[xtm-wg] More Formal Semantics in Topic Maps: OIL, FaCT, RDF ...

[Alexander Sigel <sigel@bonn.iz-soz.de>]

dear colleagues in topic map land,

thank you very much for sharing your insights with me during personal
conversations and in the tm track at the recent xml europe! it was
both inspiring and exhausting. however, i had the strong impression
that TM proponents are not fully aware of the body of work done elsewhere,
although AI approaches and RDF were quite discussed.

in order to still learn more, in this mail i start a thread on:

 |  alternatives in more formal means to express inferencable semantics
 |  in topic maps and similar mechanisms (for practical ontology-based
 |  applications)

  1. TMs + KL-ONE-alikes/description logics (DL) + frames + CG?
  2. TMs vs. RDF?
  3. TMs + semantics vs. "semantic web"/OIL?

maybe the questions have some implications for tm-org efforts?

i aim at sophisticated tool support for human indexers. part of that
are heuristic information agent (transfer) strategies which work on
multi-view ontological domain models (applied to the social sciences).

1. TMs + KL-ONE-alikes/description logics (DL) + frames + CG ?
--------------------------------------------------------------
tm schemata leverage tms from untyped semantic networks to typed ones.
the power of typed links is long known since wood's classical paper
of what's in a link. in that ai context, the KL-ONE system and alike (e.g.
SB-ONE) had been invented to represent concepts and their interrelations
in a formal way. this system discerns between the terminological and
the assertional box (TBox and ABox), but e.g. already woods had questioned
the utility of that. important is that the field of terminological or
description logics derived from that, while woods later on used simpler
means to express terminologies. one questions is: do we really need
the expensive subsumption service for semantic retrieval?

the main problem in the practical application of DL has been the
untractability of subsumption problem (i.e. too slow or undecidable).
however, recent advantages in DL seem to make it promising to couple
topic maps with DL inferencing power. the FaCT system is a free tool
which could be used for this purpose.

although several means to express semantics (semantic networks,
frames, CG) have been proven to be equivalent to first-order logics (FOL),
frame-based languages are advantageous because of their ability to
more naturally express gestalt structures. i.e. frames superimpose
a prototypical structure upon linear predicates. i doubt if the modelling
of n-ary predicates as demonstrated by steve pepper are the full solution
towards frames. e.g. how can i express complex nested relations?

in addition to that, there are several conceptual graph (CG) tools available
which could be used on the user interface side.

2. TMs vs. RDF?
---------------
i am not satisfied with the explanations i could find about the relationship
between TMs and RDF (or other schema languages or e.g. Schematron) and
hope we will see some clarification there. if RDF is more general, i.e.
TMs which allow scopes could be implemented in RDF: how would implementations
in other schema languages be different?
this still leaves the issue of interoperability with people using RDF
to express their ontology brokering approach on the web and should not
be neglected because it hinders the full acceptance of TMs.

3. TMs + semantics vs. "semantic web"/OIL?
-------------------------------------------
only today have i become aware of OIL (Ontology Inference Layer), a
recent direction the "semantic web" + ontology community has taken in
the context of EU-funded projects. OIL proposes to extend RDF with
features to express ontologies and inferencing on them (for the web),
claiming this to be an improvement over e.g. Ontolingua. (it is an XOL
extension. SHOE is cited, but not CKML). the main idea is that OIL intends
to join description logics' subsumption and inferencing power with
frame-based systems, using inter alia the FaCT tool.

Question:
=========
- How can we assure that the advantages of OIL will be available in
  TMs, or that TM functionalities and products will be interoperable with 
  OIL's?
- Should tm-org (ex XTM) invite someone from the OIL group and/or the
  semantic web people to a technical meeting?
  Can you please ensure that no duplicate work is done?

(Only some) References:
-----------------------
A. Woods and James G. Schmolze. The KL-ONE family. In F.W. Lehmann, editor,
Semantic Networks in Artificial Intelligence, pages 133--178.
Pergamon Press, 1992.

William A. Woods. Understanding subsumption and taxonomy: A framework for
progress. In John F. Sowa, editor, Principles of Semantic Networks:
Explorations in the Representation of Knowledge, pages 45-- 94.
Morgan Kaufmann, San Mateo, CA, 1991.

[Brachman et al., 1983]
Brachman, R.J.; What IS-A is and isn't. An Analysis of Taxonomic Links
in Semantik Networks; IEEE Computer, 16 (10); Oktober 1983

[Brachman et al., 1985]
Brachman, R.J./ Schmolze, J.G.; An overview of the KL-ONE knowledge
representation system; Cognitive Science Vol. 9, S. 171-216; 1985

William A. Woods. What's in a Link: Foundations for Semantic Networks.
In Daniel Bobrow and Allan Collins, editors, Representation and Understanding.
Academic Press, 1975. 

http://wwwwswest.sun.com/research/people/william.woods/

Early work by Lori(n) A. Resnick (ATT Research) on CLASSIC and on distance
measures in semantic networks (as far as i can remember)

"Semantic Web"
http://www.semanticweb.org/
  e.g.:
  http://www.aifb.uni-karlsruhe.de/WBS/dfe/dagstuhl.html
    "Semantics for the Web" Workshop, Dagstuhl, Germany
  http://www.tzi.de/buster/description.html
    "Bremer Semantic Interoperability Project"

OIL
http://www.ontoknowledge.org/
                             oil/
  especially: OIL in a nutshell (Fensel et al.)

FaCT homepage
  http://www.cs.man.ac.uk/~horrocks/FaCT/

XOL - Ontology Exchange Language (originally in bioinformatics, Karp et al.)
  http://www.ai.sri.com/pkarp/xol/

Next CG conference (ICCS 2000, Darmstadt, Germany)
  http://www.mathematik.tu-darmstadt.de/ags/ag1/iccs2000/
  Sowa's homepage with excerpts from his new book:
    http://bestweb.net/~sowa/direct/index.htm

Next DL conference (DL 2000, Aachen, Germany)
  http://www-lti.informatik.rwth-aachen.de/DL2000/
  DL in general: http://dl.kr.org/dl/

Schematron
  http://www.ascc.net/xml/resource/schematron/schematron.html

--

so long
alex
----------------------------------------------
Alexander Sigel, M.A. sigel@bonn.iz-soz.de
Informationszentrum Sozialwissenschaften
Lennéstr. 30, D-53113 Bonn, Germany
+49 228 2281 170 tel, +49 228 2281 120 fax
Homepage: http://index.bonn.iz-soz.de/~sigel/

------------------------------------------------------------------------
Wrox Wireless Developer Conference, Amsterdam, July 10-12. Choose from
40+ technical sessions covering application of WAP, XML, ASP, Java and
C++ to mobile computing. Get your ticket to the future today!
http://click.egroups.com/1/5689/3/_/337252/_/961427724/
------------------------------------------------------------------------

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com