Pages of abstract infinities (AI)

I have been messing around with making a new kind of wiki that uses Markov models to store the relationships of information and what defines understanding of a concept and how that understanding is developed. It started with the einstein puzzle game which gave some hints and Markov chains and neuroanatomy and AI systems.

In any case I think I have the structure that pulls the information in and spits it back out like a dream that flows in concert with the users skill. It incorporates graphics and equations and text as well as animation to represent things which have no physical form like concepts and relationships in space and time. It respects the dependency sequence and follows the knowledge in the steps needed to relay the concepts based on the subtle dependencies. It is certainly much like the Markov chains. The goal is to reprogram bots as new information becomes available from many different temporal delays and coordinate the relationships that apply in a certain application, much like the debian archives except that it extends to the neural implementation.

The burden of so many complexities and their factorial interaction is getting oppressive and I need a helper application that delves deeper into the data and relationships to help with the time involved in the recursive solutions of the inter-relationships. The goal is to have a way for each agent element to teach every other agent element what is learned so that the necessary dependencies can be met in a single stream at whatever best data rate can be achieved. I suppose a complete wipe and dupe is what has to be done more often as untangling the spaghetti and getting certainty from that is too iffy.

0 comments:

Automated Intelligence

Automated Intelligence
Auftrag der unendlichen LOL katzen