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:

  1. Name(s)
  2. Project name (should be descriptive)
  3. Book chapter selected
  4. Name of Sample program, if any, you plan to work from.
  5. New features or ideas you wish to explore.
  6. How the work will be divided, if you are working in a partnership.
  7. (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