University of Crete HY 566 Semantic Web CS

University of Crete HY 566 -Semantic Web CS 566 – Semantic Web Knowledge Management & Semantic Web Παπαγγελής Μάνος, Κοφφινά Ιωάννα, Κοκκινίδης Γιώργος Computer Science Department - Uo. C Heraklion 5 June, 2003 Spring‘ 03 Knowledge Management & Semantic Web

University of Crete HY 566 -Semantic Web Overview § Introduction to Knowledge Management § Knowledge Management Weaknesses § Knowledge Management for Semantic Web • Ontology-based KM systems • A Framework for KM on the Semantic Web § Knowledge Representation § Knowledge Management System Example § Conclusion Remarks Spring‘ 03 Knowledge Management & Semantic Web 2

University of Crete HY 566 -Semantic Web Contents § Introduction to Knowledge Management § Knowledge Management Weaknesses § Knowledge Management for Semantic Web • Ontology-based KM systems • A Framework for KM on the Semantic Web § Knowledge Representation § Knowledge Management System Example § Conclusion Remarks Spring‘ 03 Knowledge Management & Semantic Web 3

University of Crete HY 566 -Semantic Web What is Knowledge Management (KM) § There is no universal definition of KM § KM could be defined as the process through which organizations generate value from their intellectual and knowledge-based assets § KM is often facilitated by IT § Not all information is valuable § Two categories of knowledge • Explicit - Anything that can be documented, archived and codified, often with the help of IT • Tacit - The know-how contained in people's heads Spring‘ 03 Knowledge Management & Semantic Web 4

University of Crete HY 566 -Semantic Web Technologies that support current KM Systems § § § Knowledge repositories Expertise access tools E-learning applications Discussion and chat technologies Synchronous interaction tools Search and data mining tools. Spring‘ 03 Knowledge Management & Semantic Web 5

University of Crete HY 566 -Semantic Web KM System Weaknesses § Searching Information • Word keywords don’t express the semantics § Extracting Information • Agents are not able to extract knowledge from textual representations and to integrate information spread over different sources § Maintaining • Sustaining weakly structured text sources is difficult and time-consuming • Such collections cannot be easily consistent, correct and up -to-date § Automating Document Generation • Adaptive Websites that enable dynamic reconfiguration based on user profiles require machine–accessible representation of the semi-structured data Spring‘ 03 Knowledge Management & Semantic Web 6

University of Crete HY 566 -Semantic Web Contents § Introduction to Knowledge Management § Knowledge Management Weaknesses § Knowledge Management for Semantic Web • Ontology-based KM systems • A Framework for KM on the Semantic Web § Knowledge Representation § Knowledge Management System Example § Conclusion Remarks Spring‘ 03 Knowledge Management & Semantic Web 7

University of Crete HY 566 -Semantic Web Ontology-based KM systems § Methodology for developing ontology-based KM systems § Ontologies can help formalize the knowledge shared by a group of people, in contexts where knowledge has to be modeled, structured and interlinked § Distinction between knowledge process and knowledge metaprocess § Two orthogonal Processes with Feedback Loops § Knowledge Process § Knowledge Meta-process Spring‘ 03 Knowledge Management & Semantic Web 8

University of Crete HY 566 -Semantic Web The Knowledge Process (1/4) Knowledge Creation Knowledge Import Knowledge Capture Knowledge Retrieval and Access § Knowledge Use § § Spring‘ 03 Knowledge Management & Semantic Web 9

University of Crete HY 566 -Semantic Web The Knowledge Process (2/4) § Knowledge Creation • Computer-accessible knowledge moves between formal and informal • In order to have knowledge in the middle of the two extremes the idea is to embed the structure of knowledge items into document templates Spring‘ 03 Knowledge Management & Semantic Web 10

University of Crete HY 566 -Semantic Web The Knowledge Process (3/4) § Knowledge Import • Importing knowledge into KM system has the same or more importance than creating it • For imported knowledge, accurate access to relevant items plays an even more important role than for homemade knowledge § Knowledge Capture • Knowledge capturing refers to the way that knowledge items, their essential contents and their interlinks are accessed (Onto. Annotate) Spring‘ 03 Knowledge Management & Semantic Web 11

University of Crete HY 566 -Semantic Web The Knowledge Process (4/4) § Knowledge Retrieval and Access • Typically through a conventional GUI • Ontology can be used to derive further views of the knowledge (e. g. Navigation) and additional links and descriptions § Knowledge Use • It is not the knowledge itself that is of most interest, but the derivations made from it • No single knowledge item can be useful, but the overall picture derived the total analysis Spring‘ 03 Knowledge Management & Semantic Web 12

University of Crete HY 566 -Semantic Web The Knowledge Meta-Process (1/3) § § § Spring‘ 03 Knowledge Management & Semantic Web Feasibility Study Kickoff phase Refinement Phase Evaluation Phase Maintenance Phase 13

University of Crete HY 566 -Semantic Web The Knowledge Meta-Process (2/3) § Feasibility Study • Identification of problems and opportunity areas • Selection of the most promising focus area and target solution § Kick off phase • Requirement specification • Analysis of input sources • Development of baseline taxonomy Spring‘ 03 Knowledge Management & Semantic Web 14

University of Crete HY 566 -Semantic Web The Knowledge Meta-Process (3/3) § Refinement phase • Concept Elicitation with domain experts • Development of baseline taxonomy • Conceptualization and Formalization § Evaluation Phase • Revision and Expansion based on feedback • Analysis of usage patterns • Analysis of competency questions § Maintenance Phase • Management of organizational maintenance process Spring‘ 03 Knowledge Management & Semantic Web 15

University of Crete HY 566 -Semantic Web Contents § Introduction to Knowledge Management § Knowledge Management Weaknesses § Knowledge Management for Semantic Web • Ontology-based KM systems • A Framework for KM on the Semantic Web § Knowledge Representation § Knowledge Management System Example § Conclusion Remarks Spring‘ 03 Knowledge Management & Semantic Web 16

University of Crete HY 566 -Semantic Web A Framework for KM on the SW 1. 2. 3. 4. 5. Knowledge Capturing Knowledge Repository Knowledge Processing Knowledge Sharing Using of Knowledge Spring‘ 03 Knowledge Management & Semantic Web 17

University of Crete HY 566 -Semantic Web Knowledge Capturing § Knowledge can be collected from various sources and in different formats § Four Types of Knowledge Sources • • Expert knowledge Legacy Systems Metadata Repositories Documents § Need for Knowledge Capturing Tools Spring‘ 03 Knowledge Management & Semantic Web 18

University of Crete HY 566 -Semantic Web Knowledge Repository § Use of Relational Databases • Efficient storing • Efficient Access to RDF metadata § It is an RDF Repository like RDFSuite or RDF Gateway Spring‘ 03 Knowledge Management & Semantic Web 19

University of Crete HY 566 -Semantic Web Knowledge Process § Efficient manipulation of the stored knowledge § Graph-based processing for knowledge represented in the form of rules • E. g Deriving a dependency graph Spring‘ 03 Knowledge Management & Semantic Web 20

University of Crete HY 566 -Semantic Web Knowledge Sharing § Knowledge Integration of different sources (Knowledge Base) and its utilization § Realized by searching for rules that satisfy the query conditions § Searching is realized as an inferencing process • Ground assertions (RDF triples) and domain axioms are used for deriving new assertions Spring‘ 03 Knowledge Management & Semantic Web 21

University of Crete HY 566 -Semantic Web Using of Knowledge § Finding appropriate documents is essential, but the derivation made of them adds value to KM applications § Composition of documents • Use of conditional statements § Conditional Statements leads to efficient searching for knowledge • Precondition-Action Spring‘ 03 Knowledge Management & Semantic Web 22

University of Crete HY 566 -Semantic Web Proposed KM Framework Spring‘ 03 Knowledge Management & Semantic Web 23

University of Crete HY 566 -Semantic Web Contents § Introduction to Knowledge Management § Knowledge Management Weaknesses § Knowledge Management for Semantic Web • Ontology-based KM systems • A Framework for KM on the Semantic Web § Knowledge Representation § Knowledge Management System Example § Conclusion Remarks Spring‘ 03 Knowledge Management & Semantic Web 24

University of Crete HY 566 -Semantic Web Knowledge Representation § Knowledge should be expressed by explicit semantics in order to be understood by automated tools § Select schemas and express knowledge through them § Knowledge sharing, merging and retrieval are possible if the categories used in the knowledge representation are connected by semantic links, expressed in ontologies Spring‘ 03 Knowledge Management & Semantic Web 25

University of Crete HY 566 -Semantic Web Elements of Knowledge Representation § Ontologies and Knowledge Bases • Ontologies are catalogues of categories with their associated complete or partial formal definitions of necessary and sufficient conditions • A knowledge base is composed of one ontology (or several interconnected ontologies) plus additional statements using these ontologies § Ontology Servers • Permit Web users to modify the ontology part of the KB § Knowledge within Web Documents • Permit the insertion of knowledge inside HTML documents Spring‘ 03 Knowledge Management & Semantic Web 26

University of Crete HY 566 -Semantic Web Challenges of Semantic Web § Scale of information • The information found on the Web is orders of magnitude larger than any traditional single knowledge-base § Change rate • Information is updated frequently § Lack of referential integrity • Links may be broken and information may not be found § Distributed authority • Trust of knowledge is not standard because data are obtained through different users § Variable quality of knowledge • Knowledge may differ in quality and should not be treated the same Spring‘ 03 Knowledge Management & Semantic Web 27

University of Crete HY 566 -Semantic Web Challenges of Semantic Web (cont. ) § Unpredictable use of knowledge • Knowledge base should be task-independent § Multiple knowledge sources • Knowledge is not provided by a single source § Diversity of content • The focus of interest is wider § Linking, not copying • The size of information forbid the copy of data § Robust inferencing • The degrees of incompleteness and unsoundness must be functions of the available resources • Answers could be approximate Spring‘ 03 Knowledge Management & Semantic Web 28

University of Crete HY 566 -Semantic Web Ontology § Processing and sharing of knowledge between programs in the Web § Definitions • Representation of a shared conceptualization of a particular domain • A consensual and formal specification of a vocabulary used to describe a specific domain • A set of axioms designed to account for the intended meaning of a vocabulary § An ontology provides • A vocabulary for representing and communicating knowledge about some topic • A set of relationships that hold among the terms in that vocabulary Spring‘ 03 Knowledge Management & Semantic Web 29

University of Crete HY 566 -Semantic Web Ontology Driven KR § Knowledge sharing and reuse § Enable machine-based communication § Reusable descriptions between different services § No more keyword-based approach… § …but syntactic- and semantic-based discovery of knowledge § Hierarchical description of important concepts and definition of their properties (attributevalue mechanism) Spring‘ 03 Knowledge Management & Semantic Web 30

University of Crete HY 566 -Semantic Web Languages for KR 1. XML 2. RDF / RDF Schema 3. DAML+OIL 4. OWL Spring‘ 03 Knowledge Management & Semantic Web 31

University of Crete HY 566 -Semantic Web Contents § Introduction to Knowledge Management § Knowledge Management Weaknesses § Knowledge Management for Semantic Web • Ontology-based KM systems • A Framework for KM on the Semantic Web § Knowledge Representation § Knowledge Management System Example § Conclusion Remarks Spring‘ 03 Knowledge Management & Semantic Web 32

University of Crete HY 566 -Semantic Web On-To-Knowledge § On-To-Knowledge was a European project that built an ontology-based tool environment to speed up knowledge management § Results aimed were • Toolset for semantic information processing and user access • OIL, an ontology-based inference layer on top of the Web • Associated Methodology • Validation by industrial case studies Spring‘ 03 Knowledge Management & Semantic Web 33

University of Crete HY 566 -Semantic Web On-To-Knowledge Architecture Spring‘ 03 Knowledge Management & Semantic Web 34

University of Crete HY 566 -Semantic Web On-To-Knowledge Technical Architecture Spring‘ 03 Knowledge Management & Semantic Web 35

University of Crete HY 566 -Semantic Web Tools Used § RDFferret • Combines full text searching with RDF quering § Onto. Share • Storage of the information in an ontology and querying, browsing and searching that ontology § Spectacle • Organizes the presentation (ontology-driven) of information and offers an exploration context § Onto. Edit • Inspect, browse, codify and modify ontologies Spring‘ 03 Knowledge Management & Semantic Web 36

University of Crete HY 566 -Semantic Web Tools Used (cont. ) § Ontology Middleware Module (OMM) • Deals with ontology versioning, security (user profiles and groups), meta-information and ontology lookup and access via a number of protocols (Http, RMI, EJB, CORBA and SOAP) § LINRO • Offers reasoning tasks for description logics, including realization and retrieval § Sesame • Persistent storage of RDF data and schema information and online querying of that information Spring‘ 03 Knowledge Management & Semantic Web 37

University of Crete HY 566 -Semantic Web Tools Used (cont. ) § CORPORUM toolset • Onto. Extract and Onto. Wrapper • Information Extraction and ontology generation • Interpretation of natural language texts is done automatically • Extraction of specific information from free text based on business rules defined by the user • Extracted information is represented in RDF(S)/DAML+OIL and is submitted to the Sesame Data Repository Spring‘ 03 Knowledge Management & Semantic Web 38

University of Crete HY 566 -Semantic Web Contents § Introduction to Knowledge Management § Knowledge Management Weaknesses § Knowledge Management for Semantic Web • Ontology-based KM systems • A Framework for KM on the Semantic Web § Knowledge Representation § Knowledge Management System Example § Conclusion Remarks Spring‘ 03 Knowledge Management & Semantic Web 39

University of Crete HY 566 -Semantic Web Conclusion Remarks § Current Knowledge Management technologies need to be revised § There are some architectures of Knowledge Management Systems for Semantic Web but there are only few KM applications available § Knowledge Representation has to meet the challenges that Semantic Web poses § On-to-knowledge proposes a fine architecture on which KM systems for SW can be based Spring‘ 03 Knowledge Management & Semantic Web 40