CS 5/579, Winter 2020
There will be a variety of programming assignments throughout the term, in which students will gain experience with various tools and workflows.
Each assignment will come with a due date, and submission will be done via Sakai. If you need an extension, please make your request by 5:00 PM two days prior to the assignment’s due date. In other words, if an assignment is due on a Tuesday, extension requests must be made by Sunday afternoon. This is to help prevent procrastination, to give you time to come to me for help, and to help me plan my grading time.
I will obviously be flexible about this in cases of illness, family emergencies, etc.
Most class sessions will revolve around presentation and discussion of assigned reading in the form of journal articles. Each such session will involve two students: a presenter, and a discussion leader. Each student will perform each role at least once during the class, and this will make up a substantial portion of your grade.
The presenter’s role will be to present the material from the assigned papers, and walk the class through their motivation, approach, and experiments. The discussion leader’s role will be to moderate the class discussion of the paper and its problem domain. They should be prepared with several “talking points” about the paper: questions they had (either about the authors’ approach, or about the problem area in general), ideas for additional experiments or improvements, critiques of the methodology (only if applicable), etc. Of course, the class as a whole is expected to be ready to participate in such a discussion, and should come prepared with their own questions!
Note that many of the papers assume some level of background knowledge: in databases, statistics, security, some family of algorithms, etc. As the presenter/discussion leader, it is your responsibility to familiarize yourself with the paper’s problem domain, and include sufficient material in your presentation or discussion points to help everybody get the most out of the paper. This will involve doing some additional reading beyond just the paper itself- plan your time accordingly!
For example, consider a hypothetical paper that implemented a distributed algorithm for topic modeling, and evaluated it on a newswire corpus. In addition to the contents of the paper itself, the presenter of this paper also would be expected to include some discussion of topic modeling in general (what it is, what it’s for, how it’s evaluated, etc.), an explanation of the standard version of the particular algorithm discussed in the paper and why it needed to be re-implemented in a distributed manner, and so on. If the authors included a quantitative evaluation, the presenter would be expected to describe the data set used by the authors (is it a standard data set? what’s in it? how big is it? etc.), and also compare their evaluation to that of other researchers working in a similar area- did they do things the “normal” way? How did their results compare? Note that much of this information will be findable by looking in the papers cited by the article itself.
If you are unsure about how to present a scientific paper (or even if you are), the following resources may be of assistance:
Feel free to ask me for help if you are unsure about anything in your assigned paper, or want help figuring out how to present it.