Semantic Web &

Semantic Web ProcessesA course at Universidade da Madeira, Funchal, Portugal

June 16-18, 2005

Dr. Amit P. ShethProfessor, Computer Sc., Univ. of Georgia

Director, LSDIS labCTO/Co-founder, Semagix, Inc

Special Thanks: Cartic Ramakrishnan, Karthik Gomadam

Agenda 1Part I

• What is Semantic Web? • What makes the Semantic web

• Ontologies – importance of relationships and knowledge• Representation and Languages

• Why XML is not enough • Describe semantic web resources- RDF and RDFS • OWL

• Query processing and storage Part II• Metadata, Enabling techniques and technologies

• Ontology and knowledge engineering: ontology design, ontology population maintaining, ontology freshness

• Automated metadata extraction and annotation• Computation and reasoning with focus on relationships• Example commercial Semantic Web platform

Agenda 2

Part III• Semantic web applications: search, integration, analysis

a. Pan-Web and consumer-centricb. Enterprise

Part IV• Semantic Web Services and Processes

• What are Web Services ?• What are Web processes ?• Creating Web processes: Annotation, discovery,

composition, etc.

• Semantic Web Service/Process tools

Part I• What is Semantic Web? • What makes the Semantic web

• Ontologies – importance of relationships and knowledge• Types and examples of ontologies

• Metadata and Semantic Annotation -- metadata classifications

• Representation and Languages • Why XML is not enough • RDF - Describe semantic web resources and RDFS - RDF

as a triple, RDF as a graph (show example RDF/S) • OWL

• RDF Query processing and storage

Three generation of Information Systems:Where we have come from, where we are going

, Semantic Web technologies and platforms, Semantic Web technologies and platforms

MediaAnywhereMediaAnywhereInfoQuilt, InfoQuilt,

OBSERVEROBSERVER

Semagix FreedomSemagix Freedom

Generation IIIGeneration III

2000s2000s

Semantics (Ontology, Context, Relationships, KB)

Metadata based integration, Mediator Metadata based integration, Mediator Systems, Digital LibrariesSystems, Digital Libraries

Generation IIGeneration II

1990s1990s

VisualHarnessVisualHarnessInfoHarnessInfoHarness

AdaptX/HarnessAdaptX/Harness

Metadata (Domain model)

MermaidMermaidDDTSDDTS

IntervisioIntervisio

Heterogeneous databases/Heterogeneous databases/Federated Databases ResearchFederated Databases Research

Generation IGeneration I

1980s1980s

Data (Schema, “semantic data modeling)

Gen. Purpose,Broad Based

Scope of AgreementTask/ App

Domain Industry

CommonSense

Degre

e o

f Ag

reem

en

t

Info

rmal

Sem

i-Form

al

Form

al

Agreement About

Data/Info.

Function

Execution

Qos

Broad Scope of Semantic (Web)

Technology

Oth

er d

ime

nsio

ns:

how

ag

reem

ents

are

re

ach

ed

,…

Current Semantic Web Focus

Semantic Web Processes

Lots of Useful

SemanticTechnology

(interoperability,Integration)

Cf: Guarino, Gruber

What is the Semantic Web?• "The Semantic Web is an extension of the current web in

which information is given well-defined meaning, better enabling computers and people to work in cooperation." -- Tim Berners-Lee, James Hendler, Ora Lassila, The Semantic Web, Scientific American, May 2001

• Ontologies• RDF/RDFS or OWL Syntax – machine

processable• Semantic Metadata – annotation of web

resources

“An ontology is a specification of a conceptualization” (T. Gruber)

• A conceptualization is the way we think about a domain

• A specification provides a formal way of writing it down

Building Ontologies from the Ground Up When users set out to model their professional activity – Mark Mussen

Conceptualization and Ontology

http://www.w3c.it/events/minerva20040706/guarino.pdf

Everything that can be expressed in the language

OntologyConstrainingPossible InterpretationsOf what can Be expressed

Central Role of Ontology• Ontology represents agreement, represents

common terminology/nomenclature• Ontology is populated with extensive domain

knowledge or known facts/assertions• Key enabler of semantic metadata extraction from

all forms of content:–unstructured text (and 150 file formats)–semi-structured (HTML, XML) and –structured data

• Ontology is in turn the center piece that enables–resolution of semantic heterogeneity –semantic integration–semantically correlating/associating objects and

documents

Types of Ontologies (or things close to ontology)• Upper ontologies: modeling of time, space, process, etc• Broad-based or general purpose ontology/nomenclatures: Cyc,

CIRCA ontology (Applied Semantics), SWETO, WordNet ; • Domain-specific or Industry specific ontologies

– News: politics, sports, business, entertainment– Financial Market– Terrorism– Pharma– GlycO, ProPreO– (GO (a nomenclature), UMLS inspired ontology, …), MGED

• Application Specific and Task specific ontologies– Anti-money laundering– Equity Research– Repertoire Management– Financial irregularity

Fundamentally different approaches in developing ontologies at the two end of the above spectrum

Building ontologyThree broad approaches:

• social process/manual: many years, committees

– Can be based on metadata standard

• automatic taxonomy generation (statistical clustering/NLP): limitation/problems on quality, dependence on corpus, naming

• Descriptional component (schema) designed by domain experts; Description base (assertional component, extension) using automated processes from trusted knowledge sources

Option 2 is being investigated in several research projects;

Option 3 is currently supported by Semagix Freedom

SUMO -- http://ontology.teknowledge.com/

Part of the CYC Upper Ontology

http://www.cyc.com/cyc/technology/whatiscyc_dir/whatdoescycknow

SWETO (Semantic Web Testbed Ontology) Current Status

• Developed using Semagix technology for free non-commercial usage by the SW community; some initial users

• V1.4 population includes over 800,000 entities and over 1,500,000 explicit relationships among them

• Continue to populate the ontology with diverse sources thereby extending it in multiple domains, new smaller and larger release due soon; RDF and OWL versions

• Significant information for provenance/trust support [UMBC partnership]

• 97% of disambiguation performed automatically, 2% manually; not quite a high-quality as an evaluation testset (e.g., low connectivity)

• Working on test harness, quality measures, and benchmarks

Expressiveness Range: Knowledge Representation and Ontologies

Catalog/ID

GeneralLogical

constraints

Terms/glossary

Thesauri“narrower

term”relation

Formalis-a

Frames(properties)

Informalis-a

Formalinstance

Value Restriction

Disjointness, Inverse,part of…

SimpleTaxonomies

Expressive

Ontologies

Wordnet

CYCRDF DAML

OO

DB Schema RDFS

IEEE SUOOWL

UMLS

GO

KEGG TAMBIS

EcoCyc

BioPAX

GlycOSWETO

Pharma

Ontology Dimensions After McGuinness and FininOntology Dimensions After McGuinness and Finin

Gene Ontology (GO)

• Comprises three independent “ontologies”– molecular function of gene products– cellular component of gene products– biological process representing the gene product’s

higher order role.• Uses these terms as attributes of gene products in the

collaborating databases (gene product associations)• Allows queries across databases using GO terms, providing

linkage of biological information across species

http://www.geneontology.org/

GO = Three OntologiesGO = Three Ontologies

• Molecular Function – elemental activity or task– example: DNA binding

• Cellular Component – location or complex– example: cell nucleus

• Biological Process – goal or objective within cell– example: secretion

http://www.geneontology.org/

GlycO GlycOGlycO: a domain Ontology embodying knowledge of the

structure and metabolisms of glycans Contains 770 classes – describe structural features of

glycans URL: http://lsdis.cs.uga.edu/projects/glycomics/glyco is a

focused ontology for the description of glycomics

• models the biosynthesis, metabolism, and biological relevance of complex glycans

• models complex carbohydrates as sets of simpler structures that are connected with rich relationships

GlycO statistics: Ontology schema can be large and complex

• 770 classes• 142 slots• Instances Extracted with Semagix Freedom:

– 69,516 genes (From PharmGKB and KEGG)– 92,800 proteins (from SwissProt)– 18,343 publications (from CarbBank and MedLine)– 12,308 chemical compounds (from KEGG)– 3,193 enzymes (from KEGG)– 5,872 chemical reactions (from KEGG)– 2210 N-glycans (from KEGG)

GlycO taxonomyThe first levels of the GlycO taxonomy

Most relationships and attributes in GlycO

GlycO exploits the expressiveness of OWL-DL.Cardinality constraints, value constraints, Existential and Universal restrictions on Range and Domain of properties allow the classification of unknown entities as well as the deduction of implicit relationships.

Query and visualization

A biosynthetic pathwayGNT-I

attaches GlcNAc at position 2

UDP-N-acetyl-D-glucosamine + alpha-D-Mannosyl-1,3-(R1)-beta-D-mannosyl-R2 <=>

UDP + N-Acetyl-$beta-D-glucosaminyl-1,2-alpha-D-mannosyl-1,3-(R1)-beta-D-mannosyl-$R2

GNT-Vattaches GlcNAc at position 6

UDP-N-acetyl-D-glucosamine + G00020 <=> UDP + G00021

N-acetyl-glucosaminyl_transferase_VN-glycan_beta_GlcNAc_9N-glycan_alpha_man_4

The impact of GlycO

• GlycO models classes of glycans with unprecedented accuracy

• Implicit knowledge about glycans can be deductively derived

• Experimental results can be validated according to the model

N-GlycosylationN-Glycosylation ProcessProcess (NGPNGP)Cell Culture

Glycoprotein Fraction

Glycopeptides Fraction

extract

Separation technique I

Glycopeptides Fraction

n*m

n

Signal integrationData correlation

Peptide Fraction

Peptide Fraction

ms data ms/ms data

ms peaklist ms/ms peaklist

Peptide listN-dimensional arrayGlycopeptide identificationand quantification

proteolysis

Separation technique II

PNGase

Mass spectrometry

Data reductionData reduction

Peptide identificationbinning

n

1

By N-glycosylation Process, we mean the

identification and quantification of

glycopeptides

ProPreO models the phases of proteomics experiment using five fundamental concepts: DataData: (Example: a peaklist file from ms/ms raw data)

Data_processing_applicationsData_processing_applications: (Example: MASCOT* search engine)

HardwareHardware: embodies instrument types used in proteomics (Example: ABI_Voyager_DE_Pro_MALDI_TOF)

Parameter_listParameter_list: describes the different types of parameter lists associated with experimental phases

TaskTask: (Example: component separation, used in chromatography)

ProPreOProPreO - Experimental Proteomics - Experimental Proteomics Process OntologyProcess Ontology

*http://www.matrixscience.com/

Semantic Annotation of Scientific DataSemantic Annotation of Scientific Data

830.9570 194.9604 2580.2985 0.3592688.3214 0.2526

779.4759 38.4939784.3607 21.77361543.7476 1.38221544.7595 2.9977

1562.8113 37.47901660.7776 476.5043

ms/ms peaklist data

<ms/ms_peak_list>

<parameter instrument=micromass_QTOF_2_quadropole_time_of_flight_mass_spectrometer

mode = “ms/ms”/>

<parent_ion_mass>830.9570</parent_ion_mass>

<total_abundance>194.9604</total_abundance>

<z>2</z>

<mass_spec_peak m/z = 580.2985 abundance = 0.3592/>

<mass_spec_peak m/z = 688.3214 abundance = 0.2526/>

<mass_spec_peak m/z = 779.4759 abundance = 38.4939/>

<mass_spec_peak m/z = 784.3607 abundance = 21.7736/>

<mass_spec_peak m/z = 1543.7476 abundance = 1.3822/>

<mass_spec_peak m/z = 1544.7595 abundance = 2.9977/>

<mass_spec_peak m/z = 1562.8113 abundance = 37.4790/>

<mass_spec_peak m/z = 1660.7776 abundance = 476.5043/>

<ms/ms_peak_list>

Annotated ms/ms peaklist data

Semantic annotation of Scientific Semantic annotation of Scientific DataData

Annotated ms/ms peaklist data

<ms/ms_peak_list>

<parameter

instrument=“micromass_QTOF_2_quadropole_time_of_flight_mass_spectrometer”

mode = “ms/ms”/>

<parent_ion_mass>830.9570</parent_ion_mass>

<total_abundance>194.9604</total_abundance>

<z>2</z>

<mass_spec_peak m/z = 580.2985 abundance = 0.3592/>

<mass_spec_peak m/z = 688.3214 abundance = 0.2526/>

<mass_spec_peak m/z = 779.4759 abundance = 38.4939/>

<mass_spec_peak m/z = 784.3607 abundance = 21.7736/>

<mass_spec_peak m/z = 1543.7476 abundance = 1.3822/>

<mass_spec_peak m/z = 1544.7595 abundance = 2.9977/>

<mass_spec_peak m/z = 1562.8113 abundance = 37.4790/>

<mass_spec_peak m/z = 1660.7776 abundance = 476.5043/>

<ms/ms_peak_list>

Syntax for Onologies and Metadata• Why not use XML?• Why use OWL?• Or for that matter why RDF?• So many questions …

• XML – surface syntax for structured documents– imposes no semantic constraints on the meaning of these documents.

• XML Schema – is a language for restricting the structure of XML documents.

• RDF – is a datamodel for objects ("resources") and relations between them, – provides a simple semantics for this datamodel– these datamodels can be represented in an XML syntax.

• RDF Schema – is a vocabulary for describing properties and classes of RDF resources – with a semantics for generalization-hierarchies of such properties and classes.

• OWL – adds more vocabulary for describing properties and classes:

• relations between classes (e.g. disjointness), • cardinality (e.g. "exactly one"), • equality, richer typing of properties, • characteristics of properties (e.g. symmetry), and enumerated classes.

From XML to OWL

Exp

ressive Po

wer

http://en.wikipedia.org/wiki/Semantic_web#Components_of_the_Semantic_Web

NO SEMANTICS

Relationships as first class objects– key to Semantics

SEMANTICS

From an alphabet to a Language• XML

– “XML is only the first step to ensuring that computers can communicate freely. XML is an alphabet for computers and as everyone traveling in Europe knows, knowing the alphabet doesn’t mean you can speak Italian of French.” – Business Week, March 18th 2002

– Example cited by Nicola Guarino in http://www.w3c.it/events/minerva20040706/guarino.pdf

• RDF/RDFS and OWL would therefore be akin to the language computers use to communicate

• And ontologies represented in these languages would be akin to the exact interpretations of the concepts being communicated

Syntax for Onologies and Metadata• RDF

– A simple W3C standard used to describe Web resources

– Relationships in RDF (Properties), are binary relationships between two resources or a resource and a literal

– Resources take on the roles of Subject and Object respectively.

– The Subject, Predicate and Object compose an RDF statement

http://www.w3.org/RDF/

What is RDF?

• Resource Description Framework• Proposed as the base semantic web

language• Data model for describing properties of

resources• Statements about properties and values of

web resources• Machine-understandable metadata

RDF Elements

• Resource: – Something that can be described/referenced– Identified by a URI

• Property:– Relationship from a resource to a value:

• Another resource• An atomic value/literal

• Statement:– resource -> property -> value

RDF Statement

RDF Model

• Formal Data Model– Directed labeled graph

• Nodes: resources or literals• Edges: properties (relationships/attributes)• Labels: URIs of nodes and edges

– Collection of triples• subject (resource)• predicate (property)• object (resource or literal)

• W3C recommendation

Graph Model

Triple Model

Subject Predicate Object

Shaguille O’Neal plays_for Miami Heat

Kobe Bryant plays_for LA Lakers

LA Lakers competes_with Miami Heat

Phil Jackson coaches LA Lakers

RDF Syntax

• Formal syntax• Encoded in XML• Unambiguous property names and values• RDF adds rules for interpretation• W3C recommendation

Example

<sample:Athlete rdf:about="&sample;Kobe_Bryant"> <rdfs:label xml:lang="en">Kobe Bryant</rdfs:label> <sample:plays_for rdf:resource="&sample;LA_Lakers"/></sample:Athlete><sample:Athlete rdf:about="&sample;Shaquille_ONeal"> <rdfs:label xml:lang="en">Shaquille O'Neal</rdfs:label> <sample:plays_for rdf:resource="&sample;Miami_Heat"/></sample:Athlete><sample:Team rdf:about="&sample;LA_Lakers" <rdfs:label xml:lang="en">LA Lakers</rdfs:label></sample:Team><sample:Team rdf:about="&sample;Miami Heat" <rdfs:label xml:lang="en">Miami Heat</rdfs:label> <sample:competes_with rdf:resource="&sample;LA_Lakers"/></sample:Team><sample:Coach rdf:about="&sample;sample1_Instance_8" <rdfs:label xml:lang="en">sample1_Instance_8</rdfs:label> <sample:coaches rdf:resource="&sample;LA_Lakers"/></sample:Coach>

What is RDFS?

• RDF Vocabulary Description Language • (RDF Schema)• Extension of RDF: same data model

– graph or triples

• A hierarchy of classes • A hierarchy of properties relating classes• W3C recommendation

RDF Schema

RDF Instances

&r4

&r8

&r2

BankAccount

no

FFlyer

Payment

paidby

typeOf(instance)

subClassOf(isA)

subPropertyOf

paidby

“Marwan”

“Al-Shehhi”

&r7

&r1

fname

lname

purchased

purchased

“M’mmed”

“Atta”

Ticket

Flight

forflight

String

num

ber

purchased

Client

fnameln

ame

String

String

Passenger

FFNo

String

Customer

Cash

ffliernocr

edite

dto

fflie

rno

creditedto

paidby

holder

&r9

float

amountpu

rcha

sed

for

&r5 &r6

purchased

for

CCard

fname

lname

“Abdulaziz ”

“Alomari ”

fname

lname

“XYZ123”

&r3

String

ffid

ffid

&r11

holder

paidby

holder

RDFS Core Classes

• rdfs:Class– Class of resources that are RDF classes– Instance of rdfs:Class

• rdfs:Resource– All things being described– The class type of everything in RDF(S)– Instance of rdfs:Class

• rdf:Property– Class of RDF properties– Instance of rdfs:Class

http://www.w3.org/TR/rdf-schema/

RDFS Core Properties

• rdfs:type– A resource is an instance of a class– Instance of rdf:Property

• rdfs:subClassOf– All instances of a class are also instances of another

class– Instance of rdf:Property

• rdfs:subPropertyOf– All resources related by one property are also related by

another property– Instance of rdf:Property

RDF Core Properties

• rdfs:range– All values of a property are instances of one or more

class• The value MUST be an instance of all range classes

– Instance of rdf:Property

• rdfs:domain– All resources with the given property are instances of

one or more class• The resource MUST be an instance of all domain classes

– Instance of rdf:Property

OWL, W3C definition

• “language for defining structured, Web-based ontology

which enables richer integration and interoperability of data across application boundaries”

http://www.w3.org/2004/OWL/

OWL Use Cases

• Web portals• Multimedia Collections• Corporate web site management• Design documentation• Agents and services• Ubiquitous computing

OWL Design Goals

• Shared ontologies• Ontology evolution• Ontology interoperability• Inconsistency detection• Expressivity vs. scalability• Ease of use• Compatibility with other standards• Internationalization

What’s in OWL, but not in RDF

• Ability to be distributed across many systems– By means of owl:imports (similar to ‘include’ in

C/C++)

• Scalable to Web needs (?)• Compatible with Web standards for:

– accessibility, and– Internationalization

• Open and extensible

OWL open and extensible

• RDF Schema (meta-modeling facilities, i.e. classes of classes)

• OWL Full• OWL DL (Description Logics)

• OWL Lite– targeting tool builders

owl:Class

• Sub class of Class in RDF

• Better to forget about classes of classes

• Top-most class: owl:Thing

OWL Properties

ObjectProperties

Ana owns Cuba

Is range aliteral / typed value ?

then ERROR

Data typeProperties

Ana age 25

• XML Schema data types supported– DB people happy

Transitivity of properties

X p1 YY p1 Z

implies X p1 Z

• Transitivity existed already in RDF– “subClassOf”, and “subPropertyOf”

• Example: located_in GeorgiaAtlantalocated_in U.S.A.located_in

located_in

Symmetric properties

X p1 Y

implies X p1 Y

PortugalSpainhas_border_with

SpainPortugalhas_border_with

Functional Properties

X p1 Y X p1 Z

imply Z is the same as Y (they describe the same)

• example, p1 = has_name

&r1Portugalhas_capital

&r2Portugalhas_capital

Result: &r1 and &r2 represent the same entity

Inverse Functional Properties

Y p1 AZ p1 A

imply Z is the same as Y(they describe the same)

• example, p1 = has_email

finin@umbc.edu&r1:Tim Fininhas_email

&r2:Timothy Fininhas_email

Result: &r1 and &r2 represent the same entity

OWL Cardinality

• min Cardinality• max Cardinality

• “Cardinality”– When min = max

• has Value– belongs to the class if it has the value

OWL Tools

• Pellet (umd.edu)

– DLbased reasoner implemented in Java

• Euler– an inference engine supporting logic based

proofs. Finds out whether a given set of facts support a given conclusion

• FaCT (Ian Horrocks)

– DL classifier that can also be used for modal logic satisfiability testing

RDF Storages

• Jena• Sesame• Redland• Triple• 3store• RDFSuite• RDFStore• Kowari• Yars• Brahms

developed at LSDIS

• Variety of available storages

• Different APIs and languages

• Support from RDF to OWL-full– even reasoning

• Storage and query approach: graph Vs. triple-centric

http://www.w3.org/2001/sw/Europe/reports/rdf_scalable_storage_report/

Jena

• Implemented in Java by HP Laboratories• Support for RDF, RDFS and OWL• Reasoning / inference engine• Support for reified statements• In-memory and persistent storage

(Oracle, MySQL, PostgreSQL)

• Query language: RDQL, SPARQL• Read/write RDF in RDF/XML, N3 and N-Triples

format• Triple-centric organization and API

Jena – graph abstraction

• Graph interface is separated from (persistent) triple storage layer

• Special support for different types of graphs - optimized for performance

• Support operations like add, delete, find.

“Efficient RDF Storage and Retrieval in Jena2” Kevin Wilkinson, Craig Sayers, Harumi Kuno, Dave Reynolds

Jena – query processing• Converting multiple patterns in query into one query

to DB• Use DB query optimizer instead of executing multiple

queries from Jena level• Cluster properties that are likely to be accessed

together - optimize for common patterns• Associate a table with pattern (best) or span pattern

between tables (requires join operation)• Query may span between different graphs, but it can

be optimized only if they are in the same database

Redland, Rasqual, Raptor

• Storage for RDF triples - do not implement any language by itself

• This is the main module to include in RDF manipulation system

• Implemented in pure C for portability• Rich API enables to build modules on top of it • Rasqual - RDF query module

– RDQL– SPARQL

• Raptor - a fast RDF parser

Redland

• API available in different languages– C, C#, Java, Perl, Python, PHP, Ruby, Tcl

• API for manipulating– triples, URI/literals, graphs

• Portable - can built in most OSes• Scalable to handle millions of triples

– while using of persistent storage– but indexing is very space-consuming

• Support for context and hierarchy of models

Redland - model

• Abstraction of model to support different storages

• In-memory and persistent models– BerkeleyDB, 3store,

MySQL

• Rich, triple-centric API

„The Design and Implementation of the Redland RDF Application Framework” - David Beckett

Sesame

• Implemented in Java• Database independent

– idea of SAIL (Storage Abstraction Interface Layer)

• Scalable architecture• Implementation of remote models

– can query different models over network

• Graph-centric approach• Language: RQL

Sesame - architecture• RAL - Repository Abstraction Layer

– makes Sesame storage independent– API supportes RDF Schema semantics (e.g.

subsumption reasoning)– can be stacked one on another– interface oriented for persistance storage

(DBMS, Object-Relational DB)– data returned as streams– can even use net-based RDF services (!)

• Due to poor performance, implemented cache as one of RALs– cache mainly for RDFS, as it needs code

support in reasoning (subClassOf, ...)

“Sesame: An Architecture for Storing and QueryingRDF Data and Schema Information” - Jeen Broekstra, Arjohn Kampman, Frank van Harmelen

Sesame – query module

• Query module– query plan and optimizer similar to already

known DB solutions– query is translated to a set of simple RAL calls– each leaf of the query plan can ‘evaluate itself’

and pull data from RAL– data are returned as streams– lack of optimization on storage level

Brahms

• Implemented in C++ (bindings for Java also available)

• Graph-centric approach• Designed to support large in-memory RDF graphs• Optimized for speed and memory usage

– other storages do not offer optimized in-memory implementation for large graphs

– only main memory offers fastest access - usage of persistent storage decreases performance

• In-memory storage with fast precomputed graph snapshot loading– minimize cold-start time

Brahms• Framework for fast discovery of long association

paths in large RDF bases– memory and CPU intense algorithms

• Rich API, but no query language supported– higher level query languages do not support variable

length association path queries– association path discovery algorithms operate on low-

level graph API• Overperformed Jena, Sesame and Redland during

tests for association discovery– also was able to work efficiently on much larger in-

memory graphs than other storages did not handle

“BRAHMS: A WorkBench RDF Store And High Performance Memory System for Semantic Association Discovery” (Technical report) - Maciej Janik, Krys Kochut

Why RDF languages?

• Find resources based on predicates, values, labels or associations

• SQL is not good for querying RDF data– different models: relational and graph

• XML query languages cannot deal with graph data

• Syntactic approach is not enough• Required semantic querying• Inferencing is desirable

Available query languages

• RQL• RDQL• SeRQL • Triple• SPARQL – (latest)• SquishQL• Versa• N3• RxPath• RDFQL

• Majority of languages have roots in SQL

• No single standard as SQL

• Some languages are tightly coupled with specific storages

RQL

• Based on OQL• Utilizes functional approach with support for

generalized path expressions– both nodes and edges can become variables

• Not completely compatible with RDF specification – has some additional restrictions

• Return bindings to variables (no closure)• Implemented in RDFSuite and partially in Sesame

select Res from {Res} ns:label {x} where x=“foo” using namespace ns=…

“RQL: A Declarative Query Language for RDF” - Greg Karvounarakis, Sofia Alexaki, Vassilis Christophides, Dimitris Plexousakis, Michel Scholl

RDQL

• SQL-like syntax– easy to adopt for DB users

• Can specify patterns of triples to select• Schema is not interpreted• Not closed under queries

– output as bindings to selected variables

• Implemented in Jena

select ?p, ?q where (?p <rdfs:label> “foo”) (?p <rdf:type> ?q)

“RDQL - A Query Language for RDF” (W3C Member Submission) - Andy Seaborne (HP Labs Bristol)

SeRQL

• Sesame RDF Query Language• Based on RQL and RDQL• Support for generalized path expressions

and optional matching• Query filters

– select-from-where – return variable bindings and is not closed

– construct-from-where – return matching subgraph that can be queried (closure)

“SeRQL: Sesame RDF query language” - Jeen Broekstra

Triple

• Derived from F-logic– should be easy to adopt for logic programmers

• Triples are logic expressions– S [ P O ]

• Queries and triples have the same logic representation

• Reasoning is a part of language• Does not fulfill closure property• Implemented in Triple system

FORALL X <- ( X[rdfs:label -> “foo”] )@default:ln.

“TRIPLE - A Query, Inference, and Transformation Language for the Semantic Web” - Michael Sintek, Stefan Decker

SPARQL• W3C effort to standarize query language

– best experience and requirements from different languages (like RQL, RDQL)

• Based on matching graph patterns– triples, paths, subgraphs– optional blocks and matching– matching alternatives (union) and disjunction

• Many additional operators– grouping, sorting, limit results

PREFIX foaf: <http://xmlns.com/foaf/0.1/>SELECT ?name ?mbox WHERE { ?x foaf:name ?name . OPTIONAL { ?x foaf:mbox ?mbox } }

“SPARQL Query Language for RDF” (W3C Working Draft) - Eric Prud'hommeaux , Andy Seaborne

Sample path query

“A Comparison of RDF Query Languages” - Peter Haase, Jeen Broekstra, Andreas Eberhart, Raphael Volz

Expressive power of RDF languages

“Ontology Storage and Querying” (Technical Rreport No 308) - Aimilia Magkanaraki et al.