COMP 875 Project or Survey Paper: Final Guidelines

Due date: Thursday, December 3rd, 5PM (FIRM DEADLINE)

Submission: Email me a PDF document

Guidelines in Brief

  • Turn in your report on time!
  • Make sure your writing is self-contained and accessible
  • Include lots of figures
  • Define all your notations and label all your plots
  • A little math goes a long way!
  • Proofread and spell-check everything

Detailed Guidelines: Survey Paper

  • Length and format: The final survey paper should be about 15 pages in length and should be structured like a formal academic paper (typeset in LaTeX, single-spaced, 10 or 11 point font, one-inch margins). Here is a simple template you can use (courtesy of Rahul Raguram).

  • Subject definition: In the introduction, you should clearly define the subject area of your survey. Explain the subject in a high-level, accessible manner, preferably with illustrations. Outline the main conceptual issues the survey will be dealing with.

  • Review of literature: Even if your survey focuses in detail on only one or two specific papers or approaches, at least one section of your report should be devoted to a comprehensive review of existing literature in the general research area. A good place to put this survey is right after the introduction.

  • Math: Make sure to fully explain all the equations you choose to include in the survey. Carefully define all your notation, and don't introduce any notation that's not absolutely necessary. As far as I'm concerned, a little math goes a long way!

  • Illustrations and figures: Include lots of them! If you reproduce a figure from another paper, be sure to list the source in the caption.

  • Discussion: Compare and contrast the different approaches you surveyed. Point out strengths and weaknesses. Explain the overall impact and significance of your chosen subject area and suggest possible future research directions or alternative approaches.

  • Reference list: at least 20 papers.

Detailed Guidelines: Implementation Project

  • Length and format: The report should be about 15 pages long (however, much of this can be figures). It doesn't have to be as formally written as the survey paper, and LaTeX typesetting is optional.

  • Problem definition: Give a self-contained description of your problem (I should not have to refer back to your project proposal or progress report). List hypotheses (if any) that you were trying to test. If your project is related to your research, specify how it goes beyond your RA work.

  • Description of implemented method: This should also be brief, but self-contained. If you implemented a method from some paper, it is not necessary to give all the technical details from that paper, but you should give a high-level summary so I can understand what you did without referring to outside sources.

  • Resources: Explicitly credit any code/executables/data provided by other people. Give references to all relevant papers.

  • Data and results: Give a complete description of your data (source, size of dataset, type of supervisory information, dimensionality, etc.). Show figures illustrating the data. Describe your experimental protocol and show quantitative and qualitative results. If you have plots of quantitative results, make sure you clearly label and explain all the relevant quantities.

  • Discussion: What did you learn from the experiments? Do you find the results satisfactory? If not, what accounts for the disappointing results, and what could be improved? Did you encounter any surprises or unexpected challenges? What would you do if you had more time to work on the project?