Expected Behavior of the Learning System

It is assumed that a learning system will every time unit respond to it's inputs at least with a language expression (L) and with some kind of visualization of his model of moving ojects. In the advanced mode it can respond also with a movement command (M) for it's own avatar as its own body. An avatar is a moving object in the moving area which functions like a body for the learning system. The main idea is that the learning system can move it's avatar directly at one time unit and in the next time unit it can sens a difference in the position. In case of those moving objects which can move but are no avatars the learning agent can only sense the data, but it can not directly influence these changes.

These informal problem descriptions have to be formalized and then have to be translated into possible mechanisms of the learning system.

Gerd Doeben-Henisch 2012-03-31