Skip to content
Marxos edited this page Sep 3, 2022 · 6 revisions

The knowledge base in the AI is the system of independent predicates either entered into the system manually or learned and stored by the AI:learner.

It is anticipated that this knowledge base can be converted or stored natively in its own mathematical predicate language for merging and sharing across all droids.

To avoid the exponential growth problem of too many predicates, the dr0id is in dialog with its owner to find greater (named) categorizations of its data. This scales the O(n) problem to O(log n), a major achievement, allowing everything in the universe to be cateloged, in theory, by about 60-100 questions. A trigger condition, such as greater than 10 predicates with the same object might create a question for the droid to resolve with the owner. Or a trigger where 80% of conditions are met but a stray 5% is out-of-order and forms a contradiction: "what is this element?". Such "negative" predicates set up a lot of knowledge acquisition, however.

Clone this wiki locally