You Cannot ReSIST

You Cannot ReSISTYou Cannot ReSIST

Hugh GlaserHugh GlaserElectronics & Computer ScienceElectronics & Computer Science

University of SouthamptonUniversity of Southampton

DSSE, 28th February 2007DSSE, 28th February 2007

With

• Ian Millard

• Afraz Jaffri

• Benedicto Rodriguez

• ReSIST Partners– esp. Brian Randell

ReSISTReSIST

• EU Network of Excellence– (www.resist-noe.org)– Resilient Systems– 15 partners across Europe

• Resilience Knowledge Base

• Resilience-Explicit Computing

• Dissemination

Semantic Web ChallengeWinner 2003

• CS AKTive Space– Gather data – UK People, projects, publications

• Research funding

• Top Universities

– Geographical presentation

• AKT Project (www.aktors.org)

The Challenges• Organisational Intelligence

– Who is doing what where?– What impact are they having?

• Integrating resources – – Research Councils, HEFCE,

HESA, RAE, EU, Institutional DBs and web sites, ePrints, citation services…

• Information is heterogeneous and distributed– Not under own control– Not in a common format…

• Presenting to users

http://triplestore.aktors.org/demo/AKTiveSpace/http://triplestore.aktors.org/demo/AKTiveSpace/

Conceptual Architecture

3store

AKTReferenc

e

AKTReferenc

e

data sources

gatherers and

mediators

ontology knowledge repository

(triplestore)

applications

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

So what is RDF...?

• Resource Description Framework

• W3C recommendation, comes predominantly from Semantic Web research efforts

• A modelling language designed to represent facts about resources in the World Wide Web

• Can be used to model any abstract domain, things do not have to be accessible on the web to be described in it

RDF: Basic components

• RDF graphs are formed by triples

subject predicate object

Open - SPARQL

PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>PREFIX owl: <http://www.w3.org/2002/07/owl#>PREFIX akt: <http://www.aktors.org/ontology/portal#>PREFIX akts: <http://www.aktors.org/ontology/support#>PREFIX wiki: <http://resist.ecs.soton.ac.uk/wiki/>PREFIX course: http://www.resist-noe.org/ontology/courseware#SELECT * WHERE { ?paper akt:has-title ?title . ?paper akt:has-author ?person . ?paper <http://www.aktors.org/ontology/extension#has-

abstract> ?abstract}LIMIT 10

At the Centre

Important Components3store and CRS

• Open source semantic store

• Scalable– ReSIST - 50 million facts– Garlik - 1,000,000,000s

• Uses lack of predefined schema

• Cluster

• Distributed queries next– Across external stores

Triplestore Evaluation

• Consistent Reference Service (CRS)– Where is the Semantic Web here?

• Technologies used

• Who is “Jongho Nang”?

• Which paper is “MPEG 비디오 스트림에 대한 MC-DCT 영역에서의 자막 처리 방법” ?

• Problems– Finding co-referents

• NLP, Statistical, Structural

– Making the knowledge available• Exposing a DB?• (Semantic) Web services• Home page (FOAF file)?

• Identify co-referents– Against different policies - fit for purpose

• Are false positives acceptable?

• Distributed and federated (compare OAI-PMH)• Not a Naming Authority

• Query by URI:– http://dblp.uni-trier.de/People#35ff57fb

• Response from CRS:<ns1:Bundle rdf:about="http://www.aktors.org/ontology/coref#bundle-

a458455fb66f997bf4a67488dbbdbdae"> <ns1:duplicate rdf:resource="http://citeseer.ecs.soton.ac.uk/#P147667"/> <ns1:duplicate rdf:resource="http://dblp.uni-trier.de/People#35ff57fb"/> <ns1:insertedOn>20061126</ns1:insertedOn> <ns1:hasCanon rdf:resource="http://citeseer.ecs.soton.ac.uk/#P147667"/></ns1:Bundle>

• Need to compute fixed point across CRSes• But “canon” can be used• Maintained according to need

– Automatic– By Hand

Real Semantic Web

• The RKB Explorer– Accessing live RDF

• Current ECS?

– Query over multiple SPARQL endpoints• Triplestore per major resource• External triplestores

– Universities– Funding agencies (CORDIS, NSF, etc)

• CRS– Different sources because of different curation– Different types: convenience, good engineering– Different policies: application needs, trust

• At last the Ontology is of some use.

Future for ReSIST• Extend the use of Semantics• Resilient-Explicit Computing

– Model expert knowledge– Model processes– Model components

• Support Engineer/Scientist– Move effectively between

• System design• Knowledge Base• People

– To choose cost, characteristics, etc

• Support Run-Time Deployment– Dynamic Reconfiguration

Grid

E-Scientists

Entire E-Science CycleEncompassing experimentation, analysis, publication, research, learning

Institutional Archive

LocalWebPublisher

Holdings

Digital Library

E-Scientists Graduate Students

Undergraduate Students

Virtual Learning Environment

E-Experimentation

E-Scientists

Technical Reports

Reprints

Peer-Reviewed Journal &

Conference Papers

Preprints & Metadata

Certified Experimental

Results & Analyses

Data, Metadata & Ontologies

In the Wild