|
|
|
Mini Project 1
[ CSE 573 2004 Autumn Home | Problem Sets | Mini Project 2 ]
Projects will be done in groups of two. I detail the two default projects below, but groups can always ask me if they have their own idea that they wish to explore.
The default plans are:
- Build two satisifiability solvers (e.g. DPLL and WalkSAT) and experiment with several enhancements. More information here.
- Build a program which uses min-max search to play your favorite game and experiment with enhancements. More information here.
- Friday (October 15) Email Dan and Xu with the name of your team, the names of the teammates, and a brief project plan (which project, and a week-by-week breakdown of what you expect to do that week).
- Wednesday (November 10) Final report due.
- Reports submitted by all teams
Each group is welcome to talk to other people about their project, including other groups in the class. Of course I hope that you will read some research papers on the topic. And you are also welcome to search the web for ideas and even bits of code. However, it is extremely important for you to identify (in your final report) whom you talked to (and about what), where ideas came from (if not from your head), and what code functionality you wrote versus acquired.
We will look at your code, but the most important deliverable is your project report, so don't forget to allocate sufficient time to make this your best work. The ideal goal to aspire to is a AAAI conference submission. Given the time constraints, novel, publishable results are unlikely, but see how close you can get. Your report should emulate (as best you can) aspects of a AAAI paper such as motivation, problem definition, tone, style, citations, graphical presentations of your results. Remember: even if you don't have novel ideas you can still phrase your report as duplicating (and verifying) a prior experiment, result, or claim. And if you try an new idea which doesn't work well, you can phrase the failure as a (negative) result.
What to turn in:- The code you wrote: e.g. board functions, search algorithms, heuristics, solver algorithms, problem generator(s), and any other code you used to run experiments. The code should be clear and reasonably documented with a brief user manual). Also include instructions for building the executables and any command line options the grading program should use. You may email this to Miao.
-
An article describing your proposals, experiments and results. This
article should be written as a research article for submission to a
conference such as AAAI's national conference. It should have a
maximum of 10 letter-sized pages in 12pt font with 1 inch margins,
including all tables and figures, but excluding references and the
appendix requested below. As you are aiming to approximate a AAAI
conference paper, note that such a research article typically has
sections describing: the problem it addresses and its motivation; the
new solution(s) proposed or questions answered, and the rationale for
them; empirical and/or theoretical evidence for the claims; a
discussion of relations between any new methods and previous research,
including a candid description of your approach's limitations and
disadvantages; and directions for future research. Citations made in
the body of the paper should collected in a list of references at the
end. If your approach didn't perform well, you can propose (and
possibly test) your explanation in the empirical and/or discussion
sections.
Your report should include an appendix describing what parts of the project were done by each of the two group members. This appendix should also acknowledge technical conversations with and ideas from people outside the group. The functionality and origin of any external code should also be discussed here.