Deadline Nov 5
The purpose of this proposal is to make sure that you have a plan and have done the preliminary investigation to ensure the project is doable.
We will not grade the document. This is for your project management. We hope that preparing the plan will clarify potential issues and ideas for you.
Submit a 1 page pdf document for your group. The document should contain:
Team. Mention your team in the document. Names, and ids.
A problem statement. Note that the outlines given are often relatively vague – such as “predict”, or “find communities” etc. You need to frame the problem more precisely in a way that lets you say which result is correct and which is incorrect.
Importance. Why is the problem significant? Why is your statement relevant to this motivation? Note that it may not be always possible to prove that the precise formulation is the right way to address the motivation, but you should give some argument why it is reasonable.
One or more relevant datasets. That contain the relevant information you need to solve the problem above. Note that the problem and the data need to be consistent: you cannot solve a problem for which you do not have the right data. State which attributes of the data you will use to solve the problem.
Related work. Which other papers have already addressed the same/similar problem? And optionally, how will your work be different?
An idea of how you will address the problem. What techniques will you use? Why are they relevant? Do you have an idea that has not been used, or lets you solve a problem that has not been considered before?
It may not be possible to answer all the questions above perfectly at this stage, but do try to answer them all to save issues later on.
Keep in mind that everyone will be graded separately. So it is helpful to sort out which team-member will do which part. You can optionally choose to include it in your document. Also consider whether your part is overly dependent on the success of your team-mate’s work.
Other thing to consider is trying some preliminary tests on the dataset. Does it have enough data? Does it have too much data? How long does it take to run the type of computation you plan? Remember that you may need to run many trials and may need to repeat due to errors etc.
Consider using the student.compute.inf.ed.ac.uk server. This server has many cores, so if your computation can be split into separate programs/scripts, this may be useful. But if everyone tries to use this on the last day, that will not help…
Add a note to piazza. Attach your pdf using the “insert” menu. The selected folder should be “project-plan.”
Select to post it to individuals. Add instructors and your teammates. Also mention all team-members in the body of the note. Ask questions for advice in the body of the note.