Final Project

Presentation video due date: April 25, 2023

Final report due date: May 9, 2023

Your final project is to build a series of ML models with a dataset of your choosing using either (1) at least 3 of the techniques we learned about in class or (2) a new technique which you developed yourself. You may use any libraries you want for your project implementation.

You may NOT copy code from public tutorials. There are a very limited number of datasets which tend to be used for ML tutorials, so we will know if you copy these tutorials. Most Kaggle datasets would not qualify for the final project since there are many code submissions publicly viewable associated with each Kaggle dataset. However, if the dataset you find on Kaggle has FEWER THAN 5 submissions, then that is acceptable.

If you use a dataset which can be imported from sklearn, TensorFlow, or PyTorch, then you must create a new methodology or evaluate an existing methodology which is not implemented by default in sklearn. For example, you cannot simply use sklearn on an sklearn dataset, but you can use sklearn on a new interesting dataset you find -OR- you can evaluate a new methodlogy on an sklearn dataset.

Rubric

3 points: Coding

• Project code is provided as both a .ipynb file and a PDF version of the .ipynb file: 1 point
• Project code meets all requirements described above: 1.5 points
• Code cleanliness, organization, documentation, and comments: 0.5 points

4 points: Written Report

• At least 1.5 pages single spaced 12pt font (not including figures and references): 1 point
• Introduction, Related Work, Methods, Results, Discussion, and References (at least 10 references) sections included: 1 point
• At least 1 Methods figure and 1 Results figure: 1 point
• Professionalism, grammar, and style in the format of a technical ML paper: 1 point

2 points: Discretionary Points / Course-Specific Requirements (435 vs 635)

• ICS/DATA 435 students: Your Related Work and Discussion sections should include applications of your project to the real world. Your References can be academic (e.g., research papers) and/or real-world (e.g., websites, videos, etc.).
• ICS 635 students: Your Related Work section should not simply summarize each paper, but you should synthesize the work and discuss high level themes. Your Discussion section should discuss the high-level overview of your findings, discuss limitations of your approach, and highlight opportunities for future work. All References must be from academic journal or conference papers. The overall writing should be in the style of a research paper.

1 point: Final Video

• Video can summarize any aspect about your project. There are no strict requirements other than communicating what you did. A common format can be talking through slides.
• Video should be uploaded to YouTube (can be unlisted). Submit the link on Laulima.
• Video should be no longer than 2 minutes.
• Videos will be viewed in class.
• You must show up to class during the project video presentation days to get the full point for the video, unless you are excused.

Submission instructions

Submit a PDF on Laulima. Make sure that your Colab notebook is shared with the course instructor and TA.