[x3d-public] Semantics of Script Templates, Generics, Queries, AI and Ontological AI? What is AI distinct from semantics?

[John Carlson]

Script Templates are Templates of Script, Class, Test, Function, PROTOs, Predicates, Behavior, Thread, Process, Processor, Network, System and World.   Some Scripts are called Queries or Predicate Scripts, and  have parameters Predicate Logic Expression, Constants, Input Variables return a Collection of Objects.  Some Queries return collections of Scripts.    These are called Goal Seeking Scripts, and it’s left to the caller to evaluate the return values.  The Scripts they return are actions to reach a goal.  Class Templates are called Generics, and are Scripts (often collection classes) parameterized by Types.  You pass several types to a generic and get a non-parameterized script out, with the type parameter filled in where it appears.  Function Templates are cross domain function names.   Behavioral Templates have the goal of creating role model behaviors or exemplars.   These are unit tests or generative adversarial networks (GANs). What returns unit tests (x3d-big-bang).   Something that returns Scripts to choose from (I think of this more as genetic programming or algorithms) or something constructs scripts or functions from an inputs and weights or internal states to an output states? (This seems like neural networks and theorem provers).  How exactly can you find analogical scripts? How do you construct scripts where predicates are passed as a parameter?  (this seems like a query,  an AI which answers questions by constructing analogies or generics)  This thing that either creates scripts or functions or finds existing scripts or functions, possibly through exhaustive search.  So why do neural network works work, and classical AI does not first of all, classical AI does work for Theorem Provers.   So lets call two parts of AI Theorizing and Solving (using existing actions and creating new actions between inputs and outputs) and Pattern and Solution Recognition and with Categories (finding patterns in input and input classes using self-organizing functions).  Let’s refer to x3d-big-bang as a Generative Adversarial Network (or universe) for generating test cases or processes for Theorizing, Solving and Pattern and Solution Recognition with Categories in many domains.   Ontologies are sign systems and sign worlds used to describe systems and worlds. Ontologies are NOT AI.  Specific Ontological triples are AI in math: (a computer or robot using, creating, producing, consuming and understanding Theories/Problems, Proofs/Solutions/Scripts, Categories and Patterns).  There are other specific ontological triples in other domains which are also AI.  I welcome you to bring them up.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20190323/80b56f90/attachment.html>