Lab 3: Getting into Java and AI
There are two options for Lab 3. The first is the "easiest"
- Using the Watson text, get the code working for the
Natural Language chapters. This would be Chapter 2 (the GUI), Chapter 6,
and Chapter 13 (which needs Chapter 3). Your code must be usable (with no deprecated code)
on either the latest version 1 or preferably java 1.2 (since you should
always use the newest version). Those of you who have never used much
java should probably choose this one.
Problems that unseasoned java programmers will face
- event handlers - the delegation model (components/interface "listeners")
- package understanding ... why the mwa.ai.nlp.History? Can I just comment out the package declaration in the file? Answer: It is better to understand something than ignore it. I will address in class.
- Client/Server "applets" - problems with browser security? Solution
- appletviewer
- make applications
Use the Structured Matcher code provided from the OO Programming
for AI course in the past and implement a tool for Classification using
the idea of Specialists. The Structured Matcher will do the matching
for you, you need to consider the organization for solving the
task of Classification. I will go over this explicitly in
class. The basic structure is actually quite straight forward. You can
also use as much of the past work on the Classification tools as you want.
Both of these should be submitted via a web page with the access
to runnable code and a link to a README providing discussion
about your implementation problems/solutions, concerns about code you
used, improvements that could be made, Where's the AI?, etc.