COE 292 – Introduction to Artificial Intelligence (241)

Project
Objective
The objective of this project is for students to gain hands-on experience in applying Artificial Intelligence
(AI) techniques to real-world data within their field of study. By identifying and collecting a relevant
dataset, students will explore the practical applications of AI through the implementation of two
machine learning algorithms and one deep learning algorithm for data classification problems. This
project will enable students to demonstrate their understanding of these algorithms by analyzing their
performance, interpreting results, and drawing meaningful insights from the data. Through this process,
students will not only learn to practically program a machine leveraging AI tools effectively but also
develop critical thinking skills to address domain-specific challenges using AI methodologies.

• Grouping
In this project students are expected to work in groups. The group forming policies are as follows:

1. The size of the group should not exceed 3 students where groups of 2 are also allowed. Just
make sure you register your group with your instructor as per the rules stated below.
2. All students in a group should be from the same section.
3. No preference is given based on group size and grading will not consider group size.
4. The grouping should be approved by the section instructor. Instructors may enforce some
additional rules on grouping based on the class size and dynamics in the classroom. Contact
your instructor for more details.

• Dataset
• The dataset submitted by the student must consist of at least 500 labeled examples with a
minimum of 10 distinct features (variables) per label and at least 3 unique classes for
classification. The dataset should be balanced, with no class comprising more than 45% or less
than 20% of the total labeled classes. A group must ensure that the dataset is balanced before
submitting by cleaning the data and removing the outliers using techniques described in ISE291.
The data must be relevant to at least one of the group member’s field of study, publicly
available and ethically collected, ensuring compliance with data privacy regulations. The dataset
should be formatted in a manner suitable for applying machine learning and deep learning
algorithms, with clear feature definitions and labels.
• You must use and fill the template provided named “datasetTemplate.docx” and submit it along
your data. Please do not change the file type and submit the file after renaming it to
“dataset_GROUPNumber.docx” and name your data file as “dataset_GROUPNumber.xxx” where
xxx is the file format you intend to use for the data. Note: in the final submission you will be
asked to use the same dataset as the one you submit and changes will not be allowed after the
dead line to submit the dataset (see submission details section for more information).
• The rubric for the dataset grading is shown in appendix A. Note: in case of discrepancy between
this document and any other document, then this document will prevail.

Page 1 of 6
• AI Algorithms
The group must show the use of the following algorithms in their project:
1. K-NN
2. SVM
3. Deep learning or CNN as applicable in the selected classification problem
• Expected Tasks, Requirements and Deliverables
• Choose a real-world problem related to one of the non-computing major student within the
group members.
• Find a reputed and reliable publicly available dataset for the chosen problem with the rules
specified in the dataset section above
• You must classify the data using the three different methods that are specified in the AI
Algorithms section.
• Implement, train, validate and test the selected methods using the chosen dataset, including:
o Thorough experimentation, cross-validation
o Studying the influence of different hyperparameters on the models’ performance
o Adapting and fine-tuning the selected methods to achieve the best performance.
o Explanation of failure cases: why models fail to perform well in some cases?
o Comparisons of models: based on appropriate performance metrics for the problem,
you should explain how the selected models compare? Why does a certain model
perform better than another one?
o Results analysis and discussion (interpretation) where evaluation of the performance of
each algorithm using these metrics: (1) accuracy, (2) precision, and (3) recall.
• Deliver working code for the above (as Jupiter notebooks or a python project), including a
video recorded demo of how to compile and run your code.
• Deliver a written report (based on the provided outline and format)
• Final report
• You must use and fill the template provided named “FinalReportTemplate.docx”. Please do not
change the file type and submit the file after renaming it to “FinalReport_GROUPNumber.docx”.
Note: in the final report you must use the same submitted dataset and changes will not be
allowed after the dead line to submit the dataset (see submission details section for more
information).
• The rubric for the dataset grading is shown in appendix B. Note: in case of discrepancy between
this document and any other document, then this document will prevail.
• The rubric of the final report will be shared in the third week of classes.
• Milestones and Timeline
• Team Formations: 03-Sep-2024
• Problem and Dataset Submission: 26-Sep-2024
• Code and Final Report Submission: 21-Nov-2024
• Feedback from Instructors (First Round of Grading): 28-Nov-2024
• Rebottle of Code and Final Report: 05-Dec-2024
• Demo and Presentation: 10-Dec- 2024 to 15-Dec-2024

Page 2 of 6
• Other Info
• Plagiarism is strictly prohibited and may lead to appropriate disciplinary actions as per KFUPM
policies
• You may derive inspiration from others’ works (e.g., Kaggle or Github repositories) but your
project needs to be your original work.
• Any assumptions made (in your problem formulation, model development, dataset collection,
experimentation, etc.) need to be mentioned and justified in your report.
• Grading Policy
There will be 4 submissions as follows:
1. On 03-September 2024, students must finish forming teams as per the team formation
rules.
2. On 26-September 2024, each team will submit their dataset along with the dataset template
document which will be graded as per the provided rubric in Appendix A. Once grading is
done, a multiplicative factor (𝛼) will be determined that is the obtained grade of the team
in this activity divided by the maximum grade possible in the dataset template rubric.
3. On 21-November-2024, students will submit the final report with the code. The total points
of the final report based on the rubric will be divided by the total number of points possible
in the rubric and multiplied by the number of points stated in the syllabus for the grade (we
will call this grade (𝛽)). Then on 28-November-2024 student will get a feedback on the
report. Student are then given a second chance to submit the final report and the code after
addressing the comments of the instructors that will be graded again using the same rubric
and calculate the grade in the same way discussed earlier and this grade will be referred to
as (𝛾). Your final report grade then will be calculated as max(𝛽, 𝛾) or effectively grading we
will regrade you without penalty for the final report. Note that students can resubmit a
revision of the dataset report for regrade on 07-November-2024 and it will be regraded
without penalty to students. However, no changes to the dataset is permitted after 07-
November-2024.
4. Your final grade will be calculated as: 𝑜𝑣𝑒𝑟𝑎𝑙𝑙 𝑔𝑟𝑎𝑑𝑒 𝜙 = 𝛼 × max(𝛽, 𝛾)
5. After the project presentation, each student will be assigned a factor (𝜔𝑖 ) where 𝜔ϵ[0,1].
This will be based on the instructor’s evaluation on how much each student contributed. For
excellent effort by the student, he will be assigned a 𝜔 of 1 while free riders will be assigned
a 𝜔 of 0.
6. Your final grade (𝐺) will be calculated as 𝐺 = 𝜙𝜔𝑖
7. Not submitting the dataset report by 07-November-2024 will result with a zero grade in the
entire project points.
8. Not submitting the code and Final Report on 21-Nov-2024 will result with a zero grade for
the entire project and students will not be given a chance to resubmit on the rebottle
submission date. Note that rebottle is only for groups that have submitted on 21-Nov-2024.
9. Submission of dataset and the associated report on 26-September-2024 and not revising the
dataset when performing classification will result with a bonus of 1 point added to 𝜙.
10. The first submission of the final report that score more than or equal to 90% of the final
possible grade will be given 10% bonus and result with 100% of the grade obtained in this
part of the grading process. Note that this applies only to grade (𝛽)) and not to 𝐺 where 𝐺

Page 3 of 6
 will be determined after the final presentation and the instructor assigning each student the
appropriate weight of 𝜔𝑖 .
11. Note that the bonus point of the project cannot be carried over to any other grade such as
exams, assignments or classwork.

Note: each student in a group may obtain a different grade based on the factor 𝜔𝑖 where it is
possible that all group members get 𝜔𝑖 =1. It is also possible for a group of three to have the
following factors for each student 𝜔1 = 1, 𝜔2 = 0.75, 𝜔3 = 0.1, which means that student 1 did
his work while student 2 did most of his work and student 3 did little work.

Page 4 of 6
 Appendix A: Data set Grading Rubric
We will release this rubric after the team formation.

Page 5 of 6
 Appendix B: Final Report Submission Rubric
We will release this rubric after the dataset submission date.

Page 6 of 6