Project

Overview

For graduate students (CS680), the project is
  • required
  • individual (i.e., students must be do the project by themselves)
  • worth 25% of the final course grade

For undergraduate students (CS480), the project is

  • optional
  • group-based (up to 3 people)
  • worth 5% bonus on top of the final course grade

Project Scope

The theme for course projects this year is Machine Learning for Social Good, i.e., the application of machine learning techniques to analyze social problems, in diverse areas such as health, education, protecting democracy, urban planning, assistive technology for people with disabilities, agriculture, environmental sustainability, economic inequality, natural disasters (e.g., fire prediction), social welfare and justice, ethics, privacy and security, etc.

As a pre-requsite, please read Machine Learning that Matters by Kiri L. Wagstaff as a first step, and browse the following websites/papers for inspiration for project ideas:

In this project, you will do some research to describe the social problem that you chose to address, identify relevant dataset(s), implement at least 2 machine learning techniques and compare their performance empirically, discuss the extent to which your techniques can reveal useful insights or enable novel end-user applications towards solving the social problem. There will be 4 deliverables for the project: (a) proposal (1 page), (b) milestone report (4.5 pages), poster presentation, and final report (8 pages). Details about each deliverable is provided below.

Grading

The project is graded out of 100%, with the breakdown as follows.

Late Policy: Late penalties for all deliverables: Each deliverable is graded out of 100%. There will be a penalty of -20% (of the deliverable grade) for each additional day (9:00 pm to 8:59 pm). You are not allowed to submit the next deliverable if the previous deliverable was not submitted.

Weight
Deliverable
Due Date
Proposal
--
1 page report
8:59pm Jan 29, 2019
Milestone Report
20%
4.5 page report
8:59pm Feb 28, 2019
Poster (PDF)
5%
poster (electronic copy)
8:59pm April 2, 2019
Poster (Presentation)
15%
in-class presentation
April 4, 2019
Final Report
60%
8 page report
8:59pm April 15, 2019

Project Deliverables

Important: Do not change the structure of the templates (e.g., do not add, delete or change the titles of the sections/subsections).

1. Proposal (Submit PDF to LEARN 8:59pm Jan 29, 2019)

The proposal is a short description of your project idea. It should
  • be at most 1 page long, excluding references
  • include your project title, names and email addresses of all the people in your team
  • be prepared using this template: proposal.pdf, proposal.zip

2. Milestone Report (Submit PDF to LEARN 8:59pm Feb 28, 2019)

The milestone serves as a way to keep your project on track. It should
  • be at most 4.5 pages long, excluding references
  • include your project title, names and email addresses of all the people in your team
  • be prepared using this template: milestone-report.pdf, milestone-report.zip

3. Poster Presentation (Submit PDF to LEARN and EasyChair due 8:59pm April 2, 2019, Presentation on April 4, 2019)

The poster should be 3' wide x 4' high, in order to fit the poster boards which are 4' wide x 5' high. You can adopt one of the many poster templates on Overleaf. Posters are used to support your conversation with people who are visiting your posters. Here are some advice on how to create a good academic poster. Do not cram it with too much text, equations or graphs. Think about how you want to present your poster, and add just enough information on the poster to help you explain your work. The mark for the poster presentation will include assessment of your talk (clarity, structure, timing) and visual presentation.

Peer Review: All students (including those presenting posters) are required to visit and grade a number of posters; this accounts for 3% of participation mark and final course grade for all students. In addition, a number of external judges (e.g., professors and machine learning experts from the industry) will be invited to see and evaluate the posters. We will be doing the peer reviews using a conference paper review system called EasyChair.

4. Final Report (Submit PDF to Due 8:59pm April 15, 2019)

The final report is a coherent description of how you used machine learning to address a particular social problem. It should
  • be at most 8 pages long, excluding references
  • include your project title, names and email addresses of all the people in your team
  • be prepared using this template: final-report.pdf, final-report.zip