Deliverable1 Instructions
Deliverable1_Instructions
Deliverable1_Instructions
User Manual:
Open the PDF directly: View PDF 
.
Page Count: 2
Over the course of MAIS202, you will be completing a machine learning based project from a dataset of your choice for the final project.
At the end, you will have the option to either:
1. Demo your project by integration it in a webapp (or something more advanced) or
2. Present your work as a academic project through a poster.
McGill AI Society will be hosting a science fair where you will be showcasing your work! It will be awesome!
For both cases, you will end the project by writing a blog post about it,
Submission
This is an individual deliverable. All deliverables should be electronically submitted on Github and completed
with the same academic integrity and standards expected at McGill University. Include appropriate citations.
To submit, create a repository under your own Github, you can name it whatever you want, and push your report
there. List your repository link in this spreadsheet. All of the code and report for this project should be found in
this repository. Make sure to maintain it with properly documented README and structured code.
Submit this deliverable as “Data Selection Proposal.pdf”
Max length: 1 page
Deliverable Description
The first step of the project is to choose the dataset that you want to work with and propose your project idea.
1. Choose your dataset
You can choose any public dataset of your choice (don’t forget to cite them!). There are also a couple of useful
databases that are available: Google Dataset Search and Kaggle. Explain the reasons why you choose this dataset.
2. Methodology
Describe how you plan on approaching the project. This should be a high level overview of your plans, and this
will allow us to judge the feasibility of your project. Be as thorough as you can, so we can give you critical
feedback.
i. Data Preprocessing
Is the dataset you chose feasible? What information provided is/are the most useful? How are you planning on
preprocessing the dataset to extract this information?
Here are slides to our data preprocessing workshop last semester.
PROJECT DELIVERABLE 1
Due: February 6th, 2019
1.
i.
ii. Machine learning model
What do you want predict/estimate from this dataset? Propose a machine learning model/algorithm for it, and
explain your reasoning. Have you considered other alternative models? What are the pros and cons?
iii. Final conceptualization
For demo purposes, we want you to be able to showcase your project! Indicate your choice for the final project
and explain the details requested below. This is not final, you can always change through the semester, just keep
us updated.
- Application
We want you to integrate your model in a simple landing-page webapp. For those of you who have experience,
you are welcome to integrate your model in more sophisticated technologies (eg. mobile, hardware, webapps).
Keep in mind that this is machine learning course, so we will not providing significant guidance for software dev
related topics. Work with something that you are comfortable with :). Give the general idea of your application,
and the technologies you planned on using.
Example of a simple webapp that uses Computer Vision to estimate people’s age.
There will be no limitations to the product. You are free to build anything you want! :)
- Poster Presentation
We want you to prepare a poster presentation, just like what people find at NeurIPS, and other ML conferences.
Since this is a more academic route, your goal for this project would be to try and match and/or beat the standard
baselines for the output you are trying to predict. Do some research on previous explorations of the dataset and the
accuracies that have been achieved with it (include the references in this deliverable). Report the average baseline
results which you hope to beat (eg. predict x with y% accuracy).
