more courses, more meetings.

Today I’m joining a meeting with our Associate Chair for graduate studies.  It’s part of a series of meetings aimed at gathering student input towards shaping our department.  One issue that everyone seems to agree upon is the barriers to inter-group collaboration: grad students are divided between four buildings (BA, SF, PT, CCBR), so research groups located in different buildings rarely (or never) see one another.  This is a big downer, as it inhibits new and surprising collaborations.  How can we make collaboration easier?  Aran Donohue has a few good proposals.

I propose to try and encourage inter-group collaboration as a side effect of a solution to a different problem: to few graduate classes.  Every term there are between 10 and 20 courses offered, generally more in the fall term, fewer in the winter term, only a handful (if any) in the summer.  Graduate students at U of T are required to take a wide breadth of courses, eventually covering 6 of 8 areas: Programming Languages & Methodology, Systems: Hardware & Software, Numerical Analysis & Scientific Computation, Computational Complexity, Applied Discrete Mathematics, Artificial Intelligence, Computer Graphics & Human- Computer Interaction, Information Systems.  Yet in any given term, not every area is covered by an offered course, leaving a student who is trying to fulfill breadth requirements in a pinch.

To increase the number of courses offered, plus getting people together from different areas, I suggest creating an option for senior PhD students to teach a small seminar-style course.  Students would put together a course proposal that would be submitted to the department.  Approved students would teach this course in lieu of the two TAships that their contract offers.  The courses would be small (15 or fewer), and could be either of general focus (say introduction to computer vision algorithms), or specific (techniques for modeling human motion).  General courses would satisfy a breadth requirement.  The marking scheme would adhere to some guidelines:

  1. Marks determined by one proposal (paper or project), and the paper or project itself
  2. Groups of 2-3 students
  3. The groups must be composed of no more than two students from the same research group
  4. If it is a group of 2, the students must be from a different research group
  5. All resulting projects should be written up and posted to Aran’s suggestion 2f (software portfolio).  All papers posted to Aran’s suggestion 2a (technical report database)

Now there may be larger structural problems for this.  CUPE contracts are probably not easy to modify, but maybe the 2 TAships could be rolled into one 108 hr TAship.  I think it could work.