Scribd
Upload
40 views5 pages

Ai Fall-23 Assignment

This document provides instructions for an assignment on artificial intelligence for students in the Department of Computing and Information Systems at Daffodil International University. It outlines the tasks to be completed, including a literature review on AI topics, answering theoretical questions on machine learning algorithms, coding a solution to a provided dataset using an ML algorithm on Google Colab, and producing visualizations and analysis of the model. Students must submit their assignment in Google Classroom and the LMS by November 8th, 2023 for the fall semester course on AI and participate in a final in-person viva presentation and submission of a hard copy for evaluation.

Uploaded by

xekiwi1002
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
40 views5 pages

Ai Fall-23 Assignment

This document provides instructions for an assignment on artificial intelligence for students in the Department of Computing and Information Systems at Daffodil International University. It outlines the tasks to be completed, including a literature review on AI topics, answering theoretical questions on machine learning algorithms, coding a solution to a provided dataset using an ML algorithm on Google Colab, and producing visualizations and analysis of the model. Students must submit their assignment in Google Classroom and the LMS by November 8th, 2023 for the fall semester course on AI and participate in a final in-person viva presentation and submission of a hard copy for evaluation.

Uploaded by

xekiwi1002
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Daffodil International University

Department of Computing and Information System


Faculty of Science & Information Technology
Semester Assignment, Fall 2023
Course Code: CIS (313 & 412) & CIS (313L & 412L)
Course Title: Artificial Intelligence and Lab

Last Submission Date: Nov 08th 2023. (08-11-23)


[The figures in the right margin indicate the full marks and corresponding course Learning outcomes.]

TITLE: CSAIL LAB, MIT


Alexander Amini is a Post-Doctoral student in the CSAIL lab at MIT.
He has been doing research on Autonomous car driving at MIT and
made a simulation model for an Autonomous car driving environment.
One day Alexander launched the autonomous car for test driving in the
road. He examined the AI agent that his team built on that car that
works perfectly. It can identify pedestrians and other vehicles. It also
identifies weather conditions. But suddenly after driving the test car,
the car was hit by a sandbag, because it didn’t identify the object which
type it is. The next day Amini called a meeting in the CSAIL lab and
told everyone the incident happened yesterday. Many researchers gave
their opinion on that incident. One of them talked about training and
testing imbalanced dataset problems, where someone said that the bias-
variance problem of the AI agent, and one of them said about the
limitations of AI. Today, Alexander and his team finally made an
autonomous driving car system that is really perfect like humans.

So, CSAIL lab is opening a Research Assistant position for the fresher
graduate, but you have to participate in an online contest where you
have to solve a dataset problem provided by the CSAIL lab. Now it’s
your time to prove your brilliance and get the job position.
DATASET

Dataset Link:
https://drive.google.com/file/d/1IYSwF4pTryfSvn5pD0TgFc
JYSH4HiIiu/view?usp=sharing

Dataset Description:
https://drive.google.com/file/d/1QiwmWYZsb7nfRO0Qut_l
VcRTBYC1v6IK/view?usp=sharing

THEORY PART (35 marks)

Task-1 (5 marks)
(a) Class Presentation on trending topics on Artificial Intelligence, Machine [5]
Learning and Deep Learning.

Task-2 (15 marks)

(a) What do you understand by the concept of “AI Singularity”? Do you [3]
think all the LLM models are an invention of technology singularity?

Linear Regression:

(a) Why can’t we use Linear Regression algorithm for the Classification [4]
problem?
(b) What do you understand by the concept of Elastic Net? [4]
(c) If we use Polynomial Regression algorithm, then how can we [4]
overcome Non-Convex optimization problem?

Logistic Regression:

(a) How Logistic Regression works on Multi Class Classification? [4]


(b) Compare between Logistic Regression and SVM models and give your [4]
opinion which is better under different purposes and condition?
(c) Is there any other cost function can we use in the Logistic Regression [4]
model instead of Sigmoid function?
Support Vector Machine (Classification):
(a) In Which field SVM work better than other existing machine learning [4]
models?
(b) How does SVM works on Non-Linear problem? [4]
(c) How does different Kernels works for SVM algorithm? [4]

Support Vector Machine (Regression):


(a) How SVM works on Regression problem? [4]
(b) How does different Kernels works for SVM algorithm? [4]
(c) Can SVM works on Regression problem without Normalization? [4]

KNN:
(a) What will happen if we take Manhattan Distance or Minkowski [4]
Distance instead of Euclidean Distance on the KNN algorithm?
(b) Why KNN is so much popular now a days? Give your valuable opinion [4]
with proper justification.
(c) How KNN does works on Non-Linear dataset and Imbalanced dataset? [4]

KNN (Regression):
(a) Why KNN is so much popular now a days? Give your valuable opinion [4]
with proper justification.
(b) How does KNN works on Regression problem? [4]
(c) Does KNN Regression better than Linear Regression? If so then why? [4]

Naïve Bayes:
(a) How to solve the Zero Frequency problem in the Naïve Bayes? [4]
(b) Can we use Naïve Bayes for Regression problem? [4]
(c) Can Naïve Bayes working without the Bayes Theorem principle? [4]

Neural Network(MLP):
(a) How does a Neural Network works on a Non-Linear dataset? [4]
(b) What are the similarities and dissimilarities between SVM and Neural [4]
Network?
(c) Describe the concept of overfitting in MLPs. How can you detect [4]
overfitting, and what strategies can be employed to prevent or reduce it?

Neural Network (MLP Regression):


(a) How does MLP works on Regression problem? [4]
(b) How does MLP works on a Non-Linear dataset? [4]
(c) Which is better between KNN Regression, MLP, and Linear [4]
Regression?

Task-2 (15 marks)


(a) Do you think that for every Machine Learning problem we should do [5]
hyperparameters tuning and why? Give your own opinion with proper
explanation.
(b) Discuss leave-one-out cross-validation (LOOCV). What are the [5]
advantages and drawbacks of using LOOCV compared to other cross-
validation methods like k-fold?
(c) Describe the role of ensemble methods, such as bagging and boosting, [5]
in managing bias and variance. How do these methods work, and when
should they be used?

LAB PART (25 marks)

Task-1 (15 marks)


(a) Solve the above dataset problem by doing necessary coding with python [10]
and Machine Learning algorithm you choose for the solution on Google
Colab. (Share the Colab link in the pdf also)
(b) Draw the Confusion Matrix, Classification Report. Also show the [5]
Accuracy/MSE of the model.

Task-2 (10 marks)


(a) Give a proper description of the dataset (Columns, Column Name, [2]
Rows, Dataset size, Dataset type)
(b) Draw Scatter Plot and best fit line graph or Train and Test loss graphs. [3]
(c) Visualize the ROC Curve and give proper explanation with ROC-AUC [3]
score?
(d) Draw a table for Training and Testing data ratio and give explanation of [2]
your split value?

NOTES:
** Submit the assignment in Google Classroom and also LMS Assignment section in due
time. (Marks will deduct for late submission)
** On the Final Viva day, you have to submit the hard copy of the assignment as well as
present physically for the Final Viva. (Absence in the Final Viva will terminate as late
submission and will deduct marks)

You might also like