Transcript Corporate Memory

Corporate Memory:
A framework for supporting tools for
acquisition, organization and maintenance
of information and knowledge
Marek Ciglan, Marian Babik, Michal Laclavik
Ivana Budinska, Ladislav Hluchy
Institute of Informatics, Slovak Academy of Sciences,
Dubravska cesta 9, Bratislava, 845 07, Slovakia
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Outline
• Motivation
– Project NAZOU & CM in NAZOU
• Corporate Memory Architecture
– Interaction Layer
– Manipulation Layer
– Physical Layer
• Corporate Memory Usage
• Conclusions
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Project NAZOU
• NAZOU - Tools for Knowledge Discovery, Maintenance
and Presentation
• Goals and Challenges:
– Acquire information from public sources
– Discover knowledge
– Make knowledge available in computer processable form
– In the form of Ontologies (enable automatic reasoning)
• For specific domains only!
– Jobs offers
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Tools Chain
Presentation
Offers
XHTML
Offers
Extraction
Offers
TXT, RDFS, OWL
Relevance
Dokuments
with offers
(X)HTML, TXT, DOC, PDF, ...
Crawler
Documents
(X)HTML, TXT, DOC, PDF, ...
Web
ISIM’06, Přerov 26.4.06; Corporate Memory
TXT
Conversion
Semantic
Annotation,
Reasoning
Corporate Memory
Corporate Memory in NAZOU
• Integration of tools at the data layer
• Virtualization of data resources
• Uniform way to access the data
• Uniform access even in the case of change of physical
data resources architecture
• Provide access to different kinds of data resources
– Plain text files
– Relational db
– Ontological data
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Architecture
Corporate Memory
Interaction layer
XML-RPC Connector / WS Connector (SOAP)
Java Connector
Manipulation layer
File API
Reasoning
SQL/ DB API
Jena/Sesame
Physical layer
File storage
RDQL/Ontology API
RDB storage
ISIM’06, Přerov 26.4.06; Corporate Memory
RDF DB/ RDF-RDB
Mapping
Corporate Memory
Interaction Layer
• Local java interface
• Java xml-rpc
• WS-interface
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Interaction Layer
• Local java interface
• Java xml-rpc
• WS-interface
Q: Why to provide remote interaction with CM?
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Interaction Layer
• Local java interface
• Java xml-rpc
• WS-interface
Q: Why to provide remote interaction with CM?
A: To enable distribution of computational and
storage resources.
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Distribution of resources
App1
App2
App3
Corporate Memory
Files
ISIM’06, Přerov 26.4.06; Corporate Memory
RDBM
Onto
Corporate Memory
Distribution of resources
App1
App2
Corporate Memory
App3
Files
ISIM’06, Přerov 26.4.06; Corporate Memory
RDBM
Onto
Corporate Memory
Distribution of resources
App1
App2
Corporate Memory
Files
App3
ISIM’06, Přerov 26.4.06; Corporate Memory
RDBM
Onto
Corporate Memory
Distribution of resources
Corporate Memory
App1
Files
App2
App3
ISIM’06, Přerov 26.4.06; Corporate Memory
RDBM
Ontologies
Corporate Memory
Manipulation Layer
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Physical Layer
• Files repository
– vfat, ext3
• RDBM
– MySql
• Ontological storage
– Jena
– Sesame
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
ONTO Client
• Needs:
– Common formalized version of offers
– Reasoning, inference
– Common offers presentation
• Operations (example)
– Insert() – inserts OWL models in XML/RDF format into Onto
CM
– getXML() – returns plain XML of resource from Onto CM
– executeRDQLQuery()- returns RDF ID result list for given
RDQL query
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Files Client
• Needs:
– Access to acquired original data files
– Access plaintext form of original data
– Store textual output of tools
• Operations (example):
– insert files to CM (as data stream / from uri)
– deliver file (as stream / request URI)
– list, cp, mv, delete
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
db client
• Needs:
– Store indices of content of documents and offers
– Provide fast full-text search
– Provide functionality for textual mining tools
• Operations:
–
–
–
–
execute sql queries
enable definition of query templates
execution of stored procedures
update/insert/delete statements and templates
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Conclusions
Corporate Memory
•
•
•
•
Utility for integration of tools at the data layer
Virtualization of data resources
Uniform way to access the data
Provide access to different kinds of data
resources
– Plain text files
– Relational db
– Ontologies
ISIM’06, Přerov 26.4.06; Corporate Memory
Corporate Memory
Thank you for your attention!
ISIM’06, Přerov 26.4.06; Corporate Memory