Let's pretend for a little bit. Suppose that you are me - the professor for last semester's AI course and now this semester's Intelligent Systems course. A student comes to talk to me (who is now YOU) and the following conversation occurs:
Prof : What do you think of this class?
Student : I kind of like it, but it seems like we do a lot of reading on things unrelated to our actual project.
Prof : Yeah, but don't you think that is helpful? It helps you understand your project and perhaps apply some of the things we discussed in AI?
Student : Really? Heck, I was even wondering why we needed AI before we took this class. It seems to me that you could simply hand me this project [insert your project here] and I could have done this without any knowledge of advanced or even simple AI techniques.
How would you, the Prof, respond to this statement?
Write a two- to three-page essay on why it might be important to have a broad based understanding of AI-techniques before attempting to write a large-scale intelligent system like the one you are working on this semester.
The process:
Think about the problem that you (the student) are about to solve.
Take some time to go over the design specification that you have just turned in as your mid-term deliverable. Think carefully about the problem as a whole and about the code that you are about to write in order to create a system to solve the problem as a whole.
Think about the variety of techniques we have discussed over the last semester and a half?
You might start by reviewing the four main units we talked about in AI last semester. Then you might go over the various readings that we have discussed this semester both from the AI proceedings you wrote in the fall through the specialized readings each team proposed and presented. What do they have in common? What do they offer you as a student? What do they offer you as a system-builder? Why should JARM Flix care about NLP if it isn't part of their project? Why should Team Rodan care about advanced recurrence networks or Monte Carlo methods for generating Bayesian values?
Write your essay
No detail is too small; no idea, too big.