Indian Maritime University

(A Central University, Govt of India)

End Semester Examinations – June 2024

Programme Name: B Tech (ME)

Semester: VI

Subject Code: UG11T4601

Subject Name: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING

Date: 27.05.2024 Max Marks: 70

Duration: 03 Hrs Pass Marks: 35

General Instructions

(i) All Sections (A, B & C) are to be attempted.

(ii) Options, if any, are specified in respective section.
(iii) Scientific calculator is permitted

Section A

MCQs –All Questions are Compulsory. (10×01 mark = 10 Marks)

1. Which of the following is the branch of Artificial Intelligence?

a) Machine Learning b) Cyber forensics
c) Full-Stack Developer d) Network Design

2. In which of the following situations might a blind search be acceptable?

a) Real-life situation b) Complex game
c) Small search space d) All of the mentioned

3. Which one of the following terms is not used in the Bayes’ Theorem?
a) Prior b) Unlikelihood
c) Posterior d) Evidence

4. What kind of algorithm is logistic regression?

a) Cost function minimization b) Ranking
c) Regression d) Classification

5. If the odds in favour of an event be 5/7, find the probability of the

occurrence of the event?
a) 5/7 b) 1
c) 7/12 d) 5/12
6. What kind of neural network is most frequently applied to the image
classification?
a) RNN (Recurrent Neural Network)
b) CNN (Convolutional Neural Network)
c) FNN (Feedforward Neural Network)
d) LSTM (Long Short-Term Memory)

7. Which is true for neural networks?

a) It has set of nodes and connections
b) Each node computes it’s weighted input
c) Node could be in excited state or non-excited state
d) All of the mentioned

8. Which kind of data does Reinforcement Learning (RL) use?

a) Labelled data b) Unlabelled data
c) None d) Both

9. Which of the following is not an application of artificial intelligence?

a) Computer vision
b) Natural Language Processing
c) Data base Management System
d) Digital assistants

10. A certain Professor at the Stanford University coined the word “Artificial
Intelligence” in 1956 at a conference held at Dartmouth College. Can you
name the Professor?
a) David Levy b) John McCarthy
c) Joseph Weizenbaum d) Hans Berliner

Section B

Five Questions of 02 Marks each

11. What are various applications of AI?

12. How will you measure the problem-solving performance?

13. What is ‘Overfitting’ in machine learning and how can you avoid?

14. What is an activation function in neural networks and different types?

15. What are the advantages of heuristic function?

 Section C

Seven Questions of 10 Marks each of which any 05 questions to be

answered.

16. a) Discuss the historical evaluation of Artificial Intelligence.

b) What is the difference between Weak AI and Strong AI, and which one is
more achievable?
(7 + 3 marks)

17. Briefly discuss the following (give suitable example for each):
(i) Supervised learning
(ii) Unsupervised learning
(10 Marks)

18. a) Define conditional probability and explain Bayes rule in detail.

b) Explain in details about Logistic Regression in Machine Learning
(5 + 5 Marks)

19. a) Explain Algorithmic Bias with examples? Is it avoidable?

b) What is a filter bubble and how does it work? Give an examples.
(5 + 5 Marks)

20. a) Differentiate Blind Search and Heuristic Search.

b) Explain the Mini-Max algorithm
(5 + 5 Marks)

21. a) Explain in detail about Perceptrons and its types?

b) What is Convolutional Neural Network (CNN)?
(5 + 5 Marks)

22. How AI is changing the Maritime Industry and What Are the Advantages?
(10 Marks)
 Indian Maritime University
(A Central University, Govt of India)
End Semester Examinations – June 2024
Programme Name: B Tech (ME)
Semester: VI
Subject Code: UG11T4601
Subject Name: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING

Date: Max Marks: 7

Duration: 03 Hrs Pass Marks: 35

Answer Key
Section A

MCQs –All Questions are Compulsory. (10×01 mark = 10 Marks)

1. a) Machine Learning
2. c) small search space
3. b) Unlikelihood
4. d) Classification
5. d) 5/12
6. b) CNN (Convolutional Neural Network)
7. d) All of the mentioned
8. c) None
9. c) Data base Management System
10. b) John McCarthy

Section B
Five Questions of 02 Marks each
11. What are various applications of AI?
 Robotic vehicles
 Speech recognition
 Autonomous planning and scheduling
 Game playing
 Spam fighting
 Logistics planning
 Robotics
 Machine Translation

12. How will you measure the problem-solving performance?

 Solution:

We can evaluate an algorithm’s performance in four ways:

Completeness: Is the algorithm guaranteed to find a solution when
there is one?

Optimality: Does the strategy find the optimal solution?

Time complexity: How long does it take to find a solution?

Space complexity: How much memory is needed to perform the

search?

13. What is ‘Overfitting’ in machine learning and how can you avoid?
Solution:
In machine learning, when a statistical model describes random error
or noise instead of underlying relationship ‘overfitting’ occurs. When a
model is excessively complex, overfitting is normally observed, because
of having too many parameters with respect to the number of training
data types. The model exhibits poor performance which has been
overfit.

By using a lot of data overfitting can be avoided, overfitting happens

relatively as you have a small dataset, and you try to learn from it. But
if you have a small database and you are forced to come with a model
based on that. In such situation, you can use a technique known as
cross validation. In this method the dataset splits into two section,
testing and training datasets, the testing dataset will only test the
model while, in training dataset, the datapoints will come up with the
model. In this technique, a model is usually given a dataset of a known
data on which training (training data set) is run and a dataset of
unknown data against which the model is tested. The idea of cross
validation is to define a dataset to “test” the model in the training
phase.

14. What is an activation function in neural networks and different types?

Solution:
The activation function decides whether a neuron should be activated
or not by calculating the weighted sum and further adding bias to it.
The purpose of the activation function is to introduce non-linearity
into the output of a neuron.

Popular types of activation functions and when to use them

 Binary Step Function
 Linear Function
 Sigmoid
 Tanh
 ReLU
  Leaky ReLU
 Parameterised ReLU
 Exponential Linear Unit

15. What are the advantages of heuristic function?

Solution:

Heuristics 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 explained. This
drastically reduces the required searching time.

Section C
Seven Questions of 10 Marks each of which any 05 questions to be
answered.
16. (a) Discuss the historical evaluation of Artificial Intelligence.

Following are some milestones in the history of AI:

 The first work which is now recognized as AI was done by

Warren McCulloch and Walter pits in 1943. They proposed a model
of artificial neurons.
 Donald Hebb demonstrated in 1949an updating rule for modifying
the connection strength between neurons.
 The Alan Turing who was an English mathematician and pioneered
Machine learning in 1950.
 In 1956 the word "Artificial Intelligence" first adopted by American
Computer scientist John McCarthy.
 Joseph Weizenbaum created the first chatbot in 1966, which was
named as ELIZA.
 In 1972 the first intelligent humanoid robot was built in Japan
which was named as WABOT-1.
 The duration between years 1974 to 1980 was the first AI winter
duration.
 AI came back with "Expert System" in 1980.
 The duration second winter between the years 1987 to 1993
 In the year 1997, IBM Deep Blue beats world chess champion.
 AI entered the home in the form of Roomba, a vacuum cleaner in
2002.
 Companies like Facebook, Twitter, and Netflix also started using
AI.
 In the year 2011, IBM's Watson won jeopardy, a quiz show, where
it had to solve the complex questions as well as riddles.
  Google has launched an Android app feature "Google now"in the
year 2012
 In the year 2014, Chatbot "Eugene Goostman" won a competition
in the infamous "Turing test."
 In the year 2018 the "Project Debater" from IBM debated on
complex topics.
 Nowadays companies like Google, Facebook, IBM, and Amazon are
working with AI and creating amazing devices.

b) What is the difference between Weak AI and Strong AI, and

which one is more achievable?

Weak AI Strong AI

Limited to perform Perform intelligent human level

specific tasks activities

Have the ability to learn, think

Programmed for fixed
and perform new activities like
function
humans

It doesn’t have any

It poses creativity, common
consciousness or
sense and logic like humans.
awareness of its own.

They have a goal to

complete a task with They have a goal to solve
creative and accurate problems at a faster pace.
solutions.

There are no real examples of

Examples of weak AI strong AI because it is a
include Alexa, Siri and hypothetical theory. Some
Google Assistant. fictional examples are Wall-E and
Big Hero 6.

17. Briefly discuss the following (give suitable example for each):
(i) Supervised learning
(ii) Unsupervised learning

Solution
(i) Supervised Machine Learning:

 Supervised learning is the type of machine learning in which

machines are trained using well "labelled" training data, and on
 basis of that data, machines predict the output. The labelled data
means some input data is already tagged with the correct output.
 In the real-world, supervised learning can be used for Risk
Assessment, Image classification, Fraud Detection, spam
filtering, etc.

Steps Involved in Supervised Learning:

 First Determine the type of training dataset

 Collect/Gather the labelled training data.
 Split the training dataset into training dataset, test dataset,
and validation dataset.
 Determine the input features of the training dataset.
 Determine the suitable algorithm for the model, such as support
vector machine, decision tree, etc.
 Execute the algorithm on the training dataset and validate the
data set.
 Evaluate the accuracy of the model by providing the test set. If
the model predicts the correct output, which means our model is
accurate.

Types:

Regression: It is used for the prediction of continuous variables,

such as Weather forecasting, Market Trends, etc. Example: Linear
Regression, Regression Trees and etc.

Classification: Classification algorithms are used when the output

variable is categorical, which means there are two classes such as
Yes-No, Male-Female, True-false, etc.

Example: Random Forest, Decision Trees and Logistic Regression

(ii) Unsupervised Machine Learning

Unsupervised learning is different from the Supervised learning

technique; as its name suggests, there is no need for supervision. It
means, in unsupervised machine learning, the machine is trained using
the unlabelled dataset, and the machine predicts the output without
any supervision. In unsupervised learning, the models are trained with
the data that is neither classified nor labelled, and the model acts on
that data without any supervision.

Unsupervised Learning can be further classified into two types, which

are given below:

 Clustering
 Association
 Advantages:

 These algorithms can be used for complicated tasks compared to

the supervised ones because these algorithms work on the
unlabelled dataset.
 Unsupervised algorithms are preferable for various tasks as
getting the unlabelled dataset is easier as compared to the
labelled dataset.

Disadvantages:

 The output of an unsupervised algorithm can be less accurate as

the dataset is not labelled, and algorithms are not trained with the
exact output in prior.
 Working with Unsupervised learning is more difficult as it works
with the unlabelled dataset that does not map with the output.

Applications of unsupervised Learning: Network Analysis,

Recommendation Systems

18. a) Define conditional probability and explain Bayes rule in

detail.

Conditional probability:

Conditional probability is a probability of occurring an event when

another event has already happened.

Let's suppose, we want to calculate the event A when event B has

already occurred, "the probability of A under the conditions of B", it can
be written as:

Where P (A/ B) = Joint probability of A and B

P (B) = Marginal probability of B.

If the probability of A is given and we need to find the probability of B,

then it will be given as:

It can be explained by using the below Venn diagram, where B is

occurred event, so sample space will be reduced to set B, and now
we can only calculate event A when event B is already occurred by
 dividing the probability of P(A⋀B) by P( B ).
Bayes' theorem in Artificial intelligence

Bayes' theorem:

Bayes' theorem is also known as Bayes' rule, Bayes' law, or Bayesian

reasoning, which determines the probability of an event with uncertain
knowledge. In probability theory, it relates the conditional probability
and marginal probabilities of two random events. Bayes’ theorem was
named after the British mathematician Thomas Bayes. The Bayesian
inference is an application of Bayes’ theorem, which is fundamental to
Bayesian statistics. It is a way to calculate the value of P(B|A) with the
knowledge of P(A|B).

➢ Bayes' theorem allows updating the probability prediction of

an event by observing new information of the real world.

➢ Example: If cancer corresponds to one's age then by using

Bayes' theorem, we can determine the probability of cancer more
accurately with the help of age.

➢ Bayes' theorem can be derived using product rule and

conditional probability of event A with known event B: As from
product rule we can write:

The above equation (a) is called as Bayes' rule or Bayes' theorem.

This equation is basic of most modern AI systems for probabilistic
inference.
It shows the simple relationship between joint and conditional
probabilities. Here,

➢ P (A|B) is known as posterior, which we need to calculate, and

it will be read as Probability of hypothesis A when we have
occurred an evidence B.
➢ P (B|A) is called the likelihood, in which we consider that
hypothesis is true, then we calculate the probability of evidence.

➢ P(A) is called the prior probability, probability of

hypothesis before considering the evidence

➢ P (B) is called marginal probability, pure probability of

evidence.

In the equation (a), in general, we can write P (B) = P(A)*P(B|Ai),

hence the Bayes' rule can be written as:

Where A1, A2, A3,…… , An is a set of mutually exclusive and exhaustive

events.

(b) Explain in details about Logistic Regression in Machine

Learning

Logistic Regression in Machine Learning

 Logistic regression is one of the most popular Machine

Learning algorithms, which comes under the Supervised
Learning technique. It is used for predicting the categorical
dependent variable using a given set of independent
variables.
 Logistic regression predicts the output of a categorical
dependent variable. Therefore, the outcome must be a
categorical or discrete value. It can be either Yes or No, 0 or
1, true or False, etc. but instead of giving the exact value as
0 and 1, it gives the probabilistic values which lie
between 0 and 1.
 Logistic Regression is much similar to the Linear Regression
except that how they are used. Linear Regression is used for
solving Regression problems, whereas Logistic regression
is used for solving the classification problems.
 In Logistic regression, instead of fitting a regression line, we
fit an "S" shaped logistic function, which predicts two
maximum values (0 or 1).
 The curve from the logistic function indicates the likelihood
of something such as whether the cells are cancerous or not,
a mouse is obese or not based on its weight, etc.
 Logistic Regression is a significant machine learning
algorithm because it has the ability to provide probabilities
and classify new data using continuous and discrete
 datasets.
 Logistic Regression can be used to classify the observations
using different types of data and can easily determine the
most effective variables used for the classification.

Logistic Function (Sigmoid Function):

 The sigmoid function is a mathematical function used to map

the predicted values to probabilities.
 It maps any real value into another value within a range of 0
and 1.
 The value of the logistic regression must be between 0 and
1, which cannot go beyond this limit, so it forms a curve like
the "S" form. The S-form curve is called the Sigmoid
function or the logistic function.
 In logistic regression, we use the concept of the threshold
value, which defines the probability of either 0 or 1. Such as
values above the threshold value tends to 1, and a value
below the threshold values tends to 0.

Assumptions for Logistic Regression:

 The dependent variable must be categorical in nature.
 The independent variable should not have multi-collinearity.

Logistic Regression Equation:

The Logistic regression equation can be obtained from the Linear

Regression equation. The mathematical steps to get Logistic
Regression equations are given below:

o We know the equation of the straight line can be written as:

o In Logistic Regression y can be between 0 and 1 only, so for this

let's divide the above equation by (1-y):

o But we need range between -[infinity] to +[infinity], then take

logarithm of the equation it will become:
 The above equation is the final equation for Logistic Regression.

Type of Logistic Regression:

Logistic Regression can be classified into three types:

o Binomial: In binomial Logistic regression, there can be only

two possible types of the dependent variables, such as 0 or 1,
Pass or Fail, etc.
o Multinomial: In multinomial Logistic regression, there can be 3
or more possible unordered types of the dependent variable, such
as "cat", "dogs", or "sheep"
o Ordinal: In ordinal Logistic regression, there can be 3 or more
possible ordered types of dependent variables, such as "low",
"Medium", or "High".

19. a)Explain Algorithmic Bias with examples? Is it avoidable?

5 Marks

Algorithmic bias results in unfair outcomes due to skewed or limited

input data, unfair algorithms, or exclusionary practices during AI
development.

Algorithmic bias refers to the systemic and repeatable errors in a

computer system that create unfair outcomes, such as privileging one
arbitrary group of users over others. It's a prevalent concern today,
with artificial intelligence (AI) and machine learning (ML) applications
increasingly permeating every aspect of our lives.

Algorithmic Bias Explained

Imagine a simple decision-making tool, like a sorting hat that puts

people into different categories. But what if the hat has been only
exposed to a specific type of person while learning its task? Then it
might misjudge individuals who don't fit the 'usual' criteria, showing
bias towards those who do. That's the crux of algorithmic bias.

This bias originates from skewed or limited input data, unfair

algorithms, or exclusionary practices during AI development. It's crucial
to address this issue as AI systems are now involved in significant
domains like healthcare, finance, and criminal justice, where biased
decisions can lead to harmful effects.

Several factors contribute to algorithmic bias:

 Data bias.

 Prejudice in design.
  Socio-technical factors.

Examples of Algorithmic Bias

Real-world examples can clarify the concept of algorithmic bias:

 Hiring algorithms.

 Facial recognition systems.

Best Practices to Avoid Algorithmic Bias

Addressing algorithmic bias involves conscientious efforts at different

stages of AI system development:

 Diverse and representative data.

 Bias auditing.

 Transparency.

 Inclusive development teams.

b) What is a filter bubble and how does it work? Give an examples.

5 Marks

Internet activist Eli Pariser coined the term “filter bubble” circa 2010.
He noticed that search engines, social media sites, and other platforms
use algorithms to personalize and present the content based on your
previous activity, usually filtering out the content that offers
contrasting views or opinions.

Because of the filter bubble, you mostly see the posts reinforcing your
beliefs. Diverse content that could challenge what you believe in is just
not there. That’s why the filter bubble skews the information you
encounter online and can distort your perception of reality.

How does a filter bubble work?

Internet search engines and social media platforms usually track your
activity on their site and, with the help of third-party cookies, also off
their site. They can log:
 The posts you see and interact with.
 The time you spend on each post before moving on.
 Your location (based on GPS or your IP address).
 Your web searches and browsing history.
  Websites you visit before coming to their site.
 Websites you visit after leaving their site.
 Anything and everything else you do on their site.

Platforms use this information to deduce many things about you – from
your age and gender to your relationship status and hobbies. Add it to
the information you provide to the companies willingly by posting,
commenting, or simply registering for their service.

Examples of filter bubbles:

Web search results and social media feeds are the most common
examples of online filter bubbles. But filter bubbles now exist beyond
these platforms. Algorithms dictate suggested movies and series on
streaming sites, songs on Spotify, videos on YouTube, and even what
content you see first on some news sites.

20. a)Differentiate Blind Search and Heuristic Search. 5 Marks

Parameters Blind search Heuristic search

It is also known It is also known Informed
Also known
Uninformed Search Search
as

Using It doesn’t use knowledge It uses knowledge for the

for the searching searching process.
Knowledge
process.
Performance It finds solution slow as It finds a solution more
compared to an informed quickly.
search.
Completion It is always complete. It may or may not be
complete.
Cost Factor Cost is high. Cost is low.
It consumes moderate It consumes less time
Time
time because of slow because of quick
searching. searching.
No suggestion is given There is a direction given
Direction
regarding the solution in about the solution.
it.
It is lengthier while It is less lengthy while
Implementati
implemented. implemented.
on

It is comparatively less It is more efficient as

efficient as incurred cost efficiency takes into
is more and the speed of account cost and
Efficiency finding the Breadth-First performance. The incurred
solution is slow. cost is less and speed of
 finding solutions is quick.
Computation Comparatively higher Computational
al computational requirements are
requirements requirements. lessened.
Size of Solving a massive Having a wide scope in
search search task is terms of handling large
problems challenging. search problems.
Example Example
a) Breadth first search a) Best first search

Examples of b) Uniform cost search b) Greedy search

Algorithms
c) Depth first Search c) A* search
d) Depth limited search d) AO* Search
e) Iterative deepening e) Hill Climbing
search Algorithm
f) Bi – Directional Search

b) Explain the Mini-Max algorithm 5 Marks

➢ Mini-max algorithm is a recursive or backtracking algorithm which

is used in decision- making and game theory. It provides an
optimal move for the player assuming that opponent is also
playing optimally.
➢ Mini-Max algorithm uses recursion to search through the game-
tree.
➢ Min-Max algorithm is mostly used for game playing in AI. Such as
chess, Checkers, tic-tac- toe, go, and various tow-players game.
➢ This Algorithm computes the minimax decision for the current
state.
➢ In this algorithm two players play the game, one is called MAX
and other is called MIN.
➢ Both the players fight it as the opponent player gets the minimum
benefit while they get the maximum benefit.
➢ Both Players of the game are opponent of each other, where MAX
will select the maximized value and MIN will select the minimized
value.
➢ The minimax algorithm performs a depth-first search algorithm for
the exploration of the complete game tree.
➢ The minimax algorithm proceeds all the way down to the terminal
node of the tree, then backtrack the tree as the recursion.

Working of Min-Max Algorithm:

 The working of the minimax algorithm can be easily described
using an example. Below we have taken an example of game-tree
which is representing the two-player game.
  In this example, there are two players one is called Maximizer and
other is called Minimizer.
 Maximizer will try to get the Maximum possible score, and
Minimizer will try to get the minimum possible score.
 This algorithm applies DFS, so in this game-tree, we have to go all
the way through the leaves to reach the terminal nodes.
 At the terminal node, the terminal values are given so we will
compare those value and backtrack the tree until the initial state
occurs.

21. a) Explain in detail about Perceptrons and its types?

5 Marks

Perceptron is Machine Learning algorithm for supervised learning of

various binary classification tasks. Further, Perceptron is also
understood as an Artificial Neuron or neural network unit that helps to
detect certain input data computations in business intelligence.
Perceptron model is also treated as one of the best and simplest types
of Artificial Neural networks. However, it is a supervised learning
algorithm of binary classifiers. Hence, we can consider it as a single-
layer neural network with four main parameters, i.e., input values,
weights and Bias, net sum, and an activation function.

Basic Components of Perceptron

Mr. Frank Rosenblatt invented the perceptron model as a binary

classifier which contains three main components. These are as follows:
 Input Nodes or Input Layer:
This is the primary component of Perceptron which accepts the
initial data into the system for further processing. Each input node
contains a real numerical value.

Weight and Bias:

Weight parameter represents the strength of the connection
between units. This is another most important parameter of
Perceptron components. Weight is directly proportional to the
strength of the associated input neuron in deciding the output.
Further, Bias can be considered as the line of intercept in a linear
equation.

Activation Function:
These are the final and important components that help to
determine whether the neuron will fire or not. Activation Function
can be considered primarily as a step function.

Types of Activation functions:

 Sign function
 Step function, and
 Sigmoid function

Types of Perceptron Models

Based on the layers, Perceptron models are divided into two types.
These are as follows:
1. Single-layer Perceptron Model
2. Multi-layer Perceptron model

b) What is Convolutional Neural Network(CNN)? 5 Marks

A Convolutional Neural Network (CNN) is a type of deep learning

algorithm that is particularly well-suited for image recognition and
processing tasks. It is made up of multiple layers, including
convolutional layers, pooling layers, and fully connected layers. The
architecture of CNNs is inspired by the visual processing in the
human brain, and they are well-suited for capturing hierarchical
patterns and spatial dependencies within images.

Key components of a Convolutional Neural Network include:

 Convolutional Layers: These layers apply convolutional o
perations to input images, using filters (also known as kernels)
to detect features such as edges, textures, and more complex
patterns. Convolutional operations help preserve the spatial
relationships between pixels.
 Pooling Layers: Pooling layers downsample the spatial
dimensions of the input, reducing the computational complexity
and the number of parameters in the network. Max pooling is a
 common pooling operation, selecting the maximum value from
a group of neighbouring pixels.
 Activation Functions: Non-linear activation functions, such as
Rectified Linear Unit (ReLU), introduce non-linearity to the
model, allowing it to learn more complex relationships in the
data.
 Fully Connected Layers: These layers are responsible for
making predictions based on the high-level features learned by
the previous layers. They connect every neuron in one layer to
every neuron in the next layer.
CNNs are trained using a large dataset of labelled images, where the
network learns to recognize patterns and features that are
associated with specific objects or classes. Proven to be highly
effective in image-related tasks, achieving state-of-the-art
performance in various computer vision applications. Their ability to
automatically learn hierarchical representations of features makes
them well-suited for tasks where the spatial relationships and
patterns in the data are crucial for accurate predictions. CNNs are
widely used in areas such as image classification, object detection,
facial recognition, and medical image analysis.
The convolutional layers are the key component of a CNN, where
filters are applied to the input image to extract features such as
edges, textures, and shapes.
The output of the convolutional layers is then passed through
pooling layers, which are used to down-sample the feature maps,
reducing the spatial dimensions while retaining the most important
information. The output of the pooling layers is then passed through
one or more fully connected layers, which are used to make a
prediction or classify the image

22. How AI is changing the Maritime Industry and What Are the
Advantages? 10 Marks

Like many sectors, the maritime industry is embracing modern

technology to advance and streamline operations. Individuals
working and studying in this field should understand how
implementing new technology changes operations for better
preparation and adoption.

1. Navigation and Route Optimization

AI equipment studies past behaviors and learns from them to
apply actions itself. By reviewing routes and navigational
patterns, AI can navigate vessels without human intervention,
opening the possibility for autonomous ships or those with
minimal crew members. These operations allow Deckhands to
 focus on other essential tasks on the ship, ensuring safe cargo
delivery and increased efficiency.

Information like fuel use and water traffic can offer many
applications for vessel efficiency and streamlined operations.
Route optimization can help crews reach their destination
quicker and optimize time on the water. AI can assist in
planning and operating, using details about local weather
patterns and forecasts, traffic patterns and fuel use to find the
quickest and safest routes for crews.

2. Fuel Consumption:
Sailing and internal operations require fuel and energy sources
to work. AI and machine learning can track how much fuel your
vessels consume and suggest strategies to reduce use for
better resource and cost management. Fuel consumption can
also give insights into inefficient processes, helping drive
improvements across ship operations.

3. Equipment and Vessel Maintenance:

Maintenance is an essential aspect of ship efficiency and safety.
When equipment functions at its best, crews can experience
better time and fuel efficiency. Even smaller damages and wear
can impact entire systems, causing them to overwork and use
more energy and fuel, increasing vessel costs and
consumption. Damage that goes unnoticed can become a
liability for passages, causing potential dangers.

AI uses sensors to analyze the performance of your equipment

to determine where there might be issues. When it detects
uncharacteristic changes in fuel or energy use and heat
production, it can alert crews that the machine needs
inspection and maintenance. Because crew members have
many responsibilities, AI equipment monitoring can provide
better upkeep equipment by increasing visibility and
awareness. Equipment can receive the attention it needs
sooner rather than waiting until the next maintenance, service
or inspection cycle.

4. Port Density and Traffic:

AI analyzes information from many sources, including your
radar, sonar and GPS scanners, to provide more specified
results to your area. When docking at ports and traveling the
waters, it records other vessels and stores it with your
historical data. As it gains an understanding of regular
patterns, it can help vessels navigate water safely by
encountering fewer ships. This data benefits ships in multiple
 ways. You can complete trips more efficiently by knowing you
can dock sooner. This data can also promote greater safety on
the water. Even with all available equipment, collisions and
accidents are frequent. Analyzing and using traffic pattern data
can help your crew avoid heavily trafficked areas and reduce
collision risks.

The Advantages of AI:

 Reducing costs
 Improving efficiency
 Supporting sustainability
 Eliminating human error
 Indian Maritime University

(A Central University, Govt of India)

Supplementary Examinations – September/October 2024

Programme Name: B Tech (ME)

Semester: VI

Subject Code: UG11T4601

Subject Name: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING

Date: 24.10.2024 Max Marks: 70

Duration: 03 Hrs Pass Marks: 35

General Instructions

(i) All Sections (A, B & C) are to be attempted.

(ii) Options, if any, are specified in respective section.
(iii) Scientific calculator is permitted

Section A

Ten MCQs/Fill in the Blanks of 01 Mark each – Choose the correct answer as
applicable.

1. What is Artificial Intelligence?

(a) Artificial Intelligence is a field that aims to make humans more
intelligent.
(b) Artificial Intelligence is a field that aims to improve the security.
(c) Artificial Intelligence is a field that aims to develop intelligent machines.
(d) Artificial Intelligence is a field that aims to mine the data.

2. Which of the following is the branch of Artificial Intelligence?

(a) Machine Learning (b) Cyber forensics
(c) Full-Stack Develop (d) Network Design

3. Who is known as the Father of AI"?

(a) Fisher Ada (b) Alan Turing
(c) John McCarthy (d) Allen Newell

4. Among the given options, which search algorithm requires less memory?
(a) Optimal Search (b) Depth First Search
(c) Breadth First Search (d) Linear Search

5. Which algorithm is used in the Game tree to make decisions of Win/Lose?

(a) Heuristic Search Algorithm (b) DFS/BFS Algorithm
(c) Greedy Search Algorithm (d) Min/Max Algorithm
6. What is a perceptron in ANN?
(a) A double layer auto-associative neural network
(b) A neural network that contains feedback
(c) An auto-associative neural network
(d) A single layer feed-forward neural network with pre-processing

7. Choose the options below that do not fall within the category of Machine
Learning:
(a) Semi-unsupervised learning (b) Supervised learning
(c) Reinforcement learning (d) Unsupervised learning

8. Which of the below models is a generative model used in machine learning?

(a) Support vector machines (b) Naïve Bayes
(c) Logistic Regression (d) Linear Regression

9. What is the key difference between supervised and unsupervised learning?

(a) Supervised learning requires labelled data, while unsupervised
learning does not.
(b) Supervised learning predicts labels, while unsupervised learning
discovers patterns.
(c) Supervised learning is used for classification, while unsupervised
learning is used for regression.
(d) Supervised learning is always more accurate than unsupervised
learning.

10. What is the formula for Bayes’ theorem? Where (A & B) and (H & E) are events
and P(B), P(H) & P(E) ≠ 0.
(a) P(H|E) = [P(E|H) * P(E)] / P(H) (b) P(A|B) = [P(A|B) * P(A)] / P(B)
(c) P(H|E) = [P(H|E) * P(H)] / P(E) (d) P(A|B) = [P(B|A) * P(A)] / P(B)

Section B

Five Questions of 02 Marks each

11. Define the terms Weak and Strong AI with examples.

12. Give the difference between Machine Learning and Deep Learning.
13. What is ‘Overfitting’ in machine learning and how can you avoid?
14. What is an activation function in neural networks and different types.
15. What is Turing Test?

Section C

Seven Questions of 10 Marks each of which any 05 questions to be answered.

16. Discuss in details the historical evaluation of Artificial Intelligence. (10 Marks)
17. (a) Using Bayes’ theorem, describe the following terms: Likelihood Ratio, Prior
 Probability, Posterior probability and Marginal Probability. (4 Marks)
(b) What is the probability that person has disease dengue with neck pain?
Given: 80% of time dengue causes neck pain, P (dengue) = 1/30,000
and P (neck pain) =0.02. (6 Marks)

18. (a) Explain the Mini-Max algorithm. (5 Marks)

(b) Apply the minmax search algorithm and find the most convenient path for
the Max. node. (5 Marks)

19. (a) Differentiate between Supervised Learning and Unsupervised Learning.

Also, give examples of each. (5 Marks)
(b) What is a filter bubble and how does it work? Give examples.
(5 Marks)

20. Explain the informed and Uniformed search techniques. Also, write down the
differences between them. (10 Marks)

21. (a) Explain in detail about Perceptron and its types? (5 Marks)
(b) What is Convolutional Neural Network (CNN)? (5 Marks)

22. Discuss the applications of Artificial Intelligence in Maritime Sector in details.

(10 Marks)
 Indian Maritime University

(A Central University, Govt of India)

Supplementary Examinations – Sep/Oct 2024

Programme Name: B.Tech (M.E.)

Semester: VI

Subject Code: UG11T4601

Subject Name: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING

Date: Max Marks: 70

Duration: 03 Hrs Pass Marks: 35

Answer Key

Section A

Ten MCQs/Fill in the Blanks of 01 Mark each – Choose the correct answer as
applicable.

1.(c) Artificial Intelligence is a field that aims to develop intelligent machines.

2.(a) Machine Learning
3.(c) John McCarthy
4.(b) Depth First Search
5.(d) Min/Max Algorithm
6.(d) A single layer feed-forward neural network with pre-processing
7.(a) Semi-unsupervised learning
8.(b) Naïve Bayes
9.(a) Supervised learning requires labelled data, while unsupervised learning
does not.
10. (d) P(A|B) = [P(B|A) * P(A)] / P(B)

Section B

Five Questions of 02 Marks each

11. Define the terms Weak and Strong AI with examples.

Ans. Weak artificial intelligence (AI) —also called narrow AI—is a type
of artificial intelligence that is limited to a specific or narrow area. Weak AI
simulates human sensation. It has the potential to benefit society by
automating time-consuming tasks and by analysing data in ways that humans
sometimes can’t.

Characteristics of Weak AI: Limited range of functions, not self-aware, no

consciousness, Ability to learn and adapt within its programmed domain
Examples of Weak AI: Virtual assistants such as Siri and Alexa,
Recommendation engines used by Netflix and Amazon, Fraud detection software
used by financial institutions

Strong artificial intelligence (AI), also known as artificial general intelligence

(AGI) or general AI, is a theoretical form of AI used to describe a certain
mindset of AI development. If researchers are able to develop Strong AI, the
machine would require an intelligence equal to humans; it would have a self-
aware consciousness that has the ability to solve problems, learn, and plan for
the future. Strong AI aims to create intelligent machines that are
indistinguishable from the human mind.

Characteristics of Strong AI: The ability to reason and display common

sense, Self-awareness, Consciousness, Ability to learn and adapt
Examples of Strong AI: Self-driving cars, Pattern and image recognition,
Contextual recommendations

12. Give the difference between Machine Learning and Deep Learning.

Ans. Machine Learning vs. Deep Learning

Since deep learning and machine learning tend to be used interchangeably, it’s
worth noting the difference between the two. Machine learning and deep
learning, are sub-fields of artificial intelligence.

The way in which deep learning and machine learning differ is in how each
algorithm learns. "Deep" machine learning can use labelled datasets, also known
as supervised learning, to inform its algorithm, but it doesn’t necessarily require
a labelled dataset. Deep learning can ingest unstructured data in its raw form
(e.g., text or images), and it can automatically determine the set of features
which distinguish different categories of data from one another. This eliminates
some of the human intervention required and enables the use of larger data
sets.

Classical, or "non-deep", machine learning is more dependent on human

intervention to learn. Human experts determine the set of features to
understand the differences between data inputs, usually requiring more
structured data to learn.

13. What is ‘Overfitting’ in machine learning and how can you avoid?

Ans. In machine learning, when a statistical model describes random error or

noise instead of underlying relationship ‘overfitting’ occurs. When a model is
excessively complex, overfitting is normally observed, because of having too
many parameters with respect to the number of training data types. The model
exhibits poor performance which has been overfit.
By using a lot of data overfitting can be avoided, overfitting happens relatively
as you have a small dataset, and you try to learn from it. But if you have a small
database and you are forced to come with a model based on that. In such
situation, you can use a technique known as cross validation. In this method the
dataset splits into two section, testing and training datasets, the testing dataset
will only test the model while, in training dataset, the datapoints will come up
with the model. In this technique, a model is usually given a dataset of a known
data on which training (training data set) is run and a dataset of unknown data
against which the model is tested. The idea of cross validation is to define a
dataset to “test” the model in the training phase.

14. What is an activation function in neural networks and different

types.

Ans. The activation function decides whether a neuron should be activated or not
by calculating the weighted sum and further adding bias to it. The purpose of the
activation function is to introduce non-linearity into the output of a neuron.
Popular types of activation functions and when to use them:

 Binary Step Function

 Linear Function
 Sigmoid
 Tanh
 ReLU
 Leaky ReLU
 Parameterised ReLU
 Exponential Linear Unit

15. What is Turing Test?

Ans. The Turing Test is a deceptively simple method of determining whether a

machine can demonstrate human intelligence: If a machine can engage in a
conversation with a human without being detected as a machine, it has
demonstrated human intelligence.

The Turing Test was proposed in a paper published in 1950 by mathematician

and computing pioneer Alan Turing. It has become a fundamental motivator in
the theory and development of artificial Intelligence (AI).

The test is conducted in an interrogation room run by a judge. The test

subjects, a person and a computer program, are hidden from view. The judge
has a conversation with both parties and attempts to identify which is the
human and which is the computer, based on the quality of their conversation.
Turing concludes that if the judge can't tell the difference, the computer has
succeeded in demonstrating human intelligence. That is, it can think.

Section C

Seven Questions of 10 Marks each of which any 05 questions to be answered.

16. Discuss in details the historical evaluation of Artificial Intelligence.

(10 Marks)
Ans. The history of artificial intelligence is categorised as:

1. Groundwork for AI
2. Birth of AI
3. AI Maturation
4. AI Boom
5. AI Winter
6. AI Agents
7. Artificial General Intelligence
Groundwork for AI: 1900-1950: In the early 1900s, there was a lot of media
created that cantered around the idea of artificial humans. So much so that
scientists of all sorts started asking the question: is it possible to create an
artificial brain? Some creators even made some versions of what we now call
“robots” (and the word was coined in a Czech play in 1921) though most of them
were relatively simple. These were steam-powered for the most part, and some
could make facial expressions and even walk.
Birth of AI: 1950-1956: This range of time was when the interest in AI really
came to a head. Alan Turing published his work “Computer Machinery and
Intelligence” which eventually became The Turing Test, which experts used to
measure computer intelligence. The term “artificial intelligence” was coined and
came into popular use.
AI maturation: 1957-1979: The time between when the phrase “artificial
intelligence” was created, and the 1980s was a period of both rapid growth and
struggle for AI research. The late 1950s through the 1960s was a time of
creation. From programming languages that are still in use to this day to books
and films that explored the idea of robots, AI became a mainstream idea quickly.

The 1970s showed similar improvements, such as the first anthropomorphic

robot being built in Japan, to the first example of an autonomous vehicle being
built by an engineering grad student. However, it was also a time of struggle for
AI research, as the U.S. government showed little interest in continuing to fund
AI research.
AI boom: 1980-1987: Most of the 1980s showed a period of rapid growth and
interest in AI, now labelled as the “AI boom.” This came from both
breakthroughs in research, and additional government funding to support the
researchers. Deep Learning techniques and the use of Expert System became
more popular, both of which allowed computers to learn from their mistakes and
make independent decisions.
AI winter: 1987-1993: As the AAAI warned, an AI Winter came. The term
describes a period of low consumer, public, and private interest in AI which leads
to decreased research funding, which, in turn, leads to few breakthroughs. Both
private investors and the government lost interest in AI and halted their funding
due to high cost versus seemingly low return. This AI Winter came about
because of some setbacks in the machine market and expert systems, including
the end of the Fifth-Generation project, cutbacks in strategic computing
initiatives, and a slowdown in the deployment of expert systems.
AI agents: 1993-2011: Despite the lack of funding during the AI Winter, the
early 90s showed some impressive strides forward in AI research, including the
introduction of the first AI system that could beat a reigning world champion
chess player. This era also introduced AI into everyday life via innovations such
as the first Roomba and the first commercially-available speech recognition
software on Windows computers.
Artificial General Intelligence: 2012-present: That brings us to the most
recent developments in AI, up to the present day. We’ve seen a surge in
common-use AI tools, such as virtual assistants, search engines, etc. This time
period also popularized Deep Learning and Big Data.

17. (a) Using Bayes’ theorem, describe the following terms: Likelihood
Ratio, Prior Probability, Posterior probability and Marginal Probability.
(4 Marks)

P (B∨ A)P ( A )
Ans. P(A|B) =
P(B)

The above equation is called as Bayes' rule or Bayes' theorem. This equation
is basic of most modern AI systems for probabilistic inference.

It shows the simple relationship between joint and conditional probabilities.

Here, P(B|A) is called the likelihood, in which we consider that hypothesis is

true, then we calculate the probability of evidence.

P(A) is called the Prior probability, probability of hypothesis before

considering the evidence. It is probability of an event which is computed before
observing new information.

P(A|B) is known as Posterior, which we need to calculate, and it will be read

as Probability of hypothesis A when we have occurred an evidence B. This
probability is calculated after all evidence or information has considered. It is a
combination of prior probability and new information.

P(B) is called marginal probability, pure probability of an evidence.

(b) What is the probability that person has disease dengue with neck
pain? Given: 80% of time dengue causes neck pain, P (dengue) =
1/30,000 and P (neck pain) =0.02. (6 Marks)

Ans. Let A be the proposition that person has neck and B be the proposition that
person has dengue. Therefore,

P(A|B) = 80% = 0.8

P(B) = 1/30000
P(A)= 0.02
 P ( A∨B) P (B)
P(B|A) = = (0.8 x 1/30,000)/0.2 = 0.00133
P( A)

18. (a) Explain the Mini-Max algorithm. (5 Marks)

Ans. Mini-max algorithm is a recursive or backtracking algorithm which is used in
decision-making and game theory. It provides an optimal move for the player
assuming that opponent is also playing optimally.
 Mini-Max algorithm uses recursion to search through the game-tree.
 Min-Max algorithm is mostly used for game playing in AI. Such as Chess,
Checkers, tic-tac-toe, go, and various two-players game. This Algorithm
computes the minimax decision for the current state.
 In this algorithm two players play the game, one is called MAX and other is
called MIN.
 Both the players fight it as the opponent player gets the minimum benefit
while they get the maximum benefit.
 Both Players of the game are opponent of each other, where MAX will select
the maximized value and MIN will select the minimized value.
 The minimax algorithm performs a depth-first search algorithm for the
exploration of the complete game tree.
 The minimax algorithm proceeds all the way down to the terminal node of the
tree, then backtrack the tree as the recursion.

(b) Apply the minmax search algorithm and find the most convenient
path for the Max. node. (5 Marks)

Ans.
Convenient Path of Max. node: A-B-D

19. (a) Differentiate between Supervised Learning and Unsupervised

Learning. Also, give examples of each. (5 Marks)

Ans. Supervised machine learning: Supervised learning, also known as

supervised machine learning, is defined by its use of labelled datasets to train
algorithms to classify data or predict outcomes accurately. As input data is fed
into the model, the model adjusts its weights until it has been fitted
appropriately. This occurs as part of the cross validation process to ensure that
the model avoids overfitting or underfitting. Supervised learning helps
organizations solve a variety of real-world problems at scale, such as classifying
spam in a separate folder from your inbox. Some methods used in supervised
learning include neural networks, naïve Bayes, linear regression,
logistic regression, random forest, and support vector machine (SVM).
In the real-world, supervised learning can be used for Risk Assessment,
Image classification, Fraud Detection, spam filtering, etc.
Example: Random Forest, Decision Trees and Logistic Regression

Unsupervised machine learning: Unsupervised learning, also known as

unsupervised machine learning, uses machine learning algorithms to analyse
and cluster unlabelled datasets. These algorithms discover hidden patterns or
data groupings without the need for human intervention. This method’s ability to
discover similarities and differences in information make it ideal for exploratory
data analysis, cross-selling strategies, customer segmentation, and image and
pattern recognition. It’s also used to reduce the number of features in a model
through the process of dimensionality reduction. Principal component analysis
(PCA) and singular value decomposition (SVD) are two common approaches for
this. Other algorithms used in unsupervised learning include neural networks, k-
means clustering, and probabilistic clustering methods.
Applications of unsupervised Learning: Network Analysis, Recommendation
Systems

(b) What is a filter bubble and how does it work? Give examples.

(5 Marks)

Ans. A filter bubble is an algorithmic bias that skews or limits the information an
individual user sees on the internet. The bias is caused by the
weighted algorithms that search engines, social media sites and marketers use
to personalize user experience
Internet search engines and social media platforms usually track your activity on
their site and, with the help of third-party cookies, also off their site. They can
log:
 The posts you see and interact with.
 The time you spend on each post before moving on.
 Your location (based on GPS or your IP address).
 Your web searches and browsing history.
 Websites you visit before coming to their site.
 Websites you visit after leaving their site.
 Anything and everything else you do on their site.
Platforms use this information to deduce many things about you – from your age
and gender to your relationship status and hobbies. Add it to the information you
provide to the companies willingly by posting, commenting, or simply registering
for their service.

20. Explain the Informed (Heuristic) and Uniformed search techniques. Also,
write down the differences between them. (10
Marks)
Ans. Informed Search in AI: The algorithms of an informed search contain
information regarding the goal state. It helps an AI make more efficient and
accurate searches. A function obtains this data/info to estimate the closeness of
a state to its goal in the system. For example, Graph Search and Greedy Search.
Features of Informed Search in AI:
 It consists of information regarding the goal state.
 It makes a search more efficient.
 A function obtains the data/info regarding the closeness of the current
state of a search to its goal state.
 It utilises knowledge for implementing the searching process.
 A few examples include graph search and greedy search.
 It incurs less cost.
 It may be complete or incomplete.
 A solution can be found much quicker.
 This type of search consumes less time.
 Implementation of such an AI is short and quick, not at all lengthy.
 The AI gets a direct suggestion about the solution of the search/ problem.

Uninformed Search in AI: The algorithms of a uniformed AI do not consist of

any additional data/ info regarding the goal node. It only contains the
information provided during the definition of a problem. The plans required to
reach from the start state to the goal state differ only on the basis of the length
and order of the provided actions. For example, Breadth-First Search and Depth-
First Search.
Features of Uninformed Search in AI:
 It does not contain any additional data/ info.
 The information is provided to the AI during the definition of a problem.
 It can reach the goal state on the basis of the length and the order of
actions performed.
 The AI doesn’t utilise any knowledge to search for the solution to a
problem.
 A few examples of these include BFS (Breadth-First Search) and DFS
(Depth-First Search).
  This type of AI takes more time to generate a solution for any problem.
 It is always complete.
 The total cost incurred is generally more than that of Informed Search in
AI.
 It consumes a fairly moderate time to do the search.
 The implementation is lengthy.
 No suggestion is present for finding any solution.

Parameters Informed Search Uninformed Search

Utilizing It uses knowledge during the It does not require using any knowledge
Knowledge process of searching. during the process of searching.

Speed Finding the solution is Finding the solution is much slower

quicker. comparatively.

Completion It can be both complete and It is always bound to be complete.

incomplete.

Consumption of Due to a quicker search, it Due to slow searches, it consumes

Time consumes much less time. comparatively more time.

Cost Incurred The expenses are much The expenses are comparatively higher.
lower.

Suggestion/ The AI gets suggestions The AI does not get any suggestions
Direction regarding how and where to regarding what solution to find and where
find a solution to any to find it. Whatever knowledge it gets is
problem. out of the information provided.

Efficiency It costs less and generates It costs more and generates slower results.
quicker results. Thus, it is Thus, it is comparatively less efficient.
comparatively more efficient.

Length of Implementation is shorter The implementation is lengthier using AI.

Implementation using AI.

Examples A few examples include A few examples include Breadth-First

Graph Search and Greedy Search or BFS and Depth-First Search or
Search. DFS.

21. (a) Explain in detail about Perceptron and its types? (5 Marks)

Ans. Perceptron is Machine Learning algorithm for supervised learning of various

binary classification tasks. Further, Perceptron is also understood as an Artificial
Neuron or neural network unit that helps to detect certain input data
computations in business intelligence.
Perceptron model is also treated as one of the best and simplest types of
Artificial Neural networks. However, it is a supervised learning algorithm of
binary classifiers. Hence, we can consider it as a single layer neural network with
four main parameters, i.e., input values, weights and Bias, net sum, and an
activation function.

Basic Components of Perceptron

Input Nodes or Input Layer: This is the primary component of Perceptron
which accepts the initial data into the system for further processing. Each input
node contains a real numerical value.
Weight and Bias: Weight parameter represents the strength of the connection
between units. This is another most important parameter of Perceptron
components. Weight is directly proportional to the strength of the associated
input neuron in deciding the output. Further, Bias can be considered as the line
of intercept in a linear equation.
Activation Function:
These are the final and important components that help to
determine whether the neuron will fire or not. Activation Function
can be considered primarily as a step function.
Types of Activation functions:
 Sign function
 Step function
 Sigmoid function
Types of Perceptron Models: Based on the layers, Perceptron models are
divided into two types. These are as follows:
 Single-layer Perceptron Model
 Multi-layer Perceptron model

(b) What is Convolutional Neural Network (CNN)? (5 Marks)

Ans. A Convolutional Neural Network (CNN) is a specialized artificial neural

network designed for image processing. It employs convolutional layers to
automatically learn hierarchical features from input images, enabling effective
image recognition and classification. CNNs have revolutionized computer vision
and are pivotal in tasks like object detection and image analysis. Key
components of a Convolutional Neural Network include:

Convolutional Layers: These layers apply convolutional operations to input

images, using filters (also known as kernels) to detect features such as edges,
textures, and more complex patterns. Convolutional operations help preserve
the spatial relationships between pixels.

Pooling Layers: Pooling layers downsample the spatial dimensions of the input,
reducing the computational complexity and the number of parameters in the
network. Max pooling is a common pooling operation, selecting the maximum
value from a group of neighbouring pixels.

Activation Functions: Non-linear activation functions, such as Rectified Linear

Unit (ReLU), introduce non-linearity to the model, allowing it to learn more
complex relationships in the data.
Fully Connected Layers: These layers are responsible for making predictions
based on the high-level features learned by the previous layers. They connect
every neuron in one layer to every neuron in the next layer.

22. Discuss the applications of Artificial Intelligence in Maritime Sector

in details. (10 Marks)
Ans. Like many sectors, the maritime industry is embracing modern technology
to advance and streamline operations. Individuals working and studying in this
field should understand how implementing new technology changes operations
for better preparation and adoption.
1. Navigation and Route Optimization
2. Fuel Consumption:
3. Equipment and Vessel Maintenance:
4. Port Density and Traffic:
The Advantages of AI:
 Reducing costs
 Improving efficiency
 Supporting sustainability
 Eliminating human error