Date: Mon, 02 Dec 1996 15:59:27 GMT
Server: NCSA/1.4.2
Content-type: text/html
CSE 473 PROJECT
CSE 473 Project
Topic descriptions due Friday, May 17.
Preliminary demonstrations due Friday, May 24.
Final demonstrations and reports due Friday, May 31.
Rationale
To do justice to more than just a few topics in artificial intelligence
would require more than a single 3-credit, one-quarter course.
The project offers you an opportunity to choose for yourself one topic for
in-depth examination. This can be one that we don't get to cover
in the course, or it can be one we have already covered.
You have the flexibility to work either alone or in a partnership
with another student in the class. You may be able to apply some of
the representation and inference techniques we have studied
so far. In any case, you can put to good use the fluency in Lisp
that you have developed over the past few weeks.
Topics
You may select any of Chapters 5-14
of the text as your topic area. Within the chapter, select
a specific topic, such as the use of augmented transition
networks in the Stone World program within Natural Language
Understanding in Chapter 11.
If you choose a topic that we have already covered in class,
then your project should explore that topic in greater depth.
If you choose a topic that we haven't gotten to, then you may
wish to use one of the sample programs in the text as a starting
point, and then implement additional features and perform some
experiments.
On Friday, May 17, turn in a project topic description that includes
the following:
- Name(s)
- Project name (should be descriptive)
- Book chapter selected
- Name of Sample program, if any, you plan to work from.
- New features or ideas you wish to explore.
- How the work will be divided, if you are working in a partnership.
- (Not required, but strongly encouraged:)
A bibliographic reference (other than the text)
for additional reading on your
chosen topic. Give a two or three sentence summary of what this
reference contains that is relevant to your project.
Whether you select machine learning, natural-language understanding,
vision, or neural networks, you may wish to design your project
to work with input data from the World-Wide Web.
For example, you may wish to train a neural network to
distinguish between different types of HTML files, or you may wish
to do pattern recognition experiments on images downloaded from
various online galleries or home pages.
Jeremy and I strongly suggest that your project be executed in
a way that can help others learn something about AI. To make
your project useful in this regard, give informal demonstrations
to classmates from time to time, and ask them for suggestions
on how to make the program better or more understandable.
Preliminary Demonstrations
Be prepared to show a Lisp program on Friday, May 24, that
does something interesting. If you are using a sample program
from the text, your demo should include some novel new feature(s).
Alternatively, instead of giving a demo at this time, you
may turn in a 1-page progress report in class.
Final Demonstrations and Reports
On May 31, turn in both a diskette with your Lisp program,
and a hardcopy report on your project. Here are the guidelines
for
the form of the report.
tanimoto@cs.washington.edu