Research Papers

Much of what is happening in the world of big data is driven by research in both academia and industry, so we will read, evaluate, and present research papers in this class. Being able to read research papers is a helpful skill that will serve you well regardless of your career path. Even more importantly, you should be able to read a paper, extract the most relevant details, and then implement them or share them with others.

To this end, we have two types of research paper assignments in this class: presentations and discussions.

Here are some helpful resources:

Presentations

You are required to choose a paper to present to the class with up to 3 teammates. Each team will do 1 presentation during the semester. As long as the paper is related to big data, you can present it. There are many safe choices in the class paper library, but if you’re excited about something else then don’t be afraid to find your own paper!

Once you find a paper you are interested in, you should do a cursory read-through to make sure it lives up to your expectations, and then post your chosen paper (and teammates) on the class discussion board.

Here are the guidelines for the presentation:

The purpose of your presentation is to inform the class about your topic and teach them something new. Most papers cover a single “big idea” that forms the basis of their approach. You should try to extract the most important parts of the paper and present them to class alongside any background information needed to understand the context of the paper.

Here are a few things to think about when you are preparing your presentation:

Remember to think critically about the work: insight is more important than having tons of slides that just repeat what the paper says.

Evaluation

Your classmates will submit an evaluation of your presentation that answers the following questions:

  1. Was the organization of the presentation well thought-out and easy to understand? (yes/no/partially)
  2. Were the visual aids helpful, tasteful, and appropriate to the subject matter? (yes/no/partially)
  3. Did you learn something new from the presentation? (yes/no/partially)

In addition, your classmates will be able to submit constructive feedback about your presentation. Note that only constructive, helpful feedback will be passed on to the presenters; part of what we’re doing here is trying to all become better presenters.

Finally the instructor will also evaluate your presentation on:

Scores from both sources will be combined to produce a final score for your presentation.

Discussions

Beyond your paper presentation, you will also be required to participate in 3 paper discussions. Paper discussions happen before the presentation and require you to:

  1. Read the paper in advance (at least 48 hours before it is presented)
  2. Add notes, point out confusing sections, list resources you found online, etc. on the shared “Community Notes” document for the paper.
  3. Post a comment on the discussion board about the paper that contains:
    • A short, ~240-character ‘tweet’ that summarizes the paper
    • Thoughtful questions you had while reading it
    • Any critiques you have of the approach
  4. Come to class on the day the paper will be presented and participate in the discussion (your questions are already ready to go, so it won’t be difficult!)

Since there will be about 8 presentations over the course of the semester, you only need to read half of the papers. If the paper was hard to understand or you were lost, be sure to mention it in your discussion post or community notes! This will allow the presenter some time to try to find the answers for you and make sure they cover material that was difficult for everyone to understand.

Evaluation

Discussions are largely participation assignments; perform the steps above and you get full credit. However, simply repeating what was already in the paper’s abstract or asking some basic questions without doing the reading will not earn you credit. The instructor will follow up on your discussion posts and prompt for more information if they do not meet this requirement.