Supervised Machine Learning (Spring 2012)
-----------------------------------------
There were 28 registered students, 14 participated (by turning in at
least some homework), 3 were in the exam and 3 were accepted, all with
grade 5/5. Feedback from student was collected by department's
standard form (8 students responded), and some students commented the
course to the teaching assistant on their own iniative.
The main reason that can be deduced from the feedback for the
unusually high percentage of students who didn't even attempt the exam
was the theoretical emphasis of the course. Students found the topics
uninteresting, or not sufficiently motivated, or difficult, or
requiring prerequisites they didn't have. The course did not meet the
expectations based on "Introduction to Machine Learning."
The average of the numerical scores given in the feedback forms was
2.9, which is quite low by the department's standards. The variance
of the scores was high in all questions, except that the work load of
the course was consistently considered very high.
There are no current plans of when the course would be lectured again.
The theoretical emphasis is quite intended and part of the stated
goals of the course, but in case the course is repeated, perhaps this
can be communicated more clearly to the students in advance.
Regardless of that, the mathematical background of the students should
be considered more carefully. What exactly this would mean is not
clear right now, since the differences between students are large in
this regard, and we also do want to offer something to students with a
strong mathematical background.
Jyrki Kivinen, 14 March 2012