Transcript スライド 1
AGENTPRO and Points of Views on Ontology Ryusuke Masuoka Fujitsu Laboratories of America, Inc. College Park, MD February 6, 2002 AGENTPRO • Give an Enterprise Information Integration (EII) solution based on agent technology • AGENTPRO v1.0 released in August, 1999 based on SAGE • AGENTPRO v3.1 is going to be released in February, 2001 • http://www.interstage.com/agent_mf.html Architecture of AGENTPRO Browser Browser Browser Browser User agent User agent User agent User agent HTML etc. ACL Facilitator ファシリテータ Facilitator Database agent Database agent Database agent Database agent Database Database Database Database SQL etc. Structure of AGENTPRO V1 OAT Config. Config. Advertise Config. Ontology Definition Ontology Translation DB SymfoWARE Oracle Access Config. Advertise Mapping 検索 Application User Interface User Agent Facilitator (Mediator Agent) Database Agent Application Agent User Defined Logic INTERSTAGE Provided by the product File by users Program by user ACL: Agent Communication Language How It Works User Agent (Retailers) Facilitator query DB Agent App. Agent unadvertise /advertise Database Application forwarded queries local queries results results merged results SCM prototype by agent system This prototype was built for an imaginary TV manufacturer and associated parts suppliers. By agentifying databases and applications at a factory, shipment center, sales center, and parts suppliers distributed in a country, mediation by the facilitator enables the prototype to answer a query by a user at the headquarters of the TV manufacturer regarding the delivery date for a specified amount of product. Points of view on ontology • They need different views, then different words to express them. – They need to differentiate themselves from others (that include themselves from yesterday.) – That is their raison d'être in borderless world realized by the Internet. • Ontologies are not stationary. – Do not take them as given. – That is the exactly reason why we need ontologies • If they are stationary, you can just write conversion software (you can use whatever you want, C, Assembler, CORBA, …). Once you do that, you can stick to it for ten years. • Because they are dynamic, it is worthwhile to have articulation of semantics (i.e. ontology). • That drives the current trend … CORBA -> XML -> Semantic Web Points of view on ontology (cont.) • I agree with “mapping” instead of “merging” between ontologies by Leo. – Can accommodate continuous changes/evolution of ontologies • When new views are introduced, a new ontology can be used between those who adapt it, while assuring backward compatibilities with those who use older ones – Can achieve maximum interoperability • Merged ontologies (or standards) take resources (time, manpower, …) and need to ignore the differences. – Can lose some of important features • Does not totally deny the role of standards as a bottom line. (or you need n-to-n mappings.).