Project/Final Assignment: It's Up To You

What concepts or methods from the machine learning field interest you the most? This final assignment is your chance to explore these interests. Projects can include:

• Exploring variants of an algorithm studied in class. Such a project would compare the performance of the variants across a number of datasets (e.g. taken from the UCI ML repository).
• Duplicating a result reported in a publication.
• Applying ML techniques to a dataset that you have personal interest in. We can provide ideas from bioinformatics and EEG data analysis (Chuck Anderson who works on brain-computer interfaces would be happy to talk to you about that). See e.g. data from BCI competitions.
• Exploring algorithms not studied in class.
• Exploring issues such as over fitting, to obtain a better understanding of when it occurs for high dimensional data, and which algorithms are more susceptible.
• Participate in a machine learning competition among the ones available at the kaggle website.

You are required to work in pairs for the project. Three member teams will be allowed, but the need for the team should be justified to the instructor.

One report is submitted per team. Clearly define what the responsibility of each team member.

Each team will present their work in a poster session that will be held during the last week of classes.

Your project needs to demonstrate you have learned something new about machine learning.

Submit via Canvas a proposal by November 6th describing

• the machine learning problem/algorithm that you are planning on addressing in the project
• the steps you will take to investigate it
• who is on your team,
• what each team member will do, and
• a list of at least four key steps your team will complete and the dates by when they will be completed.

Only one person from your team should submit the proposal via Canvas. Proposals are limited to two pages.

Submit via Canvas the PDF file of your written report by midnight December 15th. Similarly to assignments, there is an eight page limit, using the Latex style of your choice. Remember, your report must clearly describe what each team member did; all members of the team need to participate in the writing.

As we discussed in class, the recommended structure of the report should include:

• Title
• Abstract
• Introduction (incl. problem statement, motivation for the work, review of previous work, open questions in the domain, and how you are proposing to address them)
• Methods. Make sure to give a detailed enough description of new algorithms that were not covered in class.
• Results and Discussion
• Conclusions and Future Work

Only one person from your team should submit the final report. In grading, each team member's contribution will be taken into account.

This is only for students in the on-campus section.

Here is what the grade sheet will look like for this assignment.

CS545 Final Assignment

======================================================================
Proposal:  10 points

  (6 points): The project is well-motivated (why are doing this).
  (2 points): Clear plan (what are the tasks involved).
  (2 points): Clear team member responsibilities (who will do what).

======================================================================
Poster:  30 points (only for on-campus students)

  (10 points): The poster was well presented by the team
  (10 points): The poster succinctly and clearly explains what you did
  (10 points): The poster was well laid out and makes good use of visual aids

======================================================================
Written Report:  60 points (90 points for distance-learning students)

  (10 points): Good introduction and motivation for the project.
  (15 points): Clear description of what you did and how you did it.
  (15 points): Results are clearly presented and discussed in depth.
  ( 5 points): Spelling, grammar, organization.
  (10 points): Project is sufficiently challenging.