Visa resultat

MPALG 1213-4 Artificial intelligence, TIN171|DIT410

Opening question

1. Which university do you belong to?

24 svarande

University of Gothenburg»		4		16%
Chalmers University of Technology»		20		83%

Genomsnitt: 1.83

Your own effort

2. How many hours per week did you spend on this course?

24 svarande

At most 15 hours/week»		0		0%
Around 20 hours/week»		1		4%
Around 25 hours/week»		6		25%
Around 30 hours/week»		12		50%
At least 35 hours/week»		5		20%

Genomsnitt: 3.87

- It was hard knowing how much you had to do to achieve the different grades. This made us work a lot more than we had to (as we understod it on the final supervisions).» (Around 30 hours/week)
- Very irregular time spent as our group did not plan ahead very well, but very much time was spent on both projects.» (Around 30 hours/week)
- The course demands a lot of work. In my case it was especially much, because we were effectively two working students of originally four students in my group.» (Around 30 hours/week)
- Too much..» (At least 35 hours/week)

3. How large part of the teaching offered did you attend?

24 svarande

0%»		1		4%
25%»		4		16%
50%»		5		20%
75%»		4		16%
100%»		10		41%

Genomsnitt: 3.75

- If I remember correctly, all the scheduled stuff was in the first few weeks, and then I was very inactive.» (0%)
- The lectures were irrelevant and on a low level. BIG difference from the other courses at master level. Needs to be changed!» (25%)
- I stopped attending lectures after it was made clear that topics would only be covered after we had already implemented them ourselves in the projects, which seems like a waste to me.» (25%)
- Only supervisions » (25%)
- The lectures were hard to follow. The introduction to the first assignment were given after the project started, which made the problem harder and the lecture unnecessary since we already had study it ourselves. » (50%)
- Attended supervisions, but not all lectures.» (75%)
- For many students the offered lectures felt useless. This was due to that it was not clearly said, that the lecture gives just an rough overview and the lecturer was kind of not motivated.» (100%)

Goals and goal fulfilment

4. How understandable are the course goals?

The goals are difficult to understand»		1		4%
The goals give some guidance, but could be clearer»		12		50%
The goals clearly describe what I am supposed to learn»		11		45%

Genomsnitt: 2.41

- The goals linked to are for the previous course.» (The goals are difficult to understand)
- The course syllabus is not updated! There is not just one project anymore.» (The goals give some guidance, but could be clearer)
- I don"t think the goal has been changed whilst the course content is changed. "via a project, and acquire an overview of the other areas".. but we had two projects. » (The goals give some guidance, but could be clearer)
- It is not 100% what is expected from the projects.» (The goals give some guidance, but could be clearer)

5. Are the goals reasonable considering your background and the number of credits?

24 svarande

No, the goals are set too low»		1		4%
Yes, the goals seem reasonable»		21		87%
No, the goals are set too high»		2		8%

Genomsnitt: 2.04

- Reasonable but not updated.» (Yes, the goals seem reasonable)
- See previous comment.» (No, the goals are set too high)
- Although, my background wasn"t really the usual.» (No, the goals are set too high)

6. Did the examination assess whether you have reached the goals?

No, not at all»		2		8%
To some extent»		10		41%
Yes, definitely»		11		45%
I don"t know/have not been examined yet»		1		4%

Genomsnitt: 2.45

- The oral examination needs to be more organized. It said that it wouldnt examine our report, yet Pasad was reading it during the examination and asked questions about it. We also got questions about alternative planing algorithms and had to answer questions from the book (which it clearly said that we wouldn"t have to do). Overall: It seemed like that you hadn"t figured out what you wanted with the oral examination. You need to decide what the oral examination is for and after that how it should be done. » (No, not at all)
- See previous comment.» (No, not at all)
- There was a very large focus on NLP.» (To some extent)
- In my case, where 70% of the work was done by me, I feel I"ve been kind of cheated from getting a real score.» (To some extent)
- Pretty good contact was had with Peter, I think he had decent check of our performance.» (Yes, definitely)

Teaching and course administration

7. To what extent has the teaching been of help for your learning?

Small extent»		10		43%
Some extent»		9		39%
Large extent»		3		13%
Great extent»		1		4%

Genomsnitt: 1.78

- Too much focus on areas which could"ve been prerequisites to the course. Could have a crash course the first week, or have required readings for those who have less knowledge about algorithms & logic » (Small extent)
- Many of the lectures felt unnecessary since they didn"t have anything to do with the projects. I guess there were relevant subjects, but when you have to focus all your energy on the projects it"s hard to appreciate lectures about other things.» (Small extent)
- We did it all on our own and only used the supervisions to understand how much we had to do. Didn"t really get any guidance in how to do the different algorithm efficent.» (Small extent)
- The order of lectures and projects is a mess. Relevant lectures should be held for the relevant project.» (Small extent)
- I didnt attend the lectures, but some help was had in the group meetings and other meet-ups. But most was solved within the group.» (Small extent)
- Group supervision sessions are good but lectures should be given about a topic before we are done working with that topic.» (Some extent)
- With some added exercises about the lectures I might have "absorbed" more of the material, currently the subjects were swapped so fast I"s hard to recollect much of it the day after.» (Some extent)
- The lectures were not really useful, but the supervision sessions especially the ones by K.V.S. Prasad were good.» (Some extent)
- For learning: good For the project: not very much» (Large extent)
- Kudos to both Peter and Prasad!» (Great extent)

8. To what extent has the course literature and other material been of help for your learning?

Small extent»		9		39%
Some extent»		6		26%
Large extent»		4		17%
Great extent»		4		17%

Genomsnitt: 2.13

- Did only use the book a little for the second project. You shouldn"t tell students they really have to buy the book, or they will have difficulties, when that"s not really true.» (Small extent)
- Only used it to get references. I think it is really bad manners to say "You wont pass the course without the book" in the beginning of the course if it isn"t true.» (Small extent)
- We used a very small part of the book.» (Small extent)
- Just for getting references. The lectures was far to trivial. » (Small extent)
- I mostly looked things up on the internet.» (Some extent)
- Extensive course literature with lots of good reading.» (Great extent)
- I didnt get the book, I got introduced to the subject a lot by my more knowing team members, otherwise "learning by doing" and online resources was most helpful.» (Great extent)

9. How understandable is the description for the "Sentimentalizer" project?

I have not seen/read the description»		0		0%
The description is difficult to understand»		1		4%
The description gives some guidance, but could be clearer»		9		37%
The description clearly describes what I am supposed to do»		14		58%

Genomsnitt: 3.54

- The description should have been clearer. It appears that some things should not have been there to begin with, such as the McNemar tests.» (The description is difficult to understand)
- State clearly that data mining and processing for different features is a quite large part here, not only the machine learning algorithms.» (The description gives some guidance, but could be clearer)
- The description had some clear ideas of what should be done, but in the supervision we were told that less could be done and that we still could achieve a VG. Work on them!» (The description gives some guidance, but could be clearer)
- I am doing this course evaluation a few months later, I have no recollection of the description being exceptionally good or bad. We decided what to do ourselves and Peter was cool with it basically.» (The description gives some guidance, but could be clearer)
- I think it was pretty good» (The description clearly describes what I am supposed to do)
- Only thing though, when we tried to include all the tasks in the instructions, our (first) report grew much longer than ~10 pages. We were not sure how formal and how detailed descriptions were expected. Our second report was however more focused, part because of experience but mostly because of lack of time.» (The description clearly describes what I am supposed to do)

10. To what extent has the "Sentimentalizer" project been of help for your learning?

Small extent»		0		0%
Some extent»		8		33%
Large extent»		11		45%
Great extent»		5		20%

Genomsnitt: 2.87

- I learned the concepts of machine learning but I lacked interest to learn the specifics in detail.» (Some extent)
- Implementing the algorithms were great fun, there was however, in my opinion, to much "fluff" going on. It would be nice to be given some skeletons, for example a test suit generator, so more time could be dedicated to the algorithms and their concepts.» (Some extent)
- This topic was not so interesting because it was basically a repetition of basic machine learning, that I have learned before in the according course. So it was mainly programming.» (Some extent)
- Project = entire course.» (Large extent)
- I was already familiar with the concepts of machine learning. The processing of the raw data was however instructive. » (Large extent)
- Very interesting and fun! I didn"t mind spending so many hours on it» (Great extent)

11. How was the workload for the "Sentimentalizer" project?

Too low»		0		0%
Low»		0		0%
Adequate»		4		16%
High»		15		62%
Too high»		5		20%

Genomsnitt: 4.04

- The parts of the project not related to algorithms took far more time than expected.» (High)
- It was ok, sadly my team mates didn"t put enough effort to make this project more than a G. Their poor work didn"t help reflect that I personally have a good understanding of the algorithms.» (High)
- We spent a lot of extra hours on this, mostly because of the number of different test cases we had to set-up.» (High)
- Bad group work, I took on too much, also I had a bit low pre knowledge, so I had to spend much time on "easy" programming problems.» (High)
- It was too much to do with four algorithms to be tested in a lot of ways each. » (Too high)
- It"s a big project that is explained in the description. » (Too high)
- Most weekends wasn"t free from work. Of course this can depend on our ambition. The needs for different grades was not clear. » (Too high)

12. How understandable is the description for the "Shrdlite" project?

I have not seen/read the description»		0		0%
The description is difficult to understand»		2		8%
The description gives some guidance, but could be clearer»		16		66%
The description clearly describes what I am supposed to do»		6		25%

Genomsnitt: 3.16

- Some example or readings for a basic planning algorithm would be nice.» (The description gives some guidance, but could be clearer)
- The project description was fine, but there were a lot of work to setup the porject with webserver and everything and make it work. There should be more guidance for this I think.» (The description gives some guidance, but could be clearer)
- We were a bit unsure about how formal the report was expected to be.» (The description gives some guidance, but could be clearer)
- It"s not perfectly clear what is absolutely necessary.» (The description gives some guidance, but could be clearer)
- A better explanation of what is requested could be better. It should suffice to make sure in the first supervision that the students really know what is expected.» (The description gives some guidance, but could be clearer)
- Most of all difficult to get it started, team members fixed that for me, after some discussion in group and with supervisor the project got more understandishable.» (The description gives some guidance, but could be clearer)

13. To what extent has the "Shrdlite" project been of help for your learning?

Small extent»		2		8%
Some extent»		7		29%
Large extent»		8		33%
Great extent»		7		29%

Genomsnitt: 2.83

- I felt that the project didn"t really require us to learn a lot about planning algorithms. We implemented a fairly simple one by ourselves but half of the project was essentially about identifying blocks given some predicates, not about specific algorithms.» (Small extent)
- Interesting» (Large extent)
- Here we got to implement a couple more algorithms and learn PDDL, which was also great fun. I liked the fact that we could learn a wider range of algorithms and implement some of them.» (Large extent)
- We were overdue with project 1 so time was short here, a good project, but lacking the same motivation from me.» (Large extent)
- Learned a lot about planning, that was interesting.» (Large extent)
- I found it very fun and interesting. » (Great extent)

14. How was the workload for the "Shrdlite" project?

Too low»		0		0%
Low»		0		0%
Adequate»		7		29%
High»		12		50%
Too high»		5		20%

Genomsnitt: 3.91

- I pretty much did 80% of the work here. Getting a 3 sucks.» (Adequate)
- This was not as heavy as the Sentimentalizer» (High)
- The fact that is needed a webserver meant more work but didn"t feel related to the course.» (High)
- It was for us very unclear how we should even start. » (Too high)
- Well, I worked alone» (Too high)
- The time was too short to deal with this complex project. Therefore we did not manage to implement a fully working algorithm.» (Too high)

15. How did you allocate your time for this course?

- lectures/sentimentalizer/shrdlite 10/60/30»
- 10% lectures 55% Sentimentalizer 35% Shrdlite»
- 5% Lecture 65% Sentimentaliser 30% Shrdlite»
- - Lectures 5% - Project "Sentimentalizer" 45% - Project "Shrdlite" 50% - Other 0%»
- 1% lectures, 60% Sentimentalizer, 40% Shrdlite.»
- Nearly 0% lectures and then 50/50 throttle per project.»
- - Lectures: 10% - Project "Sentimentalizer": 40% - Project "Shrdlite": 50% - Other: 0%»
- 10 % Lectures, 65 % Sentimentalizer, 25 % Shrdlite»
- 15% Lectures, 45% Sentimentalizer, 40% Shrdlite»
- Lectures: 5 Sentimentalizer: 45 Shrdlite: 50»
- 20? 45 35»
- -Lectures 10% -Sentimentalizer 50% -Shrdlite 30% -Reading chapters from lectures 20%»
- Sentimentalizer 50% and Shrdlite 50%. »
- 20 - Lectures 30 - Project "Sentimentalizer" 50 - Project "Shrdlite" 0 - Other»
- 50% sentimentalizer 45% shrdlite 5% lectures 0% other »
- - 10% - 50% - 40% The reason we spent more time on the first project was because we thought a 30 page long report was a good idea...»
- Lectures 0% Sentimentalizer 65% Shrdlite 34% Other 1%?»
- 50 % sentimentalizer 35 % shrdlite 7.5 % lectures 7.5 % other»
- 5% Lectures 55% Sentimentalizer 35% Shrdlite 5% Other»

16. How well did the course administration, web page, handouts etc work?

Very badly»		1		4%
Rather badly»		2		8%
Rather well»		14		58%
Very well»		7		29%

Genomsnitt: 3.12

- Insufficient instructions on Shrdlite to start work, lost almost a week because of this. No grading on hand ins!!! Why did we get a final grade on everything when the homepage specifies that we were supposed to get several grades on each project. We need more feedback on what we do so we can improve over the course!!!» (Very badly)
- Some things were needlessly difficult to find (such as rooms for the supervision sessions) and the schedule was spread out over at least two places.» (Rather badly)
- Could have some more meetings or lectures about the projects. » (Rather well)
- Some stuff were unclear to start with, but got fixed as people complained, like how to fix the GF stuff» (Rather well)

Study climate

17. How were the opportunities for asking questions and getting help?

Very poor»		0		0%
Rather poor»		2		8%
Rather good»		9		37%
Very good»		13		54%
I did not seek help»		0		0%

Genomsnitt: 3.45

- The supervision sessions were good although they mainly covered general topics, not specific implementation problems we might run into.» (Rather good)
- Supervision was crucial» (Very good)
- Again, kudos to Peter and Prasad!» (Very good)
- Peter good» (Very good)

18. How well has cooperation between you and your fellow students worked?

Very poorly»		2		8%
Rather poorly»		2		8%
Rather well»		6		25%
Very well»		14		58%
I did not seek cooperation»		0		0%

Genomsnitt: 3.33

- Only 1 really wanted to participate, the rest only had excuses as to why they didn"t have enough time to work.» (Very poorly)
- Good in first project though» (Very poorly)
- Communication and workload questions could have been managed better between us.» (Rather poorly)
- We had ongoing discussions with Peter regarding this, it turned out decently in the end» (Rather poorly)
- Some communication problems in my group.» (Rather well)
- Cooperation was good with my fellow student. Although we had some problems at the beginning. One student left the group for personal reason after 3 weeks without any contribution. Another student could not contribute to the project much, because he was lacking programming skills.» (Rather well)
- Apart from differences in ambition (grade) we worked very well together.» (Very well)
- In the group very well but we didn"t seek help from students outside the group. » (Very well)
- As always it"s difficult to divide the work load when some don"t aim as high as the rest.» (Very well)

19. How was the course workload?

Too low»		0		0%
Low»		0		0%
Adequate»		5		20%
High»		11		45%
Too high»		8		33%

Genomsnitt: 4.12

- In reality, in a better group, the workload would have been quite ok.» (High)
- It was hard combining this course with another one. Had to put a lot of time in the first project, making me falling behind in my other course.» (Too high)
- Two projects and reports with overlapping deadlines is just too much for one course.» (Too high)
- Since the need for different grades was very unclear one could very easily spend a lot of time.» (Too high)

20. How was the total workload this study period?

Too low»		0		0%
Low»		0		0%
Adequate»		1		4%
High»		14		58%
Too high»		9		37%

Genomsnitt: 4.33

- I did only this course, as an elective course.» (High)
- I spent almost all of my time this period on this course and I still feel like I didn"t finish eveything I wanted.» (Too high)
- Two project courses. » (Too high)
- The other course also demanded a lot of extra hours of us.» (Too high)

Summarizing questions

21. What is your general impression of the course?

Poor»		2		8%
Fair»		0		0%
Adequate»		6		25%
Good»		13		54%
Excellent»		3		12%

Genomsnitt: 3.62 (bidrar till totalt genomsnitt/jämförelseindex)

- The course as intended was successfully given, but I feel that the intentions should change. I would have learned a lot more with less focus on the projects. To be honest, I feel that I learned very little about both planning and machine learning (even though I got the highest grade in both), and I didn"t learn anything at all about the other topics in the book.» (Poor)
- It could be improved with more relevant lectures and more even workload of the projects. It also feels like you might want to learn more than two things in a course. » (Adequate)
- Make the projects and lectures more focused on learning some very specific set of algorithms and make it harder! Now the projects and lectures focused on too many other things that could have been left out.» (Adequate)
- Good things to learn, aweful examination.» (Adequate)
- The subject is very insteresting, and I learned a lot of things, but it seems too broad of a subject to capture in one period, and the projects were a little too much work.» (Good)
- Might be a good idea to look over the lecture material.» (Good)
- Nice projects» (Good)

22. What should definitely be preserved to next year?

23. What should definitely be changed to next year?

24. Additional comments