Messages

We submitted exam results to the department last night, so they should (I hope) become visible for you on-line today. 23 students took the exam, and the outcome is somewhat polarized: three As, seven Bs, and one C contrast with three Fs, four Es, and five Ds. Because we felt the exam was maybe a little too long for the three hours you had available, we lowered the grade criteria by five percent. In other words, the minimum requirement to pass was 35% of the total points available. The second examinor was Lilja ?vrelid. In case you want to review your exam scores in detail, please contact the department administration. Stephan will be away from UiO until mid-August now, hence it will take a little while before we can jointly look things over; however, anyone who would like an exam review should register their request with the department as soon as possible (i.e. within the formal deadline). God sommer!

In the lecture today, we agreed to start half an hour early tomorrow, i.e. at 9:45 tomorrow, in order to have more time to talk through the exam. Please make sure to arrive on-time: in the beginning of the session, will cast a vote on which sub-problems from the sample exams to discuss jointly; and then we will start looking at exam questions in order of decreasing difficulty.

Because some students are tied up in an exam until 13:00 tomorrow, we will start the last lecture one hour later than usual, i.e. at 13:15 (in our usual room). If possible, I will want to go on until around 14:30; please let me know before the start of the lecture, in case that would cause someone bad trouble. In case you have wrap-up (not exam-related) questions you would like me to address in this session, please email!

Some of you are wondering about results on obligatory assignments, which is a natural thing to feel curious about these days. As for the second obligatory assignment (problem sets 2a and 2b), while both Arne and I (and my daughter) were ill in April, I had taken over grading of assignment 2a; the results were registered in the IFI on-line system about ten days ago, and from what I remember everyone who had submitted both 2a and 2b received enough points. I did not have the time to send out individual comments, however; so please check on-line for your results. As regards the final obligatory assignment, i.e. problem sets 3a and 3b, Arne is busy grading your submissions as we speak. We will make sure to have registered results in the on-line system by 16:00 tomorrow (Thursday, May 27), and then Arne will email individual feedback on Friday. Because one student was granted a deadline extension (due to illness), we cannot publish a...

As of this afternoon, the final problem set (Exercise 3b) and model solution for the previous assignment (3a) are linked from the course schedule. To obtain the model solution, at the IFI Linux environment, you can also run: solution3a. In the laboratory tomorrow, we will go through some fine points of lexical rules, and then jointly start working on adding semantics to our grammars.

Two updates on laboratories and the upcoming submission of problem set (3a). We managed to get a reservation on C207 (our usual laboratory room) for the extraordinary laboratory tomorrow, from 12:00 to 14:00. Because we are a little behind schedule, we postponed the submission deadline by two days, to Wednesday next Week, May 5.

Arne is regrettably sick this week, hence feedback on your submissions to problem set (2a) is delayed; Stephan will be holding the laboratory session tomorrow. We will aim to get feedback on your previous submissions to you as soon as possible, but this will likely not be in time for the upcoming submission deadline for problem set (2b). Note that there are several optional parts towards the end of (2b), so it is in principle possible to score up to 120 points on this exercise alone (and thus pass on the second obligatory assignment). In case you wanted to work out your expected points for (2a), the allocation of points to parts (1) through (4) of problem set (2a) is as follows: 30, 20, 35, and 15.

I sent out an email to everyone a little earlier today, with instructions on how to work around the problems in selecting files to load into the LKB that recently affect some of you (but, weirdly, not everyone). Please make sure you have received this message, which is the first time we use the mailing list for this class; in case you are enrolled in the class but did not receive email from me today, please let me know.

I also noticed there was a typo in the exercise, near the end, about how to submit your work. Since we are now working on problem set (2b), the command-line incantation for packing up and delivering your work to us should be: submit exercise2b.

Finally, the current problem set deliberately throws you out 'into the deep end' (of the swimming pool, I believe the metaphor goes), i.e. we ask you to work with the new TDL language, unification grammar, and LKB debugging support already before we have formally introduced everything there is to kno...

In case you want to use the ¡®automagic submission¡¯ procedure we mention in the current problem set: at the shell prompt, and assuming your current working directory is the parent to the ¡®exercise2a¡¯ folder containing your work, try the following: submit exercise2a. There should be messages about all the files that are packaged up and, finnaly, a confirmation about successfil delivery. In case you are in doubt about this method, please email both Arne and Stephan. God P?ske!

Note the upcoming deadline for problem set (2a) tomorrow; please make sure to take advantage of the laboratory session tomorrow, to complete your implementation of the CKY parser, debug, and document. Also, my apologies for being late in sending out invitations to the on-line questionnaire that we will use for mid-term evaluation of the class; you will receive email about this later today.

We have now completed the course schedule for the second half of the semester, including obligatory readings for the first few weeks after Easter, all remaining lecture dates (laboratory dates are yet to be filled in, but they should largely be predictable), and submission deadlines for the remaining three problem sets (2b, 3a, and 3b). In case any of these dates were problematic in terms of other known scheduling constraints, please let us know asap.

I had used an out-of-date link when publishing slides copies from the most recent lecture; the link from the course schedule is correct as of tonight.

We have published a model solution for Exercise 1b today, which we will discuss in detail during the laboratory tomorrow. Please review our solutions beforehand and compare to your own. Those who identify non-trivial errors in our code during the session tomorrow will be awarded between five and ten bonus points. Also, the next problem set is now available; please print and read through this before the laboratory tomorrow. The submission deadline has been pushed back a little, to Friday, March 26 (the last day before the Easter break).

We have now published a model solution for our first problem set. We will review our solution in some detail in the laboratory tomorrow; we will also (finally) demonstrate some of the extra functionality in the interface between emacs(1) and the Lisp system. In preparing for the laboratory, please compare our solution to your own and see which you like better.

It took a little while, but today the first bug in our first problem set was found: in question (9d), the last example of a single capital letter followed by a period should be ¡®?.¡¯, i.e. the capital Norwegian ? was lost in our original version of the problem set. We have now posted a corrected version to the course web page. Please remember that the submission deadline for this problem set is coming up on Monday, February 15. Good luck with completing this assignment!

Both in the lecture today and the laboratory tomorrow, there will be a few additional slides on using Lisp, everything you need to know to complete the current problem set. As a general remark, one more time: we will be using a very restricted subset of Lisp in this class, and we aim to introduce all the necessary concepts, data types, and functions on the slides. Hence, in solving the problem sets, please review the slide copies as your primary reference.

A note on communication outside the lecture and laboratory hours: we welcome any kind of feedback or questions. Specifically, do not hesitate to email us when you need help, for example when you get stuck while working on one of the problem sets. Please email both Stephan and Arne, and include as much information about the problem as possible (for example a function definition that you think should work, plus information about how you call the function and what exactly you find surprising in its behavior). For other kinds of feedback, there is also a web-based form to submit comments anonymously; everyone registered for the class should have received an email invitation to this service, with information about how to access it. If not, please contact Stephan (it would indicate that you appear to not be registered for the class).

At least two students in INF2820 are also registered for INF5830 (Semantic Technologies); at present both courses have their final exam at the exact same time. I have asked IFI administration to request rescheduling, and it is very likely this will work out.

I have updated the course web pages today, so please check out the schedule for the next few weeks and information regarding Obligatory Exercises. Also, until the very end of the lecture today, we thought obligatory registration of presence would not happen today, but actually it did, as people were leaving the room. If you left class early today or were unable to participate today, please make sure to re-register for this class. There are spaces available, and on-line registration will be open starting tomorrow. Finally, please note that our first laboratory session will be this coming week (Friday, February 5).

Due to illness, the first lecture today must be cancelled. This means we will push back the entire course schedule by one week, including the obligatory registration of presence. In other words, the first lecture will be Thursday next week, January 28, and everyone needs to be present to confirm registration for this class. My apologies for the late notice!

We have posted the list of readings for this class (please find the link on the top-level course page for this semester) and ask everyone to obtain copies of the two text books we will be using: Jurafsky & Martin (2008) and Sag, Wasow, & Bender (2003). Both should be available in sufficient quantities from Akademika, and both are books that we expect will be valuable to you in later years of study too.

This course if open to international students. All teaching materials are in English, and instruction will be in English too. The first obligatory meeting is on January 21. Please contact Stephan Oepen (by email or phone) beforehand, in case you need additional information beforehand.