Course Analysis, DD2380 Artificial intelligence, 6hp

Course given 2007/2008 in study period 1

  • Course responsible:
    Patric Jensfelt

  • Lecturers:
    Patric Jensfelt, Danica Kragic

  • Lectures:
    12 lectures (24 hours)

  • Registered students:
    55 have completed the project assignment and at least one of the homeworks. In addition to this 2 PhD students attended the course.

  • Course material:
    Book:
    Artificial Intelligence: A Modern Approach (Second Edition) by Stuart J. Russell and Peter Norvig, Prentice Hall (2003), ISBN 0-13790-395-2
    All lectures were accessible for download.
    2 project descriptions and three homwork were posted on the web.

  • Examination:
    Assessment consisted of 3 homeworks (3hp) and one project (3hp). There were strict deadlines for the homeworks and the projects announced at the beginning of the course. All assignments were graded A-F. The homeworks had to be completed individually and handed in on time. The project was completed in groups of four students. A penalty was given to homeworks handed in after the deadline. The final grade on the course was given as the averae of the 3 worst results. The motivation behind this was to make it a bit less "aggressive" than taking the lowest among the four individual grades which seem like the exact mapping of what the grading system tels you to do.

  • Course results:
    50 undergraduate students have passed the course. An addition 5 students have passed the project (PROA) but not all homeworks (INLA). In addition the two PhD students have passed the course but these are not included in the statistics below.
  • Relation to the previous years:
    The biggest change from previous year was the grading system. As a result of this the homeworks and the project assignment were graded A-F. The nature of the homeworks were also changed to adapt them to the goal directed grades. Each homework was divided into two parts. The first one consisted of questions which required only short answers. These questions were intended to test the requirements for passing the course. The student had to correctly answer a large fraction of these questions to pass the homework and for the rest of it to be graded at all. The rest of the homeworks tested various parts of the course and this part was used to determine the grade E-A.
    The grading for the project assignment was changed so that more empahsis was given to the presentation, both written and oral. In addition, each group also served as opponents to another group during the presentation and their ework as opponents where also graded. The idea was to mimick the grading process around a master thesis project.
  • Summary:
    12 lectures were held. 55% of the students attended more than 80% of the lectures.

    Three homework assignments were given out during the course. The last two of them contained problems which required the students to make an implementation to get the highest grade. The implementation part of the homeworks could be caried out in groups of two students but the rest should be completed individually.

    The course is given in English and the course evaluation seems to support this. 10 out of 20 thought that it was essitial or good that it was in English and 8 did not care. Only 2 of the 20 would have preferred the course to be given in Swedish.

    The minesweep project was selected by all students. The examination consited of an implementation of an agent that could play minesweep, a report, a presentation and act as opponent on anoter group.
  • Results of students evaluation:
    20 students filled in the evaluation (until December 1, 2007).

  • Course level:
    Most students think that the course level was medium. Since it is an introductory course, they felt that their prior knowledge was suitable to follow the course.

  • Planned changes:
    The first part of the homework may have to be made a more difficult. It was too simple to pass the homework part of the course.

    Hope to involve some of the PhD students in the group that have now taken the course in the course.

    Investigate different project which has a tighter coupling with the course material.

    Change the calculation of the final grade so that it is calculated as the average of all four indivual grades. It hurts to have your best result removed from the calculations as in the current system and reduces motivation for the student.


  • How difficult was this course?

    1. 20% (4 st) Easy.
    2. 75% (15 st) Medium.
    3. 5% (1 st) Diffucult.


  • The course is given in English. How do you feel about this?

    1. 30% (6 st) If it was in Swedish I would not have taken it.
    2. 20% (4 st) It good that it is in English
    3. 40% (8 st) I don't care
    4. 10% (2 st) I would prefer it in Swedish


  • Was the aim of the course clear from the beginning?

    1. 75% (15 st) Yes.
    2. 25% (5 st) Not sure.
    3. 0% (0 st) No.


  • Was the course interesting and meaningful?

    1. 25% (5 st) Very much.
    2. 60% (12 st) Yes.
    3. 15% (3 st) Neutral.
    4. 0% (0 st) Not particularly.
    5. 0% (0 st) No.


  • Prerequisites for the course were courses in numerics and statistics. Do you think that your prior knowledge on this topics was enough to follow the course?

    1. 90% (18 st) Yes.
    2. 10% (2 st) Not sure.
    3. 0% (0 st) No.


  • Do you find course book by Russell and Norvig suitable?

    1. 80% (16 st) Yes.
    2. 5% (1 st) Not sure.
    3. 0% (0 st) No.
    4. 5% (1 st) Bought it but did not read it.
    5. 10% (2 st) Did not buy the book.


  • How many lectures did you attend?

    1. 10% (2 st) Less than 30%.
    2. 25% (5 st) 30-60%.
    3. 10% (2 st) 60-80%.
    4. 55% (11 st) More than 80%.


  • What do you think about the quality of teaching? Were the topics clearly presented?

    1. 25% (5 st) Very good
    2. 50% (10 st) Good.
    3. 25% (5 st) Acceptable.
    4. 0% (0 st) Not so good.
    5. 0% (0 st) Bad.
    6. 0% (0 st) Did not attend.


  • Did you read the relevant book chapters before they were presented at the lectures?

    1. 0% (0 st) Yes, always.
    2. 30% (6 st) Sometimes.
    3. 35% (7 st) Rearly.
    4. 35% (7 st) Never.


  • Did you think that the homeworks helped you learn the material in the course?

    1. 50% (10 st) Yes very much.
    2. 40% (8 st) Yes.
    3. 10% (2 st) Not really.
    4. 0% (0 st) Just a waste of time.


  • Did you think that the project helped you learn the material in the course?

    1. 35% (7 st) Yes very much.
    2. 25% (5 st) Yes.
    3. 40% (8 st) Not really.
    4. 0% (0 st) Just a waste of time.


  • Did you attend the minesweep "competition"?

    1. 80% (16 st) Yes.
    2. 20% (4 st) No.

  • Was it a good idea to have a competition in minesweep?

    1. 55% (11 st) Yes, great idea.
    2. 35% (7 st) Yes.
    3. 10% (2 st) Not really.
    4. 0% (0 st) Just a waste of time.


  • How many other courses did you attend in parallel to this course?

    1. 20% (4 st) One.
    2. 65% (13 st) Two .
    3. 10% (2 st) Three.
    4. 5% (1 st) More than three.


  • This is a 4p course. Compared to other similar courses, is 4p suitable?

    1. 80% (16 st) 4p is OK.
    2. 0% (0 st) Should be less than 4p.
    3. 15% (3 st) Should be more than 4p.


    things you liked/appreciated about the course:

    I liked that the final examination was a project that was then evaluated in form of a competition. I think that is a very good motivation.
    ---
    Minesweeper and the handwriting recognition. Good course book.
    ---
    The teachers were available when some questions needed to be asked. Help hours were relevant.
    ---
    I liked this course. We have AI courses on my home university as well, but I am sure that they are boring compare to this one.
    ---
    The minewsweep competition was a great idea.
    ---
    The minesweeper project was really funny and interesting!
    ---
    The project was funny, the work in groups is always interesting...
    ---
    The peek into many areas of AI.
    ---
    The content.


    things you would like to see changed next year:

    Remove part A of the homeworks. Maybe homework sets 1 and 2 could be merged and include some slightly larger task similar to the handwriting system. I know the course isn't supposed to go very deep into the many topics it deals with, but it's no fun if it's too shallow either. It's possible the course could benefit from having a few more things excluded (e.g. basic logic, since that's a mandatory course for D students anyway).
    ---
    The homework needs to be changed. Especially, the last homework, which was the hardest, should not be due at the same time as the project.
    ---
    The statistics and HMM were quite hard compared to the other parts.
    ---
    The last homework was a bit tedious, the last part of it should have more weight on it.
    ---
    Don't move the deadlines. Set a deadline and then enforce it.

    It would be nice to have projects other than minesweeper to choose from.