Ontology families for AI-systems non-404

[X DdC]

ontolo...@googlegroups.com

Dear All,

Apology for the earlier 404 -link.
[[ It had         https://www.ontooo.com/ontology.pdf
   instead of  https://www.ontooo.com/Ontology.pdf
   Yet another instance of my un-intelligence. ]]

Wrote a paper:
     Ontology families for AI-systems
Abstract:
     The recent chat-systems exhibit a discontinuity against the AI
     achievements since 1970.  They support ‘chat’ interaction with
     customers by relying on having been trained on massive amounts of
     text. They can generate useful replies (beyond earlier Google
     results), which has amazed the general audience. At the same time
     these chat-systems cannot apply the crown-jewel algorithms
     (search, planning, reasoning, etc.) on simple problem
     specifications of these algorithms. We provide numerous examples
     showing that the Large Language Models system architecture is the
     root problem.   In each case, we argue that an ontology is
     required to handle these examples. In addition, we argue that
     large AI-systems need multiple ontologies.  These ontologies are
     the bridge to neuro-symbolic computing. Pursuing Artificial
     General Intelligence (AGI) requires – we claim –
     meta-ontologies. At the same we provide arguments against the
     ultimate feasibility of AGI. We permit ourselves to take the
     stance that AI-systems should avoid certain topics.

It is available at:
  https://www.ontooo.com/Ontology.pdf
Comments needed/requested ...
dennis de champeaux

[Mike Peters]

A fascinating read, Dennis
Cheers
Mike Peters
Ajabbi

https://www.blog.ajabbi.com/
New Zealand