It appears that exam results were lost in the internal mail between Forskningsparken and IFI. My apologies for the delay; I have now re-sent those results, and hence they should really become visible for you in a day or two.
We have completed grading the final exams (Christer Johansson, UiB served as the second examiner), and results should show up in the usual places in the next few days. In the meantime, please fill in the on-line course evaluation form for this class! It appeared that I had given you a few more questions than was feasible for a three-hour exam; few people gave answers to all sub-parts of the exam, and noone came close to a complete score. Hence, we adjusted the mapping from total points to grades accordingly, and in the end everyone passed. Out of the seven who took the exam, there was one A, three Bs, one C, one D, and one E. In case you want to look closer at your results, please make contact with me. Finally, please consider working as a tutor for INF2820 this spring; God jul!
At last, we have published the sample exam that will provide the basis for our final, preparatory lecture this week. Note that I would maybe consider this exam a little more voluminous than I would expect to give you for a real, three-hour exam. I recommend that you print the sample exam, first work on it without consulting any external tools, and only afterwards look up what you would have had to know to answer all questions. Our final meeting has now been scheduled for Friday, November 28, from 10:15 -- 12:00. We will meet in room 109 in Vilhelm Bjerknes hus, i.e. the first floor of the building where we had our laboraties on Mondays.
The fourth and final model solution is now available. We will discuss this in some detail during the laboratory on Monday. Please note that our model solutions are exam-relevant. Not so much in terms of specific coding details of the implementation, but rather in terms of general design decisions, for example the choices of data and control structures.
Lars will email more details, but just to announce quickly that we have now registered assignment results in the on-line university system. Essentially, everyone who had qualified before this past problem set, and everyone who submitted results for the final assignment have accumulated the minimum number of points to take the exam. We will post our model solution prior to the last laboratory session, this coming Monday. Also, we need to reschedule our last lecture, because IFI holds an off-campus internal seminar that day. I will email more, but the lecture will have to be Wednesday afternoon next week, or sometime on Thursday or Friday. Watch your email regarding this rescheduling.
We have now released the revised and extended fourth problem set. This version corrects the minor errors announced earlier today, and it adds the additional, third part of the assignment.
There are (at least) two corrections to be made to our latest problem set. First, the abstract learning Goals stated at the top of the page are accidentally pasted over from the previous assignment; our actual goals with this problem set are to understand better chart parsing, PCFGs, and graph unification. Second, in part (2d) we use the verb adores instead of saw; only the latter verb is actually known to our grammar, so please adjust the example accordingly.
And here it is, our fourth and finalproblem set. Please note that we have pushed back the submission deadline for this exercise to Thursday, November 20. We will make available feedback on whether or not people have qualified for the exam early the following day, i.e. before noon on Friday, November 21. This is important because that Friday is the last day to un-register from the class and the exam (trekkfrist). Finally, also note that there will be a second part to this problem set, to be posted on the course web site before our next laboratory session. Please start work on this problem set immediately; we have already covered everything you need to part (2), which accounts for 40 percent of the total points associated with this assignment.
Finally, the model solution for the third problem set is now on-line; my apologies for the delay in posting this! We will review this code in some detail during the laboratory session today. The fourth and final problem set is almost ready for release and will be posted later tonight or mid-day tomorrow.
A summary of responses to our mid-way evaluation is now available. All in all, the feedback was maybe a little more positive than I had anticipated. I will aim to tilt a little more towards the use of slides and put some additional effort in making explicit where the various techniques we learn are used in the real world. In terms of difficulty of the first two problem sets, it appears that things are about the way I think they should be. I will be happy to discuss these results more, during class hours (if you feel there is common interest) or individually.
We had to re-schedule our lecture for the coming week: instead of Tuesday, October 28, we will meet on Friday, October 31. The lecture will be from 10:15 to 12:00, in Helga Engs hus, Seminarrom 232. The laboratory session, on the other hand, will meet as usual, this coming Monday. Please remember that the third problem set is due on Tuesday, October 28, before 12:00 noon.
The third problem set is now available. As announced earlier, the submission deadline for this problem set is Tuesday, October 28 (before 12:00 noon). Please print a copy of the problem set prior to our next laboratory session on Monday, read through the entire document, solve the theoretical part (2), and start implementing the HMM data structure. The laboratory session will be a lot more productive if everyone is familiar with the problem set beforehand (and Lars and I will be distinctly grumpy if people are not)!
As is common at UiO, we ask for student feedback on the course this far. Such feedback will help us adjust the course better to your expectation (for the second half of this semester) and, more generally, will help improve the course for future years. Those who were not present for the lecture today, please retrieve a brief, anonymous questionaire and use a few minutes to reflect on your course experience this far. We will summarize results on Tuesday, October 14, hence please return additional questionaires to Jan Tore, Lars, or Stephan no later than Monday, October 13. For example, bring your completed questionaire to the next laboratory, or just drop it into our mailbox at IFI (just hand it to the receptionist for the administration) or the one at FP (on the fourth floor, right above the 100m-skogen student laboratory).
Please remember that our second obligatory assignment is due tomorrow (Tuesday, November 7) before the lecture. As part of the lecture tomorrow, we will conduct a half-way course evaluation, asking for your feedback on lectures and exercises so far, and suggestions for making the course more effective (or, maybe, more fun). It is desirable that as many as possible contribute their experience and opinion, hence we would be especially happy if many of you showed up for the lecture tomorrow.
The files brown.lsp and krugman.txt mentioned towards the end of the second problem set are now available in our course setup on the IFI Linux machines, i.e. they should magivally appear in your home directory next time you log in. For those not using our course environment, all files needed for this asssignment are available as a single tar(1) archive as well.
We have managed to re-schedule our laboratory slot and keep the same room (C210 in Vilhelm Bjerknes hus). Starting this coming Monday already, September 22, we will meet from 16:15 to 18:00. In the next laboratory, we will introduce a few more Lisp concepts (abstract data types; see the slide copies from this week), jointly review some of the solutions to the first problem set, and then start work on the second problem set. We have fallen behind the laboratory schedule a little, hence the second problem set will have a new submission deadline of 12:00 noon on Tuesday, October 7. We will post both a model solution and the next problem set to the course page before our next laboratory session this coming Monday.
We will most likely re-schedule our laboratory sessions to Mondays, 16:00 -- 18:00, as discussed in class today. Please email Stephan immediately, and no later than noon tomorrow (Wednesday), in case this time is not possible for you. Also, I have started using the email list for the course, and everyone should have received two emails from me (regarding re-scheduling and completion of the first problem set) in the past twenty four hours. Should you not have received these emails from Stephan, please contact the IFI student administration (to investigate why you are not subscribed to the INF4820 email list).
Because some people did not get a chance to attend the laboratory session today, we will push back the submission deadline for the first obligatory problem set by a few days: the new and final deadline is Friday, September 19, before 12:00 noon.
Due to the scheduling conflict of our Monday laboratory slot, we will try to find another slot that works for everyone during the lecture tomorrow (Tuesday). Hence, it is important that everyone come for the lecture tomorrow. If you cannot make it, please email me before the lecture with an overview of all times, Monday to Friday, between 8:00 and 18:00 each day, that are impossible (i.e. due to other classes, but not because of a general aversion to early hours, for example) for you. Re-scheduling at this point will not be easy, but we should give it our best shot!
There will be a small amount of lecturing (on input and output in Lisp) during the laboratory session today, plus a bit of a demonstration of the integrated development facilities in emacs. For the rest of the session, we will work through our firstobligatoryproblemset, which is due tomorrow.
Regarding the assigned reading for this coming Tuesday, Akademia announces today that, finally, copies of Jurafsky & Martin (2008) will be in store, starting around noon this coming Monday. Please make sure to get your copy of the book next week, though for your convenience we also provide an electronic copy of Chapter 2 of the book.
Finally, our first problem set is now available. The Lisp system should be ready to use on all IFI Linux machines, following the instructions on the exercise sheet. In case something does not work for you the way (you think) it should, please let us know, for example by email even before our first laboratory session this coming Monday. Godhelg!
My apologies, I am late with releasing our first problem set (including instructions on how to run the Lisp system at IFI). Watch this space, we will definitely post things tomorrow (i.e. Thursday, September 4).