Each student enrolled in CSE 415 will complete a project, either
as an individual, or as part of a team of two students.
This is an opportunity to choose one of the application areas of
AI and get into it in depth.
Here are
the guidelines:
- Create an "intelligent agent" that demonstrates one of
the techniques from chapters 7-14 of the text.
- Implement the agent in Common Lisp, following a common framework
for use by the whole class.
- The interface should be conversational, along the lines of those
used by the SHRINK, LINNEUS and Stone World programs in the text.
Dates:
- By May 22, come up with an idea for a possible project, and share
it with prospective partners or, if you want, two or three random
people in the class. Post your idea by midnight on the
ideas
web page.
- By May 25 at midnight, submit a project proposal using the
project proposal
form.
The proposal should identify: names of teammates,
proposed title of project, 4 or 5 line scenario (i.e., fragment of
a sample conversation that you HOPE your agent will be able to carry
on), plan for what each team member will do.
- By May 30 at midnight.
Brief progress reports -- post on the web.
- June 5. Demonstrations in class.
- June 10. Project writeups due at 4:30 PM as hardcopy. Either deliver
them to Sieg 314 by that time, or turn them in at the (planned) review
session (tentatively scheduled to take place 4:30-5:30 Monday, June 10).
|
Project writeups should include the following...
- Names of each team member.
- Project title.
- Name or description of the main technique being used.
- Technical description (of approximately one page) of your project,
explaining how it processes user input and produces its answers or
responses. Explain any adaptations you may have made to the technique.
Also describe the range of inputs that your agent can handle.
- Simple example: Give a transcript of a short session with your
agent (approx. three to four exchanges, for most projects, but a complete
session in any case), and then explain each item in detail.
- Longer example: Give a transcript for a longer session that
shows off most of the features of your system. Point out any particularly
interesting behavior, if you wish.
- Code excerpt. Present about one page of Lisp code that you feel
is representative of your contribution to the project. (Don't present
code from the book or the Internet). Explain what part of the agent
is implemented with this code, and describe briefly how the code works.
- Retrospective: Each team member should make a short statement
(e.g., one to three sentences) about what was learned in doing the project.
- References: List any references used, including any web sites,
texts, etc.
- Demo instructions: Give instructions for loading and running
your agent.
|
