Numbers

Due: February 8, 2009 23:59:59 EST

Please submit a short text file via tsquare. Please include your name in the body of the text.

This assignment is worth 0% of your final grade; however, if you don't do it you will have violated a requirement of your duties as a reviewer and will suffer greatly for it, i.e. you will not pass the class.

Why?

Reviewers typically have background knowledge for the papers they review. You will probably not have this luxury. So, as a favor to you, we are requiring certain things from you so that you'll start thinking about this stuff early in the semester, gain the required background, and know what's going on when you see the final paper submissions. How nice of us.

What to turn in

A list of keywords. Seriously, that's it.

Below you will find a short list of keywords that represent the topics covered by the group projects. It is up to you to select four or five of these keywords and send your list to the TA to help us in matching reviewers to papers. You can rank the keywords if you like. The reviews will be single-blind, i.e. the reviewers will know which group they are reviewing, but the groups will not know who is reviewing them. Each project will have a lot of reviewers assigned to it, and each reviewer will have exactly two projects to review. But all of that comes later. Just a heads up.

Keywords

• Supervised learning, classification.
• Activity discovery. This has an unsupervised flavor.
• Activity or object recognition. This has a supervised flavor.
• Segmentation (in the computer vision sense).
• Image Analysis, classification.
• Probabilistic methods. This includes Bayesian ways of thinking about things.
• Feature selection.
• Symbolic vs. numerical learning.
• Dealing with noise.
• Game playing.
• Reinforcement learning.
• Anomaly detection.
• Ensemble methods, e.g. boosting.
• Prediction.
• Time series prediction.
• Support Vector Machines.

Keep in mind the keywords you submit will determine what type of projects you will have to review, so do your research now to make sure you know what each keyword really means. That way you won't get stuck with something that really bores you AND you won't miss out on something that really excites you.