Presentation on theme: "Dr. Leo Obrst Information Semantics Command & Control Center July 17, 2007 Ontologies Can't Help Records Management Or Can They?"— Presentation transcript:
Dr. Leo Obrst Information Semantics Command & Control Center July 17, 2007 Ontologies Can't Help Records Management Or Can They?
2 Ontologies Can't Help Records Management Records Management & Archiving have problems very different from other applications So, Records Management does not address: –Structured data: Database storage, access, query/retrieval –Unstructured data: meta-data tagging/indexing, information extraction, categorizing, retrieval –Media types: text, graphic, audio, voice, video, etc. –Configuration management, versioning –Policy and governance, business rules, –Intellectual property, privacy, security –Software services and systems that support all the above If this is so, then ontologies cant help Records Management Or does Records Management address these?
3 Records Management, Archiving, & Ontologies Records Management & Archiving has most of the same issues as does "live" data: –How can you find what you are interested in? –How can you deal with media types: current, past, future? –How do you protect past content and yet adhere to then and future intellectual property and security issues? –All of the stuff on the previous page applies Ontologies and other semantic technologies can assist with all of these issues –You can utilize tagging/indexing of data (structured, unstructured, semi- structured) with respect to given semantics: ontologies, conceptual models, thesauri, taxonomies, but –You have to retain references to those resources, their media types, and the tagging/indexing appropriate then and –Ways to access those resources now, where now is some indeterminate time in the future –Therefore, some way of indexing archives with respect to time is required; however, this is not the obvious statement it seems –You must keep track of the versions of the resources, their tagging/indexing methods (and computer software that can generate and retrieve indexes)
4 Mapping Issues Involved in the issues of both major points are mapping issues: –The ontology resources may require other resources they either: Import Map to (align with, utilize, etc.) And along with those resources' media types and software accessors –How do you create an architecture which can map to resources that are in the future, and not yet conceived? Assume that extensibility issues will be sufficiently dealt with by the future archivers for all of the past resources, accessors, etc., or Build a general enough architecture today that will permit, in fact mandate, this? –All of the above is configuration management and versioning, except not only applied syntactically, but semantically.
5 Ontology Spectrum: Application Logical Theory Thesaurus Taxonomy Conceptual Model Expressivity Categorization, Simple Search & Navigation, Simple Indexing Synonyms, Enhanced Search (Improved Recall) & Navigation, Cross Indexing Application Enterprise Modeling (system, service, data), Question-Answering (Improved Precision), Querying, SW Services Real World Domain Modeling, Semantic Search (using concepts, properties, relations, rules), Machine Interpretability (M2M, M2H semantic interoperability), Automated Reasoning, SW Services Ontology weak strong Concept (referent) - based Term - based More Expressive Semantic Models Enable More Complex Applications