CS5/624: Problem Solving with Large Clusters

Project Proposals

Handed out: April 6, 2016
Due: 23:59 on April 18, 2016

Description

Please prepare a short (1-2 pages) proposal for what you would like to do as your final project for CS624. The goal of the final project is to make use of the technologies and approaches we are learning about to answer a research question of your choice. Your resesarch question may be related to your "day job", PhD work, etc., or it may simply be something of interest to you personally.

For this project, you do not necessarily need to propose a hypothesis-driven research question- a thorough, systematic empirical evaluation of several different algorithms or tools is perfectly appropriate, provided that it is well-motivated. Similarly, replicating a result reported in a paper (e.g., "using tool X to fit a model gives a ABC% speedup over tool Y") may also be appropriate. All of that said, coming up with good hypotheses is a difficult but important skill, and this is a good opportunity to practice. :-)

Your proposal should explain your research question and its motivation, and also include the following elements:

After preparing your proposal, the next step is to conduct a pilot study. The goals of the pilot study are to, at a minimum:

  1. obtain whatever data you'll need for the full project,
  2. get that data into a form that you will be able to work with it, and
  3. take the first few steps towards addressing your research question.
This might mean building a simple baseline system to compare your new idea or algorithm against, generating some of the descriptive statistics that you'll need to guide your next steps, etc.

You will be preparing a writeup of your pilot study, as well as a short (10 minutes) presentation to be given in class on May 4th. Your pilot study writeup should essentially make up most of the "background" section of your final project writeup, as well as part of the "methods" and "results". It should also include discussion about what went right, what went wrong, and what the next steps are to move from the pilot phase to the final phase.