CS 7641
Machine Learning
Reviewer Activites: Review of the Final Presentation

Numbers

Due: Friday, April 24, 2009, 23:59:59 EST
You will submit your suggestions in the usual way. As before, submit your reviews as PAPERCODE.txt and do not include your name on them.

This assignment is worth 2.5% of your final grade, i.e., ~17% of your "reviewer activities" grade. Again, if you fail to complete this activity, you will have failed to fulfill your duties as a reviewer, which will reflect poorly on you at the end of the semester.

Why?

We have finally reached the culmination of the semester---the final presentations---which means you have been granted a great opportunity to see many technical talks in the next few class periods and see first-hand what makes a good talk. It should be fun. The review form asks you to rate the talks based on two overarching criteria:

  1. Did the group address the criticisms brought up by the reviewers? The groups have only had two weeks or so to do this, so figuring out how they decided to budget the remaining time to tweak the final package may require asking some probing questions. You'll notice that the presentation review form is similar to the paper review form. This was intentionally done to highlight the fact that you will be focusing on what has changed (what has gotten better, what has stayed the same) since the final paper submission.

  2. How well did the presenter present the talk? This is intended to measure how well the group conveys their work. Key elements of this include how clearly they motivate the problem, how clearly they motivate their approach, and how convincing their analysis of the results are. Note that "failure" can also be interesting given a good analysis of what happened. Failure won't get you tenure, though.

What to turn in

For each paper you have been assigned, fill out a presentation review form in response to the corresponding presentation. You must briefly justify each one of your responses.