CSC242: Intro to Artificial Intelligence
Spring 2012
Upper Level Writing
Students may elect to use this course to fulfill their Upper Level
Writing requirement. It's not a bad idea, since I like reading essays
and AI has plenty of great topics to write about.
Please see the Computer Science Department's requirements for Upper
Level Writing, part of
the URCS
Undergraduate Handbook (Section 6 as of this writing). To fulfill
these requirements, you will submit two drafts of your essay for
review before the final version is due at the end of term. Your essay
should be no less than 8 pages at the first draft (and shouldn't get
any shorter subsequently).
Schedule
Here's the schedule for ULW in 242:
- Feb 1: Choose your topic and send a one paragraph to one
page description to me via email.
- Mar 1: First draft due. No less than 8 pages. I will review
and give you feedback.
- Apr 1: Second draft due. Ditto on feedback.
- May 1: Final version due.
Topics
For topics, browse the textbook for interesting subject areas,
especially those not covered in the syllabus. The
Association for the Advancement of Artificial
Intelligence has some very nice AI
Topics pages with lots of ideas and suggestions for further
reading or viewing.
One warning: Like AI problems, AI topics are big and
complicated. Whatever you choose, you will need to focus. If this is
not a strength for you, make an outline and edit aggressively.
Suggestions
Given a topic, I suggest that you follow one of the following
approaches for writing your essay:
- The historical approach: What were the original motivations
for the problem, what has been done, what's being done today, and
what's left to do. Have the original goals been met? Why or why
not?
- The survey approach: Produce a catalog of AI techniques or
applications of the techniques to your topic. You will need to come
up with some kind of organization to make this into a readable
narrative. A ten-page bullet list is not an essay.
- The research approach: Focus on some narrower topic or
sub-problem and dig into a bit more deeply. You'd still have some
history and something on applications, but they would be background
and motivation for the technical part. This would be like a research
paper, only it wouldn't be your original research (unless you get
inspired ;-)).
- The implementation approach: You could code up something
beyond what we do in our projects, and report on that. Again, you
will still need background and motivation for your problem (more
than you would in a project write-up). Your results needn't be
original research, but they would need to be presented
professionally as in a research paper.
Of course other approaches are also possible. Just keep your reader
(me) in mind.
Last update: Thursday, 26-Jan-2012 10:30:37 EST