitclass = vine ANDSurface = rough ANDColour = tan

THEN Fruit = cantaloupe

Rule 9: IF Fruitclass = tree ANDColour = orange ANDSeedclass = stonefruit

THEN Fruit = apricot

Rule 10: IF Fruitclass = tree ANDColour = orange ANDSeedclass = multiple

THEN Fruit = orange

Rule 11: IF Fruitclass = tree ANDColour = red ANDSeedclass = stonefruit

THEN Fruit = cherry

AI Final Sem. 1 2012/2013 Page: 9

Page 10: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Rule 12: IF Fruitclass = tree ANDColour = orange ANDSeedclass = stonefruit

THEN Fruit = peach

Rule 13: IF Fruitclass = tree ANDColour = red OR yellow OR green ANDSeedclass = multiple

THEN Fruit = apple

Rule 14: IF Fruitclass = tree ANDColour = purple ANDSeedclass = stonefruit

THEN Fruit = plum

i) Use FORWARD CHAINING to describe the production system table including its

working memory, conflict set and rule fired to establish a fruit. Initial data given is :

Shape = round Diameter > 4 inches Surface = smoothColour = yellow

Terminate when the final value for Fruit in the working memory. [6 marks]

Iteration #

Working memory Conflict set

Rule fired

AI Final Sem. 1 2012/2013 Page: 10

Page 11: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

ii) Given the fruit to search is apple, use BACKWARD CHAINING to describe the

production system table including its working memory, conflict set and rule fired to

establish the initial data for this fruit.

State the initial facts required to establish that the fruit searched is an apple. [9 marks]

Iteration #

Working memory Conflict set

Rule fired

The initial facts required to establish fruit to search is apple are:

AI Final Sem. 1 2012/2013 Page: 11

Page 12: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Question 3 (b) – 5 marks

i) How is blackboard systems related to production systems? Give a diagrammatic

example of such an application. [2.5 marks]

ii) How can the concept of agents be applied in the blackboard architecture? [2.5

marks]

AI Final Sem. 1 2012/2013 Page: 12

Page 13: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Question 4 (a) – 10 marks

Develop a knowledge representation by drawing a SEMANTIC NETWORK to model the

classification of musical instruments as described below.

Musical Instruments are classified according to five major categories namely

Idiophones, Membranophones, Aerophones, Chordophones and Electrophones.

Idiophones are played by striking, shaking, plucking, or rubbing. Examples of

struck idiophones are bells and cymbals. Shaken Idiophones are rattles.

Examples of plucked idiophones are harps.

Membranophones are played by striking or rubbing. Examples are kettledrums

and snare drums. They are played by striking on them. Membranophones are

made from stretched skin or membrane.

Aerophones are played by blowing or breathing. Trumpets and Horns players

must blow or compress their lips to set the air into motion. For flutes, the

player’s breath is directed against the sharp edge of a hole to set the air into

motion.

Chordophones may be plucked or bowed. Examples are Harps.

Electrophones are played using electronic circuits. Examples are Electric guitar.

AI Final Sem. 1 2012/2013 Page: 13

Page 14: SCSJ3553 - Artificial Intelligence Final Exam paper - UTM

Question 4 (b) – 10 marks

Model your knowledge by assembling the partial description given about Multiple

Intelligences below into a STRUCTURED FRAME-BASED representation.

Gardner (1983) introduced the Theory of Multiple Intelligences which meets

with the criteria for a behavior to be intelligent. He believes that there are 8

types of intelligences: Spatial, Linguistic, Logical-mathematical, Bodily-kinesthetic,

Musical, Interpersonal, Intrapersonal and Naturalistic.

Spatial intelligence deals with spatial judgment, and ability to visualize with

the mind’s eye. Potential careers for them are Architect, Artist, and Engineer.

Linguistic intelligence deals with spoken or written words. Potential careers for

them are Writer, Lawyer and Teacher.

Logical-mathematical intelligence deals with logic, abstractions, reasoning and

numbers. Potential careers for them are Scientist, Mathematician, Computer

programmer, and Engineer.

Intrapersonal intelligence deals with introspective and self-reflective capacities.

Potential careers for them are Philosopher, Writer, Theorist and Scientist.

Naturalistic intelligence deals with the ability to recognize and categorize

plants, animals and other objects in nature. Potential careers for them are

Farmers, Botanists and Chef.

Note: The JOB frame will only cater for job titles that are in italics i.e. Engineer, Writer, Scientist and Chef.

AI Final Sem. 1 2012/2013 Page: 14