|
Soon, I will be needing to turn in a proposal for a senior project I will be having to work on all semester long. I have a couple ideas of what I want to do, however, I do not know whether they'll prove to be big or good enough to be called a "Senior Project".
I want to make an AI program for a game of checkers, as well as make its respective GUI, but I do not know if that will suffice. I somehow see the problem to be relatively easy yet I could still write a 30 page report and give a 30 minute presentation on the entire project's life-cycle.
If for some reason this ends up not being complex enough for a senior project, I was thinking of turning it into a self-learning checkers AI. However, that would require having knowledge in some area such as Neural Networks or Machine learning, both which are grad. school courses, to be able to carry out its development.
I'd like to know what you guys think and if you'd consider my first proposal to be good enough to be my senior project! I will be graduating this December and I want to do something different from everyone else around here (web applets).
I know I can make a Chess or a Go AI and that Checkers is "easy" compared to those board games and that it was also already solved a while ago and all of that good stuff. But I want to keep this last semester as easy as I possibly can given I'm also taking 5 other courses T_T.
Cheers!
  
|
|
|
Hong Kong20321 Posts
|
United States24665 Posts
Checkers with AI on a modified board might be more interesting.
|
i've heard that making a go AI is very very very hard.
|
United States24665 Posts
On August 31 2010 00:29 intoyourrainbOW wrote:
i've heard that making a go AI is very very very hard.
By hard I assume you mean hard to make it as competitive as the other top GO ai's. After all the best computer go players can barely beat other mid-level amateurs haha.
|
There's some decent lectures on youtube about neural networking and machine learning if you decide to go the second route. Whichever you choose, good luck!
|
I think it depends on how much of the project you want to be the gui. A really nice gui will be more work than a good checkers ai. You won't make hardly any progress on a go ai in one semester if you are going for anything beyond a 9x9 board. I did a go learning ai my senior year of college (after having spent my junior year in Japan playing a ton of go) and it was not impressive at all given how much work I put into it.
If you really are interested in ai, especially if you plan to go to grad school, do the self learning checkers. You'll gain a lot of knowledge from it and the presentation will be very easy since the project will have a really natural progression.
Good luck!
|
EsX_Raptor, where do you go to school?
|
go is very hard. there are too many possible moves.
|
Do it on the new SC2 ai, how the computer opponent no longer 'cheats'. It has to scout you and based on what you are doing that it sees, it plays accordingly... that's amazing.
|
the AI class in my school had to make a checkers AI as one of their projects for the class, so there's definitely no way that checkers is hard enough for a senior project. I don't know how your school does senior projects, but I thought the point of senior projects was to do something that hasn't been done before. Not necessarily come up with a completely new idea, but maybe to do something in a new way, or apply something differently than it was being used before. Also, my school required senior projects to have a hardware and software component, so a checkers AI and GUI wouldn't fall under that criteria. Again, I don't know what your school's requirements are for a senior project, but a checkers AI is not a new or original idea at all and can be coded in just a couple of weeks.
|
United States4053 Posts
I can tell you, making a checkers AI probably isn't incredibly difficult. My friend made a very challenging chess AI by assigning each piece a point value and telling it to perform the move that led to the highest net point gain after 5 turns.
|
I had a senior project last year. I built a laser harp. If your current plan doesn't work out, you could consider that. Although it didn't take us all semester, and the financing of it is somewhat of a problem.
|
|
|
|
|
|
EPT
