TRIPLE a Rule Language for RDF

1
TRIPLE a Rule Language for RDF

• Stefan Decker
• Stanford University

2
Many Previously Unknown Communication Partners
3
Too many languages, to many formats
4
Layers help to manage Complexity
5
1. Step

• Define uniform, underlying syntax
• Lowest common denominator labeled
  graphs(semi-structured Data) -gt RDF

Relational Database
Structured Text (e.g., Vcard)
Person
begin vcardfn Stefann
DeckerStefanend vcard
Person
row
row
vcard1
fn
n
L-name
L-name
ID
ID
F-name
F-name
Stefan
DeckerStefan
1
Decker
Decker
Stefan
Birgit
2
6
RDF

• Graph-based universal syntax

(Agent-) Applications
RDF-Layer (Single dataformat, Query and storage
System)
Scheduling Service
Insurance Ratings
Calendar
Semantics in a global, open environment?
7
Step 2 Layers on Top of RDF

• Tim Berners-Lee
• Axioms, Architecture and Aspirations
• W3C all-working group plenary Meeting
• 28 February 2001

8
Motivation Why Rule Languages for the Web?

• Plethora of data available
• Data needs to be adapted and combined
• Time to Market Faster to write rules than code

9
Data Models on the Web

• Plethora of data models and languages
• Relational Data, UML, ER, TopicMaps, DAMLOIL,
  XML-Schema, special purpose data models
• Semi-structured data for the integration of
  heterogeneous data
• Graph-based data models academia OEM (Object
  Exchange Model)
• W3C Standard RDF
• XML is another (restricted) form of
  semi-structured data

10
Whats Wrong With Existing Approaches to RDF
Query?

• Built-in semantics (e.g. SiLRI, RQL, DQL)
• but many RDF-based languages with different
  semantics (DAMLOIL, RDF Schema, UML/RDF,
  TopicMaps/RDF,)
• For each language a specialized query language
  ????

11
RDF Models

• No support for RDF models
• one large heap of RDF data

model 2
model 1
model 3
12
Idea
model1
Set of RulesDescribing Semantics of a Data Model
model2
13
TRIPLE Requirements

• Support for Specification not Programming
• Bottom-up evaluation or tabling
• Expressive rule bodies
• Negation
• Support for Object Invention
• Skolem functions are required
• Effective Evaluation and Change
• Well-founded vs. stables semantics (intuitive
  behavior when changes occur!

14
TRIPLE Language Overview

• Native support for
• Resources namespaces, abbreviations
• Models (sets of RDF statements) and Semantic
  Spaces
• Reification
• Rules with expressive bodies (full FOL syntax)
• Transformations
• Syntactical extension of Horn Logic
• Syntactically inspired by F-Logic
• subjectpredicate?object (molecules)

15
Language Description I

• Namespace and resource abbreviations
• rdf http//www.w3.org/1999/02/22-rdf-syntax-ns
  .
• isa rdfsubClassOf.
• Statements, triples, molecules
• subjectpredicate?object
• subjectp1?o1 p2 ?o2 ...
• s1p1 ? s2p2?o
• Models, model expressions, parameterized models
• sp?o_at_m triple lts,p,ogt in model m
• sp?o_at_(m1 ? m2) model intersection, union, diff.
• sp?o_at_sf(m1, X, Y) Skolem function

16
Language Description II

• Reification
• stefanbelieves ? ltOraisAuthorOf?homepagegt
• Logical formulae
• usual logical connectives and quantifiers ? ? ?
  ? ?, lt-gt -gt lt-
• all variables introduced via ? (or ?)
• Clauses
• facts sp1?o1 p2 ?o2 ....
• rules ?X s1p1?X ? s2p2?X ? ... .
• Blocks
• _at_model clauses
• ?Mdl _at_model(Mdl) clauses

17
Example Dublin Core
dc http//purl.org/dc/elements/1.0/. dfki
http//www.dfki.de/. _at_dfkidocuments
dfkid_01_01 dctitle ? TRIPLE
dccreator ? Michael Sintek
dcsubject ? RDF dcsubject ? triples
... .
namespace abbreviations
block
Michael Sintek
TRIPLE
fact
dctitle
dccreator
dfkid_01_01
dcsubject
dcsubject
...
RDF
triples
Query FORALL D lt- Ddcsubject ?
RDF_at_dfkidocuments.
18
Layered Architecture Semantic Spaces

• TRIPLE supports the definition of semantical RDF
  extensions in a modular way
• RDF Schema (and other simple frame
  systems)semantics can be directly defined in
  TRIPLE as a parameterized model (see next slide)
• OIL, DAMLOIL (i.e., expressive ontology
  languages, DL)requires interaction with foreign
  reasoning components (e.g., DL classifier)
• Goal use various semantics in one inference
  (e.g., for information integration)

19
TRIPLE/RDFS RDF Schema Model
namespace abbreviations

• rdf 'http//www.w3.org/...rdf-syntax-ns'.
• rdfs 'http//www.w3.org/.../PR-rdf-schema-...'
  .
• type rdftype.
• subPropertyOf rdfssubPropertyOf.
• subClassOf rdfssubClassOf.
• FORALL Mdl _at_rdfschema(Mdl)
• FORALL O,P,V OP-gtV lt-
• OP-gtV_at_Mdl.
• FORALL O,P,V OsubClassOf -gtV lt- EXISTS W
  (OsubClassOf -gtW AND WsubClassOf -gtV).

resource abbreviations
model block
copy triples from Mdl
transitivity of subPropertyOfand subClassOf
20
UML in TRIPLE
namespace abbreviations

• rdf 'http//www.w3.org/...rdf-syntax-ns'.
• uml http//www.omg.org/uml/1.3/Foundation.Core
  .'.
• FORALL Mdl _at_uml(Mdl)
• FORALL O,P,V OP-gtV lt-
• OP-gtV_at_Mdl.
• FORALL X,Z g(X,Z)rdftype-gtumlGeneralization
  
• umlGeneralization.child-gtX
  umlGeneralization.parent-gtZlt- EXISTS
  Y,G1,G2
• G1umlGeneralization.child-gtXumlGeneralizatio
  n.parent-gtY AND G2umlGeneralization.child-gtY
  umlGeneralization.parent-gtZ .

model block
copy triples from Mdl
transitivity of umlGeneralization
21
TRIPLE and DAMLOIL

• daml_oil(Mdl) model by accessing a DL classifier
  (e.g., FaCT) (in general external software
  component)
• access only allowed in rule bodies
• results in hybrid rule language similar to Carin,
  but more pragmatic approach

22
TRIPLE/DAMLOIL Example

• daml 'http//www.daml.org/.../damloil'.
• animals 'http//www.example.org/animals'.
• _at_animalsontology
• animalsAnimalrdftype -gt damlClass.
• animalsHerbivorerdftype -gt damlClass
• rdfssubClassOf -gt animalsAnimal.
• animalsCarnivorerdftype -gt damlClass
• rdfssubClassOf -gt animalsAnimal
• damldisjointWith -gt animalsHerbivore.
• animalsOmnivorerdftype -gt damlClass
• rdfssubClassOf -gt animalsHerbivore
• rdfssubClassOf -gt animalsCarnivore.
• FORALL Ont _at_check(Ont)
• FORALL C unsatisfiable(C) lt-
• CdamlsubClassOf -gt damlNothing_at_daml_oil(On
  t).

Animal
s
s
Herbivore
Carnivore
damldisjointWith
s
s
Omnivore
s rdfssubClassOf
find all unsatisfiable classes(will detect
Omnivore)
23
Generalization Mediation Architecture with
TRIPLE
Agent-Applications
TRIPLERDF
Software Component
WebService A
TopicMap
24
Realization Compilation to Horn logic

• First implementation (and informal semantics) by
  mapping to Horn Logic / XSB system (Prolog with
  tabled resolution)
• Lloyd-Topor transformation for quantifiers etc.
• RDF-specific transformations given as rewrite
  rules

25
Future and Ongoing Work on TRIPLE

• Direct Semantic Characterization
• Classic Negation
• ECA Rules
• Pure Java Version
• Scalability Issues

26
Outlook Ontology Versioning as an Application
of TRIPLE

• Ontology building is a collaborative effort
  like software engineering
• Versioning issues come up
• CVS (Concurrent Versioning System) essential of
  open source software projects (APACHE)
• CVS works on text files Which subclasses of
  Person changed in the last Week is not possible

27
TRIPLE as a Basis off a Multiple Semantics Query
Ontology Versioning System (OVS)
Collaboration Primitives (similar to CVS)
TRIPLE
DAMLOIL
UML
RDF Schema
28
Ontology Driven Peer-to-Peer Systems

• Ongoing Project (http//edutella.jxta.org)
• Peer-to-Peer concept not restricted to file
  transfer, but arbitrary services
• Scenario Peer-to-Peer Network of all hotels,
  clustered by an Ontology
• Peers need a query infrastructure, ideally based
  on an application independent data format -gt
  TRIPLE

29
Conclusions

• TRIPLE is a new RDF-specific query and inference
  language
• Allows specification of/access to multiple
  semantics
• Every Horn Logic inference engine can be used
• First implementation available (for a subset of
  TRIPLE called TRIPLE0)http//triple.semanticweb.
  org
• RDF Exchange syntax is defined
• Part of RuleML initiativehttp//www.dfki.uni-kl.
  de/ruleml/