The title of my talk in that planning session was "Relating Classical
and Nonmonotonic Logic." As a result of the discussion with the other
participants, I added six more slides to my talk, and I changed the
title to emphasize the point I was trying to make: (01)
http://www.jfsowa.com/talks/sharing.pdf
Knowledge Sharing Among Heterogeneous Agents (02)
My new title emphasizes the goal: Use a classical logic, such as
Common Logic, as the basis for message passing among agents that
use an open-ended variety of different logics -- or no explicit
logic of any kind. That was the original goal of the proposal
for the Semantic Web by Tim Berners-Lee in February 2000. (03)
My primary claim is that knowledge sharing among agents that use any
kind of reasoning method -- classical, nonmonotonic, probabilistic,
statistical, heuristic, case-based, fuzzy, or quirky (procedural) --
can and should be based on passing messages that are stated in
classical FOL, possibly supplemented with metadata stated in FOL. (04)
Slide 20 (copy below) summarizes the conclusions. Slide 19 says
that there is more to come. I would appreciate any comments,
suggestions, or complaints. (05)
John
______________________________________________________________________ (06)
Source: Slide 20 of http://www.jfsowa.com/talks/sharing.pdf (07)
FOUNDATION FOR KNOWLEDGE SHARING (08)
Model-theoretic semantics is key to all versions of logic: (09)
● Classical logic has a fixed set of axioms (assumptions and
constraints) and a fixed set of models. (010)
● Nonmonotonic methods may change the assumptions, constraints,
and models during the reasoning process. (011)
● Heuristics, probability, statistics, case-based reasoning, and
fuzzy logic use a wider range of metalevel reasoning. (012)
● But the semantics of every method of reasoning is based on an
evaluation function that determines truth values in terms of models. (013)
Basis for message passing among heterogeneous agents: (014)
● All agents can reason with statements in FOL or some subset of FOL. (015)
● All agents can assert statements in FOL. (016)
● Messages should consist of FOL statements with FOL metadata. (017)
● The metadata may state assumptions, constraints, and qualifications,
such as probabilities, error bounds, or fuzzy values. (018)
● Any agent may use its own methods to interpret any message. (019)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (020)
|