Скачать презентацию An Environment for Merging and Testing Large Ontologies Скачать презентацию An Environment for Merging and Testing Large Ontologies

1e1f3a97b96a4181a3bd1aded0ac4458.ppt

  • Количество слайдов: 29

An Environment for Merging and Testing Large Ontologies Deborah Mc. Guinness, Richard Fikes, James An Environment for Merging and Testing Large Ontologies Deborah Mc. Guinness, Richard Fikes, James Rice*, Steve Wilder Associate Director and Senior Research Scientist Knowledge Systems Laboratory Stanford University Stanford, CA 94305 650 -723 -9770 [email protected] stanford. edu *Commerce. One, Mountain View, CA

Motivation: Ontology Integration Trends u Integrated in most search applications (Yahoo, Lycos, Xift, …) Motivation: Ontology Integration Trends u Integrated in most search applications (Yahoo, Lycos, Xift, …) u Core component of E-Commerce applications (Amazon, e. Bay, Virtual Vineyards, REI, Vertical. Net, Commerce. One, etc. ) u Integrated in configuration applications (Dell, PROSE, etc. )

Motivation: Ontology Evolution u Controlled vocabularies abound (SIC-codes, UN/SPSC, Rosetta. Net, Open. Directory, …) Motivation: Ontology Evolution u Controlled vocabularies abound (SIC-codes, UN/SPSC, Rosetta. Net, Open. Directory, …) u Distributed ownership/maintenance u Larger scale (Open Directory >23. 5 K editors, ~250 K categories, 1. 65 M sites) u Becoming more complicated - Moving to classes and slots (and value restrictions, enumerated sets, cardinality)

Chimaera – A Merging and Diagnostic Ontology Environment Web-based tool utilizing the KSL Ontolingua Chimaera – A Merging and Diagnostic Ontology Environment Web-based tool utilizing the KSL Ontolingua platform that supports: u merging multiple ontologies found in distributed environments u analysis of single or multiple ontologies u attention focus in problematic areas u simple browsing and mixed initiative editing

The Need For KB Merging u Large-scale knowledge repositories will contain KBs produced by The Need For KB Merging u Large-scale knowledge repositories will contain KBs produced by multiple authors in multiple settings u KBs for applications will be built by assembling and extending multiple modular KBs from repositories u KBs u u developed by multiple authors will frequently Express overlapping knowledge in a common domain Use differing representations and vocabularies u For such KBs to be used together as building blocks - Their representational differences must be reconciled

The KB Merging Task u Combine KBs that: u Were developed independently (by multiple The KB Merging Task u Combine KBs that: u Were developed independently (by multiple authors) u Express overlapping knowledge in a common domain u Use differing representations and vocabularies u Produce merged KB with u Non-redundant u Coherent u Unified vocabulary, content, and representation

How KB Merging Tools Can Help u Combine u input KBs with name clashes How KB Merging Tools Can Help u Combine u input KBs with name clashes Treat each input KB as a separate name space u Support merging of classes and relations Replace all occurrences by the merged class or relation u Test for logical consistency of merge (e. g. instances/subclasses of multiple disjoint classes) u Actively look for inconsistent extensions u u Match u Find name clashes, subsumed names, synonyms, . . . u Focus u vocabulary attention Portions of KB where new relationships are likely to be needed E. g. , sibling subclasses from multiple input KBs u Derive u relationships among classes and relations Disjointness, equivalence, subsumption, inconsistency, . . .

Merging Tools u Merging can be arbitrarily difficult u KBs can differ in basic Merging Tools u Merging can be arbitrarily difficult u KBs can differ in basic representational design u May require extensive negotiation among authors u Tools u KB can significantly accelerate major steps merging using conventional editing tools is Labor intensive u Difficult Error prone u Hypothesis: tools specifically designed to support KB merging can significantly u Speed up the merging process u Make broader user set productive u Improve the quality of the resulting KB

Our KB Analysis Task u Review KBs that: u Were developed using differing standards Our KB Analysis Task u Review KBs that: u Were developed using differing standards u May be syntactically but not semantically validated u May use differing modeling representations u May have different purposes u Produce KB logs (in interactive environments) u Identify provable problems u Suggest possible problems in style and/or modeling u Are extensible by being user programmable

Chimaera Usage u HPKB program – analyze diverse KBs, support KR novices as well Chimaera Usage u HPKB program – analyze diverse KBs, support KR novices as well as experts u Cleaning semi-automatically generated KBs u Browsing and merging multiple controlled vocabularies (e. g. , internal vocabularies and UN/SPSC (std products and services codes)) u Reviewing internal vocabularies

Discussion/Conclusion • Ontologies are becoming more central to applications, they are larger, more distributed, Discussion/Conclusion • Ontologies are becoming more central to applications, they are larger, more distributed, and longer-lived • Environmental support (in particular merging and diagnostic support) is more critical for the broader user base • Chimaera provides merging and diagnostic support for ontologies in many formats • It improves performance over existing tools • It has been used by people of various training backgrounds in government and commercial applications and is available for use. • http: //www. ksl. Stanford. EDU/software/chimaera/ movie, tutorial, papers, link to live system, etc.

Extras Extras

The Need For KB Analysis u Large-scale knowledge repositories will contain KBs produced by The Need For KB Analysis u Large-scale knowledge repositories will contain KBs produced by multiple authors in multiple settings u KBs for applications will be built by assembling and extending multiple modular KBs from repositories that may not be consistent u KBs developed by multiple authors will frequently u Express overlapping knowledge in different, possibly contradictory ways u Use differing assumptions and styles u Have different purposes must be reviewed for appropriateness and “correctness” u KBs

What is an Ontology? Catalog/ ID Thesauri “narrower term” relation Terms/ glossary Informal is-a What is an Ontology? Catalog/ ID Thesauri “narrower term” relation Terms/ glossary Informal is-a Frames General Formal is-a (properties) Logical constraints Formal instance Disjointness, Value Inverse, part. Restrs. of…

Ontologies and importance to E-Commerce Simple ontologies provide: u. Controlled shared vocabulary (search engines, Ontologies and importance to E-Commerce Simple ontologies provide: u. Controlled shared vocabulary (search engines, authors, users, databases, programs all speak same language) u. Organization (and navigation support) u. Expectation setting (left side of many web pages) u. Browsing support (tagged structures such as Yahoo!) u. Search support (query expansion approaches such as Find. UR, e-Cyc) u. Sense disambiguation

Ontologies and importance to E-Commerce II u. Foundation for expansion and leverage u. Conflict Ontologies and importance to E-Commerce II u. Foundation for expansion and leverage u. Conflict detection u. Completion u. Regression testing/validation/verification support foundation u. Configuration support u. Structured, comparative search u. Generalization/ Specialization u…

E-Commerce Search (starting point Forrester modified by Mc. Guinness) u Ask Queries - multiple E-Commerce Search (starting point Forrester modified by Mc. Guinness) u Ask Queries - multiple search interfaces (surgical shoppers, advice seekers, window shoppers) - set user expectations (interactive query refinement) - anticipate anomalies u Get Answers - basic information (multiple sorts, filtering, structuring) - modify results (user defined parameters for refining, user profile info, narrow query, broaden query, disambiguate query) - suggest alternatives (suggest other comparable products even from competitor’s sites) u Make Decisions - manipulate results (enable side by side comparison) - dive deeper (provide additional info, multimedia, other views) - take action (buy)

A Few Observations about Ontologies u u u Simple ontologies can be built by A Few Observations about Ontologies u u u Simple ontologies can be built by non-experts u Consider Verity’s Topic Editor, Collaborative Topic Builder, GFP interface, Chimaera, etc. Ontologies can be semi-automatically generated u from crawls of site such as yahoo!, amazon, excite, etc. u Semi-structured sites can provide starting points Ontologies are exploding (business pull instead of technology push) u most e-commerce sites are using them - My. Simon, Affinia, Amazon, Yahoo! Shopping, , etc. u Controlled vocabularies (for the web) abound - SIC codes, UMLS, UN/SPSC Open Directory, Rosetta Net, … u Business ontologies are including roles u DTDs are making more ontology information available u Businesses have ontology directors u “Real” ontologies are becoming more central to applications

Implications and Needs u Ontology Language Syntax and Semantics u Environments for Creation and Implications and Needs u Ontology Language Syntax and Semantics u Environments for Creation and Maintenance of Ontologies u Training (Conceptual Modeling, reasoning implications, …) u Issues: u Collaboration among distributed teams u Diverse training levels u Interconnectivity with many systems/standards u Analysis and Diagnosis u Scale