Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

PART I. EXPLANATION. 10 POINTS EACH.

1. Why are representations so important in Artificial Intelligence? What


risks are inherent in using the wrong representation?

ANSWER:

Humans excel at comprehending, thinking, and analyzing information. Humans know things
and use that knowledge to execute different activities in the natural world. However,
information interpretation and logic deal with how computers do all of these things. As a
result, the following is a summary of knowledge representation:
Information representation and reasoning (KR, KRR) is a branch of artificial intelligence that
studies how AI agents think and how their thoughts influence their intellectual actions.

It's in charge of encoding information about the natural world so that a robot can
comprehend and use it to solve complex real-world challenges like diagnosing a medical
disorder or interacting with humans in natural language.

It's also a way of describing how artificial intelligence should reflect information. Knowledge
representation is more than simply putting data into a folder. Nonetheless, it allows an
intelligent computer to benefit from that intelligence and experience to act intelligently as a
person.
The danger faced by a financial reporting mistake due to a cause other than a lack of
internal control is an inherent risk.

In a financial audit, the intrinsic risk is more likely to emerge where contracts are complicated
or when financial estimates require a high level of judgment.

2. Explain why goal trees are so useful to artificial intelligence research.


Give illustrations of how they are used.

ANSWER:

Goal-based agents take into account long-term decisions and the desirability of the result,
making them easy to train and adaptable to changing conditions.

Problem-solving and preparation agents are the two types of goal-based agents. Each state
of the universe is regarded by problem-solving agents as inseparable, with no internal
structure apparent to the problem-solving algorithms. Every state is divided into variables by
planning agents, who create relationships between them.

Five components can be used to formalize a problem:

1. The agent's starting state.


2. The agent's future behavior, referring to any of the states in which the agent is currently
located.
3. The transformation model explains the purpose of each operation. Let's call it RESULT(s,
a), where s is the current state of the action, and an is the action done by the handler.
4. The target measure is used to see if the present condition is a good one.
The cost of each path is determined by the path cost equation, which is reflected in the
output metric.
5. The cost function is defined as c(s, a, s'), where s denotes the current state, and an
indicates the action taken by the agent to enter state s.'

3. Explain the relationship between graphs, semantic nets, semantic trees,


search spaces, and search trees.

ANSWER:

1. Weighted and unweighted trees and graphs are also possible.

2.In most trees, a node is designated as the root. In general, Graph does not label a node as
the root. However, if you use it for anything specific, such as traversing a graph, you should
mark it. A rooted graph has a source.

3. All graphs and trees depict relationships between data points. An edge between them
represents a connection between any two edges.

4. It is correct that all trees can be interpreted as graphs. However, trees cannot be used to
represent all graphs. A forest is a graph, but not every Graph is a tree. A loop in a graph
prevents it from being interpreted as a tree. A chart, on the other hand, can be disconnected,
while a tree cannot.

5. All graphs and trees can be searched. Only the traversal algorithms vary, but graph
searches are possible.

6. Loops cannot exist in a tree, but they do live in a graph. There must be strictly n-1 nodes
in a tree with n nodes, and the tree must be connected. It implies that trees do not have
cycles. These limitations do not exist in a graph.

4. Explain the connection between frames and object-oriented structures in


programming languages, such as Java and C++.

ANSWER:

(Not "frames" in AI) I assume you say "stack frames" or "activation frames."

For each method invocation, "stack frames" or "activation frames" are generated in
languages such as C, C++, Java, Algol, LISP, and Pascal. LIFO (Last in First Out) life cycles
apply to these frames. The last structure is replaced as a function or process returns.
(Multiple frames are removed by a global jump or catch/throw.)
The "indefinite lifetime" of an object is used to describe how long an object can last. They
are generated and survive until they are lost (in C++) or garbage collected (in garbage
collection) (Java, Smalltalk, SCHEME, Objective-C).

5. Explain the connection between decision trees and the Describe and
Match algorithm. How efficient do you think this algorithm is? Can you
think of any ways to improve it?

ANSWER:

Every day I work on data science problems handling large and huge sets of data. To get a good
knowledge about the data it must be in a good structured and sorted order. Here sorting is used
in everyday life whether we use it or not. We sort our tasks based on their importance. We get
Facebook and Twitter posts sorted from new to old. So in this way sorting is a part in our daily
life.

**Algorithm of Quick Sort:

// A is array

Read Array

def QSort(Array, f, l):

// First call should be QSort(A, 0, size(Array))

if p < r:

q = Partition(Array, f, l)

QSort(Array, f, l)

QSort(Array, q+1, l)

def Partition(Array, f, l):

// x is the pivot element

x = Array[f]

i=f-1

j=l+1

while 1:

j=j-1

while Array[j] <= x:

j=j-1
i=i+1

while Array[i] >= x:

j=j-1

// Comparing the indexes

if i < j:

swap(Array[i], Array[j])

else:

return j

PART II. ASSESSMENT (50 points)


1. As explained from the discussion above about Search Trees and Its example
which is the Tower of Hanoi.

You might also like