So let’s see, we have half a dozen picture sites (ex. Flickr), a half a dozen blog sites (ex. Blogger), half a dozen social network sites (ex. Facebook) and just about everyone has a different collection of sites that they use. Some use Xanga and Facebook, some use Flickr and Blogger, some just like LiveJournal and mySpace. Its such a mess. I think there is a high demand for a portal of sorts which can tie all of these services together into one. Even go as far as tying all of the APIs that are out there into one application.
Imagine: you log in and you see your friends from all the sites that you have accounts in like Flickr, Blogger, Facebook, and mySpace. Then you request more information about a person and you get a meshup version of their profile based on all the services which they use (and you have access to as well). You can then continue to request more information such as calling up googleMaps and so on.
Considering the abundance of information which is out there, if a service could consolidate all of this information in a slick way, it would make heads spin. This would be the holy grail of web 2.0. Its even worth leaving school and working just on this for a semester, perhaps this is the new pot of gold waiting to be discovered.
I have been pondering the possibility of creating a distributed and collaborative system to detect plagiarism in assignments submitted by students in universities. No one can really trust the effectiveness of automated systems such as turnitin.com for as reliable as it is: its far from a perfect system. The lack of trust in automated systems results in graders having to manually ‘double-check’ the results to be sure that the results are reliable. What I propose is a collaborative plagiarism detection system where assignments are checked twice: once by a computer thereby eliminating clear insistences of lack of plagiarism and a second check made collaboratively by those who use such system.
The system will work by allowing graders to check as many sets of assignments as they wish(thought by the automated part of the system to be potentially plagiarized) and decided if it is plagiarized in their opinion or not. The key here however is that the sets shown to the grader should be no longer than a few lines long, this will eliminate bias based on content surrounding the scene of the potential crime. The effectiveness of this will be based on participation. The more graders that exist and the more sets that are evaluated the more effective it will be. Due to the short nature of each set it should take a very short moment for one to evaluate it and therefor creating a fast speedy system for identifying cheaters.
After a very long cup of coffee this morning, an eureka moment took place. The full description of the problem and solution I’m still writting because it is rather tricky to explain. But the turning point was when I remembered that old Physics unit on lenses. Suppose you have an object from a picture with coordinates (x,y), how do you know how much to pan and tilt your head to face the object. The key here is that you are going from a 2D coordinate system to a 3D one in a sense.
Now if I can only finish understanding Optics and Lasers by this weekend, I’ll be set.
Have you ever wandered why cars in the 21st century are sill unable to drive by themselves? Or why pilots are still needed to land airplanes? Or why train conductors still exist? One of the reasons is due to the fact that no method has yet been found to detect objects accurately and in a timely fashion. It doesn’t help if the car detects a kid crossing the street 2 minutes later, because by that time the kid will have become road kill.
My project attempts to solve this problem by combining information from two different sensors; an Infra-red sensor (which is very precise but slow in acquiring data) and a set of color images (which is far from accurate). Objects will than be identified from the Infra-red sensor and the images separately and then merged together to more clearly identify objects in a given environment. The goal is that by merging objects from two different sources the precision will be far greater than any current single sensor algorithm.
Read the rest of this entry »