Home Prospectus FAQ Staff Schedule Style Guide Grading Marks

Answers to FAQ

• You may use your work to extend or complement the work you do in another class or another context.
• You are welcome to work together in teams, just like in the real world. If you work in a team, you need to share the writing and oral presentation.
• You can get an A in the subject with any project type.

Delivery options

You do a project only if you are registered for the graduate version, 6.833. The preferred way to discharge your project obligation is:

• Do something interesting that is related to the subject, loosely defined.
• Deliver a 15 minute oral presentation on what you have done.
• Submit your presentation as if it were a report done by a study group. That is, submit a one-page executive summary plus your slides cum heavy legends, as explained in more detail after spring break.

The less preferred way is:

• Do something interesting that is related to the subject, loosely defined.
• Submit a written paper of 10–15 pages or so in the form of a journal article.

Project types

Projects can be, but are not limited to, any of the following types:

• Best: A complete research project. You both have idea and do the implementation, and report on your experimental results. This is the heroic way, inasmuch as your project is not all you do in the subject. You might accomplish more by working in teams, but then you are exposed to weak-link problems exacerbated by the part-time, conflicting pull nature of student labor. Caution is advised, along with especially clear and modest goals.

• Very good: A reimplementation. You pick one of the papers that describes an implemented idea and you reimplement, reporting on the surprises that invariably emerge when you understand something at the level required to write code. You also report on improvements you make.

  For example, you could redo the work reported in the Greenblatt and Vaina paper or the Borchardt paper.

  Of course, everyone seems to overestimate what they can do in a short time. On the other hand, if sufficient detail is given, and you are a good coder, this can be a lot of fun.

• Good: A research proposal. You propose a mechanism for performing some clearly defined task.

  One way to generate ideas is to write all the paper names on slips of paper, put them in a hat, and draw two at a time, asking if they make any sense. A famous mathematician (who?) claims he does new mathematics in a roughly analogous fashion.

  For example, you might draw the Rao paper and the Borchardt paper and ask if Borchardt's representation might be useful in describing Rao's routines.

  If you get inspired and passionate about something, you might just find yourself with a real thesis proposal when you are done with this. After all, Dave Huffman discovered Huffman coding while doing a term paper in Bob Fano's information theory class.

• Ok: A simulated topic exam. You read several papers and do a scholarly critique. Part of the process is to pick a set of dimensions (ugh) along which to compare and contrast the papers, recalling the spider diagram approach. You conclude with a description of the degree to which you were moved (yippee), and why.

  For example, you could contrast the work of Pearl (Bayes nets) with that of Zadah (fuzzy sets), viewed as efforts aimed at explaining commonsense reasoning; or you could compare various papers on commonsense reasoning; or you could examine the contrasting views of various anti and pro AI philosophers, such as John Searle (Berkeley) and Daniel Dennett (Tufts). When you have a rough idea of what you might want to do, I can perhaps recommend particular papers.

  The simulated topic exam is your safety net; if your other attempts do not work out, or you just cannot identify something you want to do, this is the sure-fire, finite-time, still educational, way out.