Better Artificial intelligence

Team: 44

School: LAS CRUCES HIGH

Area of Science: computer


Abstract: Jeremy Wilson
Jorge Palma
Daniel Cooper
Hyun-Seok Nam
Better Artificial Intelligence (AI)
Problem:
Within our group we have been very interested in AI. We believe AI, in recent days, has not been keeping up with the rest of the advancements of technology. We intend to construct a better AI program which will be able to interact successfully with humans and other objects.
Purpose:
Our aim, for the time being, is to create an AI program that will be able to learn like a 5 year old child. Approaching the problem this way means we will model the brain as opposed to using one of the available artificial intelligence languages.
Procedure:
First we are going to research what other people have done with AI. After we do that we plan on coding a miniature version of the final program which will include a small camera, a simple java program, and some test objects. The program will take a simple and small picture of a solid object then compare it to a previously saved picture. If the picture does not match any saved picture, it will be saved along with a name. At this point, every time you put the same object in front of the camera the program will recognize the object and give the name associated with it. The final program will use feedback between the layers of its nodes. The program will try to match its input which depends whether or not its nodes are on or off. A prediction scheme will be used to evaluate how well it can predict the input, and then the model will adjusts its output to make it better able to predict next time.


Team Members:

  Jeremy Wilson
  Daniel Cooper
  Jorge Palma

Sponsoring Teacher: Gregory Marez