Tests

Assignments

Submitting all the components of the assignments electronically is encouraged, but anything that does not imply programming can be submitted on paper.

1. Assignment 1 is due on 19 March at midnight (electronic matter) and 18 March in class (anything that is submitted on paper).

Reports and presentations

Unfortunately, there is no way to talk about AI techniques and AI applications all in one term. So the purpose of these reports is to give you as the author (and the whole class when you give your presentation) a glimpse of what are the actual uses of all the techniques we talked about during this course.

Your report (and presentation) could be based on the textbook material, but I expect you to do some independent research regarding the state-of-the art in the area and/or real-world case studies. You will get a passing mark for a report based only on the textbook alright, but not necessarily a good mark.

Your report (as well as your presentation) should normally be organized around the following points:

1. What is/are the problem(s) included in this domain, and why are AI techniques necessary.
2. What are the techniques that we talked about throughout the course and that are used in this particular domain, and how are they modified and/or put together in order to accomodate the particular domain you talk about.
3. Case studies.
4. State-of-the art for the domain.

Some topics are not suitable for such an organization, but most of them are. Do not go into too much details on any point--remember, this is supposed to present an overview of the topic, not an in-depth analysis. You will have about 30 minutes to present your topic, although the report should typically go a bit deeper into details.

Hand in

a written report, which will also be posted on the course's Web page. Provide a format suitable for displaying on the Web (HTML is of course the preferred format, but plain text and PDF are alright; I can also convert various other formats which are thus acceptable; contact me for specifics), together with an open format suitable for printing (such as a hardcopy, PDF, Postscript, LaTeX, etc.; I will not accept Word documents).

Hand in both a hardcopy or a PDF document, and a format suitable for on line posting.

A normal length for your report is ten printed pages using a 10-point font. This is just a guideline though, I will mark the report based on the content, not the length.

Your report must include a suitable list of complete references.

The report has to be handed in by 2 April 2010, except for the case in which you submit hardcopies, which have to be handed in by 1 April. Please follow the usual instructions to submit the electronic part of your work.

Presentation

You will give a presentation of about 30 minutes in class. The presentations are scheduled for 6 April (and 8 April if necessary, two presentations per lecture slot).

Bonuses

All of you, individually, will have the opportunity to mark each presentation as well as each report in whose production you did not participate. I will consider your marks as an advisory. If you attend all the presentations and you provide marks for them, you will get a bonus of 2% in the final grade for this course.

I am interested to hear about the following topics:

1. Planning (Chapter 11, maybe part of Chapter 12).
2. Utility theory (Chapter 16, maybe part of CHapter 17)
3. Learning:
   1. From observations (Chapter 18)
   2. Reinforcement (Chapter 21)
   3. Neural networks (in Chapter 20, a whole bunch of documentation also available)
4. Natural language processing
   1. Grammar-based (Chapter 22)
   2. Probabilistic (Chapter 23)
5. TAKEN Game playing, with chess and Deep Blue as case study (part of Chapter 6 and independent research)

Mid-term examination

The mid-term examination was Thursday, 11 February in class. Here are my answers. The average grade (excluding the student who did not write the exam) was 13.25 (66.25%).

The mid-term examination covers state space search, CSP, and propositional logic. The exercises in the textbook do make for good practice. Here is a collection of suggested exercises that also outlines the major directions for the mid-term questions.

Submission procedures

Recall that you are allowed to collaborate in producing your reports in groups of no more than two, and that all collaborators must be named and must be currently taking the course. The assignments are however individual.

The assignments should be handed in electronically, as one directory which in turn contains your work for the given assignment. Do not create any kind of archive; submit the directory itself, unarchived. The name of the submitted directory need not follow any particular format; it will be automatically prefixed by the name of your account at submission time.

Hand in your code and any other documentation you are required to provide or you feel like providing. Any document must be submitted in an open (non-proprietary) format. In particular, Microsoft Word or Wordperfect files are not acceptable. The preferred format is plain text, but others are fine, including but not limited to HTML, Postscript, and TEX (L^ATEX included). Good ol' hardcopies are also accepted. :-) Anything submitted as hardcopy should contain on top of the first page the name of all the collaborators.

Also provide a readme file containing the name and student number of all the collaborators, and the email(s) you want the marks and comments sent back to.

How to submit

Submit a directory containing all your work using the submit command available on the J117 machines. Use csc216 as course name. Type man submit for more details. Alternatively, you can submit by email. In such a case please archive your submission so that it expands into a subdirectory and submit the archive.