Scribd
Upload
12 views6 pages

2024 CSC14003 ReviewExercisesForFinal

The document outlines a final examination review for an Introduction to Artificial Intelligence course, covering topics such as decision tree classifiers, propositional logic, first-order logic, and neural networks. It includes specific questions requiring calculations, logical proofs, and representations in formal logic. The review emphasizes practical applications of AI concepts through exercises and problem-solving tasks.

Uploaded by

thaihoangphuc7076
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
12 views6 pages

2024 CSC14003 ReviewExercisesForFinal

The document outlines a final examination review for an Introduction to Artificial Intelligence course, covering topics such as decision tree classifiers, propositional logic, first-order logic, and neural networks. It includes specific questions requiring calculations, logical proofs, and representations in formal logic. The review emphasizes practical applications of AI concepts through exercises and problem-solving tasks.

Uploaded by

thaihoangphuc7076
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 6

Course: CSC14003 – Introduction to Artificial Intelligence

Class 22CLC – Term Ii/2023-2024

FINAL EXAMINATION REVIEW

Question 01. (Adapted from Prof. Ziv-Bar Joseph, Carnegie Mellon University, Course 10-701
Machine Learning materials.) NASA wants to discriminate Martians (M) from Humans (H)
based on these features (attributes): Green ∈ {𝑁, 𝑌}, Legs ∈ {2, 3}, Height ∈ {𝑆, 𝑇}, Smelly ∈
{𝑁, 𝑌}. Your available training data is as follows (N = No, Y = Yes, S = Small, T = Tall). Note that
it is just a made-up problem for the exercise, anything can happen!

# Height Green Legs Smelly Target: Species0


1 S Y 3 Y M
2 T Y 3 N M
3 S Y 3 N M
4 T Y 3 N M
5 T N 2 Y M
6 T Y 2 Y H
7 S N 2 N H
8 T N 3 N H
9 S N 3 N H
10 T N 3 N H

Build an ID3 decision tree classifier from the above training dataset. Attributes are evaluated
using Information Gain. Ties are broken such that an attribute with earlier alphabetical order is
preferred.

1) Present the calculations required to choose the attribute for the root node.
Target Height Green Legs Smelly

S T Y N 2 3 Y N

AE

IG

The root is ____________________________________

2) Present the calculations required to choose the attribute for the root’s left branch.
Target Height Legs Smelly

AE

IG

The left child of the root is _____________________________________

3) Present the calculations required to choose the attribute for the root’s right branch.
Target Height Legs Smelly

AE

IG

The right child of the root is _____________________________________

4) Draw the complete ID3 decision tree

Question 02. From the sentence "Heads I win, tails you lose," prove that "I win" by using
propositional logic refutation resolution.
a) Represent the given sentence in propositional logic using only the following prepositions:
Head (the coin’s head), Tail (the coin’s tail), IWin (I win), and YouLose (you lose)

b) Add some general knowledge axioms about coins, winning, and losing.

c) Convert the propositional logic sentences to their CNF equivalents.

d) Apply propositional refutation resolution to prove the above conclusion.


Question 03. Consider the following KB.
1. Buffalo(x) ∧ Pig(y) → Faster(x,y) 4. Buffalo(Bob)
2. Pig(y) ∧ Slug(z) → Faster (y,z) 5. Pig(Pat)
3. Faster(x,y) ∧ Faster (y, z) → Faster(x,z) 6. Slug(Steve)

Use first-order logic forward chaining to prove Faster(Bob, Steve). If several rules apply, use
the one with the smallest number. Show the chaining process step by step, using the numbering
of the sentences to identify how you are using the rules and facts in the KB. For either
presentation method, you will need to indicate the unifications.

Question 04. Consider the following English sentences


1. Every child loves Santa.
2. Everyone who loves Santa loves any reindeer.
3. Rudolph is a reindeer, and Rudolph has a red nose.
4. Anything which has a red nose is weird or is a clown.
5. No reindeer is a clown.
6. Scrooge does not love anything which is weird.

a) Build a FOL knowledge base from the above sentences, using only given predicates.
CHILD(x): “x is a child” LOVES(x, y): “x loves y”
REINDEER(x): “x is a reindeer” REDNOSE(x): “x has red nose”
WEIRD(x): “x is weird” CLOWN(x): “x is a clown”

b) Convert the propositional logic sentences to their CNF equivalents.


c) Apply propositional refutation resolution to prove that “Scrooge is not a child.”

Question 05. Rewrite the following sentence in first-order logic.

Politicians can fool some people all of the time, and they can fool all people some of the
time, but they can’t fool all of the people all of the time.

using only the predicates given below.


Politician(x): x is a politician Person(x): x is a person
Time(t): t is a time period Fool(x, y, t): x fools y at time t

Problem 06. Consider the multi-layer percceptron shown below.


The network parameters setting is as follows.
● The set of weights are
w1 = 0.62 w2 = 0.42 w3 = 0.55 w4 = -0.17 w5 = 0.35 w6 = 0.81
● There is no bias. There is no threshold.

● Learning rate is set to 0.25.

● Sigmoid activation function

a) Compute the output values for all the hidden and output neurons when input signals come
to neuron 1 and 2 are both 1s and output signal is 1.

Neuron 3 4 5

Output

b) Adjust the weights according to the computed values above.

Weights w1 w2 w3 w4 w5 w6

Values

Question 07. In the network shown below, all the units have binary inputs (0 or 1), unipolar
step functions and binary outputs (0 or 1). The weights for this network are w 31 = 1, w32 = 1, w41
= 1, w42 = 1 and w43 = −2. The threshold of the hidden unit (3) is 1.5 and the threshold of the
output unit (4) is 0.5. The threshold of both input units (1 and 2) is 0.5, so the output of these
units is the same as the input.

Which Boolean functions can be computed by this network? Justify your answer by showing
detailed calculations.

1 2 3 4

You might also like