ISIBANG 2007 01 31 jk

Slide1: 

Towards a knowledge exchange infrastructure for Agricultural Research and Technology The Role of the Agricultural Ontology Service Dr.rer.nat. Johannes Keizer Knowledge Exchange and Capacity Building Division Food and Agriculture Organization of the UN International Conference on Semantic Web and Digital Libraries Bangalore, 20-23 February 2007

Slide3: 

World Food Summit 1996 Reducing Hunger and Poverty in the World by 50% in 2015 "The Rome Declaration calls upon us to reduce by half the number of chronically undernourished people on the Earth by the year 2015 .... If each of us gives his or her best I believe that we can meet and even exceed the target we have set for ourselves.""The Rome Declaration calls upon us to reduce by half the number of chronically undernourished people on the Earth by the year 2015 .... If each of us gives his or her best I believe that we can meet and even exceed the target we have set for ourselves.“ "We have the possibility to do it. We have the knowledge. We have the resources. And with the Rome Declaration and the Plan of Action, we've shown that we have the will." Our Division’s Goal: Combating Hunger by Facilitation of Knowledge Exchange

This Presentation: 

This Presentation Data, Information, Knowledge and the Semantic Web, An Infrastructure for Agricultural Science and Technology The openAccess publishing Paradigm Open Archives: The AGRIS OAI Architecture Proposal Crop and Weather Data: Meteo Broker A general model The Agricultural Ontology Service The AGROVOC Concept Server AgMes NeOn Ontologies under Development The Process and the Vision

Slide6: 

Data, Information, Knowledge and the Semantic Web

Slide7: 

Data Information Knowledge Wisdom Judgement Inferences Structure Codification The Stair to Wisdom

Slide8: 

In the first 10 years the Internet has been mainly a space for publishing Operations which were done on hardcopy were digitized, but workflows and processes remained the same There is no conceptual difference between a card catalogue and an electronic catalogue as mostly used nowadays The virtual connection of resources on the Internet stayed virtual, in reality many distinct silos of Information and Knowledge were created Partly Steps backed, compared to Interoperability protocols in the Library area z39.50) confined to ILMS WebServices no real kickstart (in FAO 1!) The only real large scale integration success is Google, everything else is corporate (Amazone, ebay,) Worst Situation in Technology and Science Tim Berners Lee: The web is like a big relational database without relations On which Step of the Stair is the Internet?

Slide9: 

Web 20 tries to introduce this relational expects by human efforts Social Bookmarking is a way to create relations and networks between cataloguers on the web, Delicous and FlickR are kind of community catalogues RSS feeds try to get different information silos on the web in connection But again silos are created. Single Web2 spots do not communicate with each other, Consistence within such spots is low, scalability not proven

Slide10: 

©Berners-Lee .........the big picture

Slide11: 

A Knowledge Infrastructure for Agricultural Science and Technology The Open Access Publishing Paradigm Open Archives: The AGRIS OAI Architecture Proposal Crop and Weather Data: Meteo - Broker A general model

OAI: how to get the different repositories to communicat: 

OAI: how to get the different repositories to communicat Dublin Core Metadata Exchance Schema (1996) Open Archive Initiative Metdata Harvesting protocol (OAI MHP)

...our starting point was simple: exchange of bibliographical data between different repositories: 

...our starting point was simple: exchange of bibliographical data between different repositories © Salokhe

Why not simple DC?: 

Why not simple DC? Citation For example, the citation information using AGRIS AP is displayed as: <ags:citation> <ags:citationTitle>Journal of Agricultural Research and Extension (Thailand)</ags:citationTitle> <ags:citationTitle>Warasan Wichai Lae Songsoem Wichakan Kaset</ags:citationTitle> <ags:citationIdentifier scheme="ags:ISSN">0125-8850</ags:citationIdentifier> <ags:citationNumber>18(2) p.1-12</ags:citationNumber> <ags:citationChronology>Apr-Sep 2001</ags:citationChronology> </ags:citation> The dumbing down process would result in the information being merged into various fields and presented as: <dc:relation> Journal of Agricultural Research and Extension (Thailand); Warasan Wichai Lae Songsoem Wichakan Kaset; 18(2) p.1-12; ISSN: 0125-8850; Apr-Sep 2001</dc:relation> Relation For example, the relation information using AGRIS AP is displayed as: <dc:relation> <dcterms:isVersionOf>http://www.fao.org/agris/agmes/DC1-FAO1.doc </dcterms:isVersionOf > <ags:relationHasTranslation>ftp://ftp.fao.org/fao/W8270c.pdf </ags:relationHasTranslation> </dc:relation> The dumbing down process would result in the information being merged into various fields and presented as: <dc:relation> http://www.fao.org/agris/agmes/DC1-FAO1.doc</dc:relation> <dc:relation> ftp://ftp.fao.org/fao/W8270c.pdf </dc:relation>

Slide15: 

The AGRIS – Application Profile <?xml version="1.0"?> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:dcterms="http://purl.org/dc/terms/" xmlns:ags="http://www.fao.org/agris/agmes/schemas/1.0/ags#"> <rdf:Description about="http://www.fao.org/docrep/008/ae909e/ae909e00.htm"> <dc:title> AGRIS: Guidelines for Description of Information Objects for International Information System on Agricultural Sciences and Technology </dc:title> <ags:creator> FAO, Rome (Italy). Library and Documentation Systems Div. </ags:creator> <dc:subject>Metadata Standards; Guidelines; Dublin Core; Document-like Information Objects; Application Profile; Information Exchange</dc:subject> <dc:subject> <ags:subjectClassification> <value>C30</value> <rdfs:label>Documentation and information</rdfs:label> <rdfs:isDefinedBy rdf:resource="http://www.fao.org/agris/"/> </ags:subjectClassification> </dc:subject> <dc:subject> <ags:subjectClassification> <value>Z699.35.M28</value> <rdfs:isDefinedBy rdf:resource="http://lcweb.loc.gov/catdir/cpso/lcco/lcco.html"/> </ags:subjectClassification> </dc:subject> <dc:subject> <ags:subjectThesaurus> <rdfs:label>AGRIS; DATA PROCESSING ; METHODS ; TRAINING ; LIBRARIANSHIP ; STANDARDS; INFORMATION STORAGE </rdfs:label> <rdfs:isDefinedy rdf:resource="http://www.fao.org/agrovoc/"/> </ags:subjectThesaurus> </dc:subject> <dc:subject> <ags:subjectThesaurus> <rdfs:label>information processing; information systems; information storage; AGRIS </rdfs:label>

Slide16: 

Figure 3. The Data Providers : open archives. A data provider can export metadata in different metadata formats (Figure 3). The Unqualified Dublin Core is the basic metadata format for OAI Protocol, but the function of DC is basically to facilitate the exchange of metadata at cross-domain level, to enable the communication outside of the specific communities. For AGRIS Community, as other subject networks, it is necessary the use of a richer metadata format. For that AGRIS Iniative will encourage the export in AGRIS AP in order to guarantee a high level in the quality of description of agricultural information resources. The AGRIS OAI data flow

Slide17: 

The AGRIS OAI Search Engine International Conference on Semantic Web and Digital Libraries

Slide18: 

The AGRIS OAI Search Engine

Slide19: 

Home The AGRIS OAI - Network

AGRIS OAI: Implementation Examples: 

AGRIS OAI: Implementation Examples NARIMS - (Egyptian National Agricultural Information Management System) KAINET (Kenya Agricultural Information Network) BIBSYS (Norwegian Agricultural University) GFIS (Global Forestry Information Services) Many AGRIS centres) .. but far, far away from a critical mass

Slide21: 

Over 22,000 stations of 25 databases MeteoBroker: Heterogeneous scientific data

MeteoBroker: Heterogeneous scientific data: 

MeteoBroker: Heterogeneous scientific data A lot of digital data sets are continuously produced in agricultural experimental stations Using ordinal software such as spread sheet applications But they are likely to be kept in local stations and scientist level The data sets are isolated and hardly integrated among different locations How to ease data publication for merging and sharing for end users

MeteoBroker: semantic Organization of weather data: 

MeteoBroker: semantic Organization of weather data Separated crop data are hard to be integrated with different resources, e.g. weather data Heterogeneity e.g. Models constructed using local data are only applicable locally How to integrate crop data with weather data upon user’s request

MeteoBroker: If merging and sharing are possible: 

MeteoBroker: If merging and sharing are possible End users can freely combine separated data sets from different locations and perform analysis on them Datamining over the huge amount of data sets becomes real and we can possibly find out unknown facts Integration with completely different resources e.g. weather data becomes also possible Constructing new model becomes quite easy Integrated data help model test and verification.

MeteoBroker: The Infrastructure Grid: 

MeteoBroker: The Infrastructure Grid

Meteo Broker: the architecture: 

Meteo Broker: the architecture Intelligent Broker Inference Engine Dynamic DB Wrapper Item Definition OWL Station metadata RDF Metadata database Meteorological databases DB DB 2. Request 3. Request metadata 4. Request data 1. Register DB

Slide27: 

Common exchange layer (OWL, RDFS) (Vocabularies, Ontologies,) ©Liang/Sini/Salokhe ..a semantic webspace for Agricultural Research and Technology Data Grid

Slide28: 

©Liang/Sini/Salokhe

Slide29: 

Getting Interoperability: The Agricultural Ontology Service Agrovoc Exchange Schemata Ontologies

Slide30: 

http://www.fao.org/aims

Slide31: 

AGROVOC

Slide32: 

AGROVOC

Slide33: 

AGROVOC

Slide34: 

AGROVOC

Slide35: 

AGROVOC

Agrovoc Conceptserver Workbench: Overall design: 

Agrovoc Conceptserver Workbench: Overall design ipunt AOS/CS Workbench concordance pattern-matching multilingual input

Agrovoc ConceptServer Workbench:Features: 

Agrovoc ConceptServer Workbench: Features Text processing Corpus Creation Corpus Analysis Manage Concepts, Terms, Relationships Classification Schemes Quality Assurance Versioning and Deployment Other functionalities Search Import / Export Validations Administration Help

Slide38: 

Metadata Exchange Schemas

Geo-Political Ontology: 

Geo-Political Ontology Class Area: Groups: EconomicRegion GeographicalRegion Organization SpecialGroups Territory: Disputed NonSelfGoverning Other SelfGoverning isValidFrom (in years) isValidUntil (in years) isSuccessorOf isPredecessorOf hasOfficialName (string) sub properties for all languages hasShortName (string), subproperties for all languages hasCode sub properties for all classifications hasBorderWith dependsOn (domain: non-self-governing territories, range: self-governing-territories)

Fishery Ontology (1): 

Fishery Ontology (1) Ontologies built from existing classifications schemas and thesauri. Biological species: 44,100 11000 species, 4 langs, taxonomic and ISCAAP codes Water bodies: 1,500 300 water division, 5 codes Land areas: 25,000 250 territories, 5 langs, 2 names, 4 codes. economic regions, geographical regions, organizations ASFA thesaurus: 22,000 11000 entries, code AGROVOC thesaurus - Fisheries: 42,000 7000 terms, 6 languages Commodities: 6,000 Over 150 interesting resources identified in D7.2.1. Around 30 can be really useful and will be detailed.

Fishery Ontology (2): 

Fishery Ontology (2)

Slide42: 

Common exchange layer (OWL, RDFS) (Vocabularies,Ontologies,) ©Liang/Sini/Salokhe ..a semantic webspace for Agricultural Research and Technology Data Grid

A semantic space in the web:Network of Data and Service Providers with agreed procedures and standards – and common ontological layers: 

A semantic space in the web: Network of Data and Service Providers with agreed procedures and standards – and common ontological layers Data Provider: exposes institutional open archives of data and information Development Agencies and NGOs Research Institutions Industry Information Centres Service Provider: provides services based on these institutional open archives Libraries and other traditional Aggregators Thematic or Regional Centres of Excellence The data providers themselves

Slide44: 

2000 Brussels meeting and AGstandards (AgMES) initiative 2001 Launch of the Agricultural Ontology Service 2002 AOS Workshops 2003 Release of AgMES NameSpace as first AOS element Metadata Elements for Document Like Objects 2004 AGRIS Application Profile for ARD publications 2005 Implementation of AGRIS AP in various bibliographical databases 4 new language versions of AGROVOC Booming downloads of AGROVOC as a de facto Standard AIMS website released 2006 AGRIS repository in AGRIS AP XML published AGROVOC OWL Model ready and coding of AGV-concept server started Grant from the EU for a system of fishery ontologies in a 4 years project Ontologies for geopolitical and organization Information started ConceptPaper to transform the AGRIS repository in an Ontology The Implementation

AOS – a “business model”: 

AOS – a “business model” A consortium of Information Providers and Consumers providing a clearinghouse for semantic standards in the area of Agriculture, Food Security and Rural Development One stop access to agreed standards (Ontologies, Metadataschemas, Vocabularies…) Establishment of mechanisms to agree on common standards and procedures Collaborative efforts in maintaining standards Registration and documentation of common standards (namespaces, application profiles, protocols) Organization of seminars and workshops to further develop and promote the use of semantic standards Participation in semantic web activities to get funding for specific projects

Conclusions: 

Conclusions We are only at the beginning of the Knowledge and Information Age We have bits and pieces for a Knowledge Exchange Infrastructure in Agricultural Research and Technology- but we have to put them together Community building and collaboration is necessary to achieve semantic interoperability between different players

Thank you: 

Thank you Stefano Anibaldi {Stefano.Anibaldi@fao.org} Marta Iglesias {Marta.Iglesias@fao.org} Gudrun Johannsen {Gudrun.Johannsen@fao.org} Stefka Kaloyanova {Stefka.Kaloyanova@fao.org} Irene Onyancha {Irene.Onyancha@fao.org} Gauri Salokhe {Gauri.Salokhe@fao.org} Margherita Sini {Margherita.Sini@fao.org} Imma Subirats {Imma.Subirats@fao.org} Johannes Keizer {Johannes.Keizer@fao.org} and my team at FAO and credits to our network of collaborators: Dagobert Soergel, Anita Liang, Boris Lauser, Chang Chun, Asanee Kawtrakul, Caterina Caracciolo, Maria Grazia Bovo, Wang Zhong, Ze Li, Jai Haravu, ARD Prasad, Jayanta Chatterjee, and many others