CS369G: Project

Spring Quarter 2016, Stanford University
  • The course has a project requirement that constitutes 30% of the final grade.

  • The project is due Wed June 8 1:29pm.

  • To claim a particular paper to write a report about send an email to the staff list listing your team. To check whether a paper has been picked check this document.

Goal

The purpose of the project is to get you acquainted with recent developments related to the topics discussed in class. Students will pick a research paper from a list of possible topics and write a report about. Focusing on one paper will allow you to go into depth in a particular problem and thus see how the various ideas you have seen in the course can be applied or extended. The course project is a nice starting point for further research investigation and possibly an original contribution.

Guidelines

  • Students should form teams of two people and pick a paper to write a report about. If you want to pick a paper that is not on the list, you should email the staff list with a brief explanation of how it relates to the topics of the class. Also, after request we may allow a project submission by a single person. To check whether a paper has been picked check this document.

  • Scope: a succesful project is one that indicates that you have understood in depth the main ideas and techniques employed in the paper. You should not present all the details of the paper, but only include enough detail so that the main steps are clear and well motivated. The report should put emphasis on clarity and try to convert the intuition. The level of exposition should be suitable for a graduate student that is taking this class. To get an idea of the size of the report think of it as supporting notes for a two-hour lecture (around 10 pages).

  • To help you get started, try to answer the following questions:

    • What is the problem that this paper addresses?

    • Why is it interesting and what where the previous attempts?

    • What are the main results of the paper?

    • What are the new ideas in the paper and how where they motivated?

    • What are the main steps of the proof?

    • What are the limitations (if any) of the proof technique?

  • Grading: a report that sufficiently answers the above questions in a clear and concise manner will receive a full grade. Optionally, for bonus points you can implement an algorithm presented in the paper and include a brief experimental evaluation on real or generated data.

Project Topics

Hashing

Sketching - Sparsification

Dimensionality Reduction

Fast Numerical Linear Algebra