Cogitech IncMiscellaneous | About us | Training | Tabular Topic Maps | Cogitative Topic Map Websites | Semantic Web Glasses | XWATL
Cogitative Technologies>Semantic Web Glasses

Secure SSL Certificates.

Presentation DAML and Quantum Topic Maps at Knowledge Technologies'2002 Conference. 2002-03-15

DAML and Quantum Topic Maps

Proposed changes to RDF Topic Maps [Extreme Markup Technologies 2001] result in a more compact representation.
It will be shown how DAML-OIL can be used to enrich Topic Maps expressiveness and how DAML-OIL ontologies can be used to validate Topic Maps.
We will look at "Quantum" effects in Topic Maps and consider outlines of Quantum Topic Maps Processing Model derived from TMPM4.
Knowledge Technologies 2002; March 13

by Nikita Ogievetsky , Cogitech, Inc.
nogievet@cogx.com
©Cogitech, Inc.



Agenda

  • Overview of RDF Topic Maps
  • Issues in RTM, changes in order to move to DAML
  • Attempt is made to find a simplified Topic Maps Model derived from TMPM4.
    • Admitting that some ideas are crazy but important.
  • Overview of Quantum Topic Maps
  • DAML Validation
  • Quantum Reification.



Overview of RDF Topic Maps

  • Topics => PSI hubs
  • association =>
    • RDF node of type derived from association template (if association template exists).
    • RDF node of type rtm:association.
  • occurrence =>
    • RDF node of type derived from occurrence type (if exists).
    • RDF node of type rtm:occurrence.
  • baseName = > RDF node of type rtm:basename
  • variantName => RDF node of type rtm:variantname
  • association members => rtm:member bags, objects in association triples.
  • scope => rtm:scope bags, objects in association triples.
  • instanceOf => RDF node of type rtm:classinstance if referenced class is not an association template
  • RDF Classes and Properties declarations are created for all used association templates and association membership roles.



Issues




Issue 1. Lost Association

RTM following most of the Topic Maps data models hides if not looses explicit association between subjects constituted-by and indicated-by one and the same resource.
<association>
 <instanceOf>
  <subjectIndicatorRef xlink:href="#media-indicating-subject" />
 </instanceOf>
 <member>
  <roleSpec>
   <subjectIndicatorRef xlink:href="role-subject"/>
  </roleSpec>
  <subjectIndicatorRef xlink:href="#xyz"/>
 </member>
 <member>
  <roleSpec>
   <subjectIndicatorRef xlink:href="role-resource"/>
  </roleSpec>
  <resourceRef xlink:href="#xyz"/>
 </member>
</association>



Issue 2. Nested membership statements in Classic RTM

rtm:member - a proxy specifying whether association role players are indicated or constituted by the referenced resources.
  • Object in a roleSpec property assignment.
  • Subject in membership bag.
cogx:/marriage cogx:/husband [ rtm:/indicatedBy nikita ]



Association members in Quantum RTM

Instead of using a proxy statement, the very definition of a property indicates whether the referenced resource indicates or constitutes the property's object.
cogx:/marriage cogx:/husband nikita
How is this possible without loss of information?
<subjectIndicatingProperty rdf:about="http://www.cogx.com/family.daml#husband">
  <rdfs:domain rdf:resource="http://www.cogx.com/qtm.rdf#association"/>
  <rdfs:range rdf:resource="http://www.cogx.com/qtm.rdf#member"/>
</subjectIndicatingProperty>



Two base properties in Quantum RTM

  • subjectIndicatingProperty
<daml:ObjectProperty rdf:ID="subjectIndicatingProperty">
 <rdfs:domain rdf:resource="http://www.cogx.com/qrtm.rdf#association"/>
 <rdfs:range rdf:resource="http://www.w3.org/2000/01/rdf-schema#Resource"/>
</daml:ObjectProperty>
  • subjectConstitutingProperty
<daml:ObjectProperty rdf:ID="subjectConstitutingProperty">
 <rdfs:domain rdf:resource="http://www.cogx.com/qrtm.rdf#topicLink"/>
 <rdfs:range rdf:resource="http://www.w3.org/2000/01/rdf-schema#Resource"/>
</daml:ObjectProperty>
  • literal



Issue 3. Topic Expansion

In RTM (following TMPM4) XTM <topic> element expands into
  • multiple associations, one for each occurrence and baseName..
In Quantum Topic Maps
  • XTM <topic> element expands into multiple statements grouped by context (scope) and type.
  • there is no major difference between an association and a topic link.



Example. XTM

<topic id="XYZ">
  <instanceOf><topicRef="#male"/></instanceOf>
  <occurrence>
    <instanceOf><topicRef="#age"/></instanceOf>
    <scope><topicRef="#dating-service-AAA"/></scope>
    <resourceData>39</resourceData>
  </occurrence>
  <occurrence>
    <instanceOf><topicRef="#IQ"/></instanceOf>
    <scope><topicRef="#dating-service-AAA"/></scope>
    <resourceData>90</resourceData>
  </occurrence>
  <occurrence>
    <instanceOf><topicRef="#gender"/></instanceOf>
    <scope><topicRef="#dating-service-AAA"/></scope>
    <resourceData>M</resourceData>
  </occurrence>
  <occurrence>
    <instanceOf><topicRef="#photo"/></instanceOf>
    <scope><topicRef="#dating-service-AAA"/></scope>
    <resourceRef  xlink:href="http://www..../xyz.jpg">
  </occurrence>
  <occurrence>
    <instanceOf><topicRef="#profile"/></instanceOf>
    <scope><topicRef="#human-resources"/></scope>
    <resourceRef xlink:href="file://C:/profiles/xyz-profile.doc">
  </occurrence>
  <occurrence>
    <instanceOf><topicRef="#personal-records"/></instanceOf>
    <scope><topicRef="#human-resources"/></scope>
    <resourceRef  xlink:href="http://fo-company.com/xyz-records.html">
  </occurrence>
</topic>



Example. Quantum RTM

<topic id="XYZ">
  <indicatedBy>
    <topicLink>
      <rdf:type rdf:resource="#male"/>
      <rtm:validIn rdf:resource="#dating-service-AAA"/>
      <this:age>19</this:age>
      <this:IQ>90</this:IQ>
      <this:gender>M</this:gender>
      <this:photo rdf:resource="http://www..../xyz.jpg"/>
    </topicLink>
  </indicatedBy>
  <indicatedBy>
    <topicLink>
      <rdf:type rdf:resource="#male"/>
      <rtm:validIn rdf:resource="#human-resources"/>
      <this:profile rdf:resource="file://C:/profiles/xyz-profile.doc"/>
      <this:personal-records rdf:resource="http://fo-company.com/xyz-records.html"/>
    </topicLink>
  </indicatedBy>
</topic>
Which says:
  • In the validity scope indicated by "#dating-service-AAA" this subject is a "#male" and his "#age" is 19, his "#IQ" is 90 and his "#photo" can be found at "http://www..../xyz.jpg";
  • And in the validity scope indicated by "#human-resources" this subject is also a male and his "#profile" can be found at "file://C:/profiles/xyz-profile.doc" and his "#personal-records" can be found at "file://C:/profiles/xyz-profile.doc"



Rationale




Subjects: An event can be indicated by

  • one or more associations between actors playing certain roles in certain context (scope).
  • one or more resources
    • marriage between two people can be indicated by a marriage certificate.



Subjects: An actor can be ...

  • Indicated by an event
  • Constituted by a resource
    • [this presentation]
  • Indicated by one or more resources
    • [subject of this presentation]
  • Extensionally indicated by an association between resources and various properties playing certain roles in certain context (scope) and forming a representation of an actor.
  • Extensionally indicated by an association of roles played in associations with other actors.
    • [UFO]



Duality of topics / associations

All subjects that we operate originated as associations.
  • An association between family members indicates a Family.
  • An association (topic link) between facts, remembrances and images forms (emanates) representation of a person. (A person is indicated by this representation).
  • A simile (or set theory) association of all sets with 3 items indicates number 3.
And yet we operate with subjects, which are reified associations.
  • "Their marriage did not last long"



Two subject indicating properties in Quantum RTM

<daml:UnumbiguousProperty rdf:ID="indicatedBy">
	<rdfs:subPropertyOf rdf:resource="http://www.daml.org/2001/03/daml+oil#TransitiveProperty"/>
	<rdfs:domain rdf:resource="#topic"/>
	<rdfs:range rdf:resource="http://www.w3.org/2000/01/rdf-schema#Resource"/>
	<rdfs:comment>Subject Indicating Resource</rdfs:comment>
</daml:UnumbiguousProperty>

<daml:UnumbiguousProperty rdf:ID="constitutedBy">
	<rdfs:domain rdf:resource="#topic"/>
	<rdfs:range rdf:resource="http://www.w3.org/2000/01/rdf-schema#Resource"/>
	<rdfs:comment>Subject Constituting Resource</rdfs:comment>
</daml:UnumbiguousProperty>



Two Base Associations in Quantum Topic Maps

  • topic association - intentional association between topics. Scopable.
  • topic link - extensional association between topics. Scopable.
  • Quantum topic - independent of context. Not scopable.



Quantum Topic

  • Equally unambiguously identified by a set of
    • reified topic links;
    • reified associations;
    • resources - subject indicators;
    or constituted by a resource.



Similarity with Quantum Mechanics

  • This is similar to the duality of elementary particles, which propogate as waves but observed as particles.
  • We change waves when we observe them.
  • We change associations when we are getting their details.
An old saying says that we are changing a subject by making a statement about it.
Example1: To get details about a relationship between two people...
  • You will have to meet them and probably interfere with them in some ways.
  • The more precise description you are getting the more chances that your intrusion will change the relationship. (Imagine trying to get intimate details).
Example2: "Nothing" - is something that does not have properties.
  • Oops! it just got a property of not having properties!
Example3: Paparazzi...
Example4: Financial Stock researches...



The theory of quantum mechanics

  • ...is based on an entirely new type of mathematics that no longer describes the real world in terms of particles and waves; it is only the observation of the world that may be described in those terms. There is thus a duality between waves and particles in quantum mechanics: for some purposes it is helpful to think of particles as waves and for other it is helpful to think of waves as particles. -- S. Hawking. A brief history of time.



Space and Time VS Characteristics and Context

Hundred years ago we realized that we leave in a four dimensional Universe and that there is no real distinction between Space and Time coordinates.
In Topic Maps we operate in multidimensional Information Universe with one set of coordinates formed by subject characteristics (assertions) axis and the other set of coordinates formed by assertion context axis.



Particles' Tracks

In nuclear physics particles are recognized by analyzing tracks they leave in bubble chambers, drift chambers, RPC sandwiches, silicon trackers.



Things' Tracks

Things (subjects) are being recognized by impressions (representations) they leave on their observers.
We analyze things by the tracks that they leave in the multidimensional universe where one set of dimensions is formed by their characteristics and the other set of dimensions is formed by the context of observation (scope).
World of elementary physics is a subset of our Information Universe with three characteristics axis (X,Y,Z dimensions) and one context axis: Time.
In real live it is a job of a detective to identify and reconstruct subjects by their tracks in the InfoSpace.
  • Neural networks - a great application for Quantum Topic Maps.



Quantum Topic

quantopic? quantom? quantic? qutom?
Track of a thing in the infospace. Various characteristics registered in various contexts.
<rtm:topic id="C">
  <rtm:indicatedBy>
    <rtm:topicLink id="A">
      <rdf:type rdf:resource="#topic-maps-consultant"/>
      <rtm:validIn rdf:resource="#BV"/>
      <rtm:uniqueName>Nikita Oguievetski</rtm:uniqueName>
      <website rdf:resource="http://www.cogx.com">
      <characteristics>Sometimes samwhat makes some sense.</characteristics>
    </rtm:topicLink>
  </rtm:indicatedBy>
  <rtm:indicatedBy>
    <rtm:topicLink id="B">
      <rdf:type rdf:resource="#presenter"/>
      <rtm:validIn rdf:resource="#KT2002"/>
      <rtm:uniqueName>Nikita Ogievetsky</rtm:uniqueName>
      <presenataion rdf:resource="http://www.knowledgetechnologies.net/DAMLandRTM.html">
    </rtm:topicLink>
  <rtm:indicatedBy>
  <rtm:indicatedBy rdf:resource="urn:ssn:123-45-6789">
</topic>
Which says:
  • In the validity scope indicated by "#BV" this subject is of type "#topic-maps-consultant", it is uniquely identifiable by name "Nikita Oguievetski", its "#website" is at "http://www.cogx.com" and it has a "#characteristics" "Sometimes samwhat makes some sense.";
  • And in the validity scope indicated by "#KT2002" this subject is of type "#presenter", it is uniquely identifiable by name "Nikita Ogievetsky", it has a "#presentation" at "http://www.knowledgetechnologies.net/DAMLandRTM.html".



Syntax Changes in order to move to DAML

Syntactically a quantum topic element is indicatedBy multiple associations or topic link elements grouped by scope and class.
Membership proxy is not necessary:
  • membership type is pertinent to the particular role (predicate) class.



Two Identities

Global Level:
  • subject identifier of the quantum topic.
Topic Link Level:
  • uniqueName, applies only within a given scope.
Hense Two types of merging in Quantum Topic Maps:



Topic Link Merging

  • Based on TNC (Topic Naming Merging)
    • Merges topic links and hence quantum topics.



Subject Merging Rule

  • Topic Merging Constraint based on topic Identity;
    • Merges quantum topics directly.



Question: should we allow...

<topic id="the-anna-john-family">
	<rtm:indicatedBy>
	  <this:family-link>
		  <this:registereDate>01-02-2002</registereDate>
	  </this:family>
	</rtm:indicatedBy>
	<rtm:indicatedBy>
	  <this:family-association>
		  <this:wife rdf:resource="#ann"/>
		  <this:husband rdf:resource="#john"/>
	  </this:family>
	</rtm:indicatedBy>
</topic>
mixing topic link and association elements?:
<topic id="the-anna-john-family">
	<rtm:indicatedBy>
	  <this:family>
		  <this:registereDate>01-02-2002</registereDate>
		  <this:wife rdf:resource="#ann"/>
		  <this:husband rdf:resource="#john"/>
	  </this:family>
	</rtm:indicatedBy>
</topic>



DAML Validation




Given DAML Ontology

<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" xmlns:daml="http://www.daml.org/2001/03/daml+oil#">
	<daml:Ontology about="">
		<daml:imports resource="http://www.daml.org/2001/03/daml+oil"/>
		<daml:versionInfo>$Id: daml-family.daml,v 1.1 2000/10/07 03:21:17 connolly Exp $</daml:versionInfo>
	</daml:Ontology>
	<daml:Class rdf:ID="family">
		<rdfs:subClassOf>
			<daml:Restriction>
				<daml:onProperty rdf:resource="#wife"/>
				<daml:maxCardinality>1</daml:maxCardinality>
			</daml:Restriction>
		</rdfs:subClassOf>
		<rdfs:subClassOf>
			<daml:Restriction>
				<daml:onProperty rdf:resource="#husband"/>
				<daml:maxCardinality>1</daml:maxCardinality>
			</daml:Restriction>
		</rdfs:subClassOf>
	</daml:Class>
	<daml:Property rdf:ID="wife">
		<daml:domain resource="#family"/>
		<daml:range resource="#female"/>
	</daml:Property>
	<daml:Property rdf:ID="husband">
		<daml:domain resource="#family"/>
		<daml:range resource="#male"/>
	</daml:Property>
	<daml:Class rdf:ID="female"/>
	<daml:Class rdf:ID="male"/>
</rdf:RDF>



Given XTM document:

<topicMap  	xmlns:xlink="http://www.w3.org/1999/xlink">
	<topic id="john">
		<instanceOf>
			<subjectIndicatorRef xlink:href="http://www.cogx.com/family.daml#male"/>
		</instanceOf>
		<baseName>
			<baseNameString>John</baseNameString>
		</baseName>
	</topic>
	<topic id="anna">
		<instanceOf>
			<subjectIndicatorRef xlink:href="http://www.cogx.com/family.daml#female"/>
		</instanceOf>
		<baseName>
			<baseNameString>Anna</baseNameString>
		</baseName>
	</topic>
	<topic id="suzan">
		<instanceOf>
			<subjectIndicatorRef xlink:href="http://www.cogx.com/family.daml#female"/>
		</instanceOf>
		<baseName>
			<baseNameString>Suzan</baseNameString>
		</baseName>
	</topic>
	<association>
		<instanceOf>
			<subjectIndicatorRef xlink:href="http://www.cogx.com/family.daml#family"/>
		</instanceOf>
		<member>
			<roleSpec>
				<subjectIndicatorRef xlink:href="http://www.cogx.com/family.daml#husband"/>
			</roleSpec>
			<topicRef xlink:href="#suzan"/>
		</member>
		<member>
			<roleSpec>
				<subjectIndicatorRef xlink:href="http://www.cogx.com/family.daml#wife"/>
			</roleSpec>
			<topicRef xlink:href="#anna"/>
		</member>
		<member>
			<roleSpec>
				<subjectIndicatorRef xlink:href="http://www.cogx.com/family.daml#wife"/>
			</roleSpec>
			<topicRef xlink:href="#john"/>
		</member>
	</association>
</topicMap>



Quantum RDF document:

<rdf:RDF xmlns:rtm="http://www.cogx.com/xtm2rdf/rtm.rdf#"
		xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" 
		xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" 
		xmlns:a1="http://www.cogx.com/family.daml#">
	<a1:male rdf:ID="john">
		<rtm:uniquename>John</rtm:uniquename>
	</a1:male>
	<a1:female rdf:ID="anna">
		<rtm:uniquename>Anna</rtm:uniquename>
	</a1:female>
	<a1:female rdf:ID="suzan">
		<rtm:uniquename>Suzan</rtm:uniquename>
	</a1:female>
	<a1:family>
		<a1:husband rdf:resource="#suzan"/>
		<a1:wife rdf:resource="#anna"/>
		<a1:wife rdf:resource="#john"/>
	</a1:family>
</rdf:RDF>



Validator Error message:




Quantum RTM Reification

Quantom Topic Map reification differs from RDF reification:
For example, "website" occurrence:
<rtm:topic rdf:ID="I">
	<website rdf:resource="http://www.cogx.com"/>	
</rtm:topic>
Can be reified to a "website" association:
<rtm:association rdf:ID="rdf-r">
	<rdf:type rdf:resource="#website"/>	
	<rtm:role-topic rdf:resource="#I"/>	
	<rtm:role-occurrence rdf:resource="http://www.cogx.com"/>	
</rtm:association>
Which is different from RDF reification:
<rdf:Description rdf:ID="rdf-r">
	<rdf:type rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#Statement"/>	
	<rdf:predicate rdf:resource="#website"/>	
	<rdf:subject rdf:resource="#I"/>	
	<rdf:object rdf:resource="http://www.cogx.com"/>	
</rdf:Description>



DAML and Quantum Topic Maps reification'

It is interesting that according to Quantum Topic Maps reification mechanics, the reified node is an instance of DAML class with restrictions derived from the property definition:
For example, a property whose definition is:
<rdf:property rdf:ID="A">
  <rdfs:domain rdf:resource="#B"/>
  <rdfs:range rdf:resource="#C"/>
</rdf:property>
Reifies into an instance of a class whose definition is:
<daml:Class rdf:about="#A">
  <rdfs:subClassOf>
    <daml:restriction>
      <daml:onProperty rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#domain"/>
      <daml:toClass rdf:resource="#B"/>
    </daml:restriction>
  </rdfs:subClassOf>
  <rdfs:subClassOf>
    <daml:restriction>
      <daml:onProperty rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#range"/>
      <daml:toClass rdf:resource="#C"/>
    </daml:restriction>
  </rdfs:subClassOf>
</daml:Class>



Reification flow in Quantum Topic Maps




Assertion => Association => Subject

For example, given definition of an electron:
<qph:particle id="electron">
 	<qph:attractsTo rdf:resource="#proton"/>	
  {...}
</qph:particle>
Property => association
<qph:attractsTo id="electron-proton-attraction">
	<rtm:role-topic rdf:resource="#electron"/>	
	<rtm:role-rolePlayer rdf:resource="#proton"/>	
</qph:attractsTo>
Association => topic
<rtm:topic id="photon">
  <rtm:indicatedBy rdf:resource="#electron-proton-attraction"/>
  <rtm:indicatedBy rdf:resource="#effect-of-light"/>
</rtm:topic>



XTM Scopes and DAML

<rtm:validIn rdf:reaource="#A"/>
rtm:validIn references scope nodes:
  • Intersection of scoping topics:
<rtm:scope id="A">
  <rdfs:subclassOf> 
    <daml:intersectionOf parseType="daml:collection">
      <rtm:scope rdf:about="#about-xslt"/>
      <rtm:scope rdf:about="#about-topic-maps"/>
    </daml:intersectionOf>
  </rdfs:subclassOf> 
</rtm:scope>
  • Union of scoping topics:
<rtm:scope id="B">
  <rdfs:subclassOf> 
    <daml:unionOf parseType="daml:collection">
      <rtm:scope rdf:about="#experts"/>
      <rtm:scope rdf:about="#intermidiate"/>
    </daml:unionOf>
  </rdfs:subclassOf> 
</rtm:scope>
  • Complement:
<rtm:scope id="B">
   <daml:complementOf rdf:resource="#english"/>
</rtm:scope>



Changes from TMPM4

TMPM4 has all occurrences reified, while QTM uses topic links and associations to mark topic coordinates in the Information Space.
Due to quantum reification, there is no difference in the resulting graph.



The DAML is not enough

DAML-OIL Ontology Language mises scopes and lacks ways to express constrants relative to scopes.
Solution: add Quantum Topic Layer to DAML-OIL.
Extend DAML-OIL Validator with scoping ability.



Extensions (Off-topic)

If we can talk about subjects indicated by a resource, why can't we talk about subjects related to the resource otherwise?
  • For example:
  • Subject that was an inspiration for a resource.
<rtm:topic rdf:ID="TMPM4">
  <rtm:indicatedBy rdf:resource="http://www.topicmaps.net/pmtm4.htm"/>
  <rtm:inspirationOf rdf:resource="Quantum-Topic-Model.htm"/>
</rtm:topic>



THANK YOU!

Topic Maps Training:
  • http://www.cogx.com/training.html
XSLT Training:
  • http://www.cogx.com/training.html



References

  • XML Topic Maps (XTM) 1.0 TopicMaps.org Specification. http://www.topicmaps.org/xtm/1.0/
  • Resource Description Framework (RDF). http://www.w3.org/RDF/
  • DARPA Agent Markup Language (DAML). http://www.daml.org
  • DAML-OIL Validator. http://www.daml.org/validator/
  • Nikita Ogievetsky. XML Topic Maps through RDF Glasses. http://www.cogx.com/rtm2rdf
  • Topicmaps.net's Processing Model for XTM 1.0, version 1.0.2. http://www.topicmaps.net/pmtm4.htm
  • Eric van der Vlist. Representing XML Topic Maps as RDF. http://xmlhack.com/read.php?item=1108
  • Lars Marius Garshol. A Topic Map Data Model, An infoset-based proposal. http://www.y12.doe.gov/sgml/sc34/document/0229.htm





Budget Web Hosting and Cheap Domain Name Registration

Cogitech Inc. Made with Bexcelor & Tabular Topic Maps