Messages
Johan worked through parts of the night, and we just registered the results on our second obligatory assignment. Seven students submitted for problem set (2b), and all seven are qualified for the exam. We are not unhappy with the final retainment rate this year: 11 students had submitted something for our first problem set, (1a), and the same group of seven who now qualified for the exam had remained active after completion of the first obligatory assignment. I expect the faculty administration will publish the exam location in the course of today. Jonathon will be coming around half an hour into th exam on Monday, to answer clarification questions. Good luck!
Some of you (rightly) are beginning to wonder about the location of our final exam (this coming Monday, December 5, 14:30 to 18:30). Given the late submission deadline for our final problem set, exam registration will only be finalized tomorrow afternoon (Friday, December 2). Once you are qualified for the exam, you should be able to find information about the exam location in StudentWeb, hopefully still before the weekend, but the latest on Monday morning. Good luck with wrapping everything up!
Our last and final Problem Set (2b) is now available; the starting package for this assignment (which will be automagically dropped into your IFI home directory) also serves as our model solution for the previous problem set. We strongly encourage everyone to look this over and start work on the assignment in the next few days and in any case in advance of our laboratory session, this coming Thursday.
I hit a snag yesterday, when preparing our final problem set (2b), which is designed to build on the model solution for problem set (2a) that we (in part) reviewed in the lecture yesterday. My apologies for the delay in making these available, but I will make sure they go on-line before the end of the day today (Saturday). While it is bright and sunny today, the forecast for Sunday is less rosy; hence, maybe budget some time to start work on our final problem set tomorrow?
Following the discussion in the laboratory earlier today, we just published our model solution for problem set (1c); please see the detailed course programme for the link. The theory and implementation of estimating HMMs and of computing probabilities within the HMM framework is a topic of central importance, so even if you maybe did not work through this problem set to the very end, we would much recommend you review our model solution carefully and make sure you understand the design decisions (and trade-offs involved) made in choosing data structures and procedures.
Finally, please recall that tomorrow (Friday, October 18), we are honored to have Dr. Erik Velldal drop in for a guest lecture, who will open for you a treasure chest of advanced elements in Common Lisp, some to this very date unique in the world of programming languages.
It is being pointed out to us that the machine windows.ifi.uio.no has been non-responsive for at least part of tonight. Because we realize that some of you go through the Remote Desktop interface to work on our assignments, we extend the submission deadline for problem set (1c) by 24 hours, that is until the end of the day tomorrow, Tuesday, October 25.
We have just posted problem set (1c) and at the same time moved back the submission deadline to October 24. Next week, we will not have a lecture on Friday (that is on October 21), because of conference-related travel. Note however, that the Thursday laboratory sessions will be offered as usual; please see the detailed course schedule (updated as of today) for further information.
We have posted a model solution for problem set (1b), which we will discuss in depth during the laboratory session tomorrow. Please have a closer look at our code, compare to your own, and come to the laboratory tomorrow, bright and fresh, and with critical opinions about our solution.
Apologies, we were late this week in releasing a model solution for problem set (1a) and the next part of the first obligatory assignment, problem set (1b); both are now linked from the detailed course schedule (see the laboratory sections for this week and the following one). In the laboratory tomorrow, we will walk you through selected parts of our model solution, discuss contrastively your solutions and ours, and answer questions of all kinds. In doing so, we will comment more on some of the Lisp constructs we learned in the previous lecture. On Friday, we will lecture a little more about Lisp data structures and finer points of (local) variable binding, and then move on into our first (basic) search problem: a recognition procedure for non-deterministic finite-state automata.
Please observe that we managed to switch laboratory rooms, in order to have a native Linux environment: starting already tomorrow (September 1), we will be in room Fortress (3468), just a few doors down the corridor from our originally assigned room (Chill).
Our first problem set, assignment (1a), is now available. The text includes instructions for how to get going with our Lisp development environment at IFI. The main purpose of this assignment is to get going with simple Lisp programming and the use of emacs (for some, large parts of this problem set may seem near-trivial). Please have a go at this assignment prior to our first laboratory session on Thursday this week! Ideally, you will be near-finished when we meet in the laboratory, and we will be there to help you with any remaining issues.
The course programme will be largely similar to the 2009 and 2010 versions, albeit with a number of smaller adjustments: (weighted) search problems, dynamic programming (in Common Lisp), training and decoding of statistical models, and generally implementation techniques that scale to potentially very large data sets. Within the next few weeks, please obtain a copy of the NLP textbook we will assume for our theoretical foundations: Jurafsky & Martin (2008). Any other teaching materials (notably the Common Lisp book by Seibel (2005), as well as other Lisp references) are available on-line. All teaching and laboratory supervision will be in English. The first lecture is on Friday, August 26, at 10:15–12:00, in room Python (2269) in Ole-Johan Dahls hus. Please note that participation in the first lecture is obligatory.