CPECSC 580 Knowledge Management Dr Franz J Kurfess

CPE/CSC 580: Knowledge Management Dr. Franz J. Kurfess Computer Science Department Cal Poly © 2001 -2005 Franz J. Kurfess Knowledge Exchange 1

Course Overview u u Introduction Knowledge Processing u u Knowledge Organization u u u Classification, Categorization Ontologies, Taxonomies, Thesauri Knowledge Retrieval u u u Knowledge Acquisition, Representation and Manipulation Information Retrieval Knowledge Navigation Knowledge Presentation u Knowledge Visualization © 2001 -2005 Franz J. Kurfess u Knowledge Capture, Transfer, and Distribution u Usage u of Knowledge Access Patterns, User Feedback u Knowledge Techniques u Exchange Management Topic Maps, Agents u Knowledge Management Tools u Knowledge Management in Organizations Knowledge Exchange 2

Overview Knowledge Exchange u Transfer u Motivation u u Objectives u Knowledge u u Explicit Capture Extraction From Text Case-based Reasoning Enhancement of Existing Documents Communication u Capture of Knowledge u Basic Concepts Language and Communication u u u Natural Language Formal Languages Communication Models u Distribution u u of Knowledge Repositories Distribution Models u Important Concepts and Terms u Chapter Summary © 2001 -2005 Franz J. Kurfess Knowledge Exchange 3

Logistics u Term Project u Documentation u Schedule u Lab and Homework Assignments u Schedule u Exams u Grading © 2001 -2005 Franz J. Kurfess Knowledge Exchange 4

The Need for Knowledge Exchange WWW Introductory texts, expert hints, explanations, dialogues, comments, examples, exceptions, . . . Richer representations More ambiguous More versatile Info. extraction templates, dialogue segments and pegs, filled-out forms, high-level connections, . . . Descriptions augmented with prototypical examples & exceptions, problem-solving steps and substeps, . . . ((( )) ()))) © 2001 -2005 Franz J. Kurfess (defconcept bridge ())) Alternative formalizations (KIF, MELD, CML, …), alternative views of the same notion (e. g. , what is a threat) [Gil 2000] More formal More concrete More introspectible Knowledge Exchange 5

Knowledge Mobility u multiple views and versions of the same information u need to provide tools that establish connections among alternative versions/views of the same information u hyper-connectivity u need to provide tools that suggest further connections to related sources when users compose documents u need to annotate hyperlinks u basis to support information morphing u how one or more knowledge sources are used for alternative purposes v track alternative knowledge transformations v various renderings and implementations of a knowledge source v © 2001 -2005 Franz J. Kurfess [Gil 2000] Knowledge Exchange 6

Pre-Test © 2001 -2005 Franz J. Kurfess Knowledge Exchange 7

Motivation © 2001 -2005 Franz J. Kurfess Knowledge Exchange 8

Objectives © 2001 -2005 Franz J. Kurfess Knowledge Exchange 9

Knowledge Capture u Explicit Capture u Extraction From Text u Case-based Reasoning u Enhancement of Existing Documents © 2001 -2005 Franz J. Kurfess Knowledge Exchange 10

Explicit Capture u conventional techniques for knowledge acquisition u interviews with experts u knowledge engineers u advantages u carefully constructed u suitable knowledge representation methods u usually common-sense evaluation u sometimes formal evaluation u consistency checks, other formal aspects © 2001 -2005 Franz J. Kurfess Knowledge Exchange 11

Extraction From Text u syntactic u keywords u u essentially construction of an index descriptive features u semantic u level document structure u requires information about structure u u level tags, DDT, RDF sentence structure u natural language processing (NLP) u pragmatic u level context u thesaurus, ontology, NLP © 2001 -2005 Franz J. Kurfess Knowledge Exchange 12

Case-based Reasoning u solutions to a problem in a specific context are collected u represented in a structured format u problem, context, solution u usable by a computer-based system u cases are often represented through frames or similar mechanisms u new cases are matched against existing ones u patterns in the frames provide the basis for matching u the suitability of the solution is judged by the user © 2001 -2005 Franz J. Kurfess Knowledge Exchange 13

Enhancement of Existing Documents u in addition to the methods mentioned above, collections of documents can be enhanced u addition of meta-knowledge u integration into an existing framework/ontology manually through categorization u automatically through keyword extraction u indirectly through statistical correlations with other documents u © 2001 -2005 Franz J. Kurfess Knowledge Exchange 14

Transfer of Knowledge u Communication u Basic Concepts u Language and Communication Natural Language u Formal Languages u u Communication © 2001 -2005 Franz J. Kurfess Models Knowledge Exchange 15

Basic Concepts u communication u exchange of information u requires a shared system of signs u greatly enhanced by language u speaker produces signs as utterances u general: not only spoken language u u listener u (hearer) perceives and interprets signs © 2001 -2005 Franz J. Kurfess Knowledge Exchange 16

Purpose of Communication u sharing of information among agents or systems u query other agents for information u responses to queries u requests or commands v actions to be performed for another agent u offer v proposition for collaboration u acknowledgement v confirmation of requests, offers u sharing v of experiences, feelings © 2001 -2005 Franz J. Kurfess Knowledge Exchange 17

Communication Problems u intention u what is the expected outcome (speaker’s perspective) u timing u when is a communication act appropriate u selection u which act is the right one u language u what sign system should be used u interpretation u will the intended meaning be conveyed to the listener u ambiguity u can the intention be expressed without the possibility of misunderstandings © 2001 -2005 Franz J. Kurfess Knowledge Exchange 18

Language and Communication u Natural Language u used by humans u evolves over time u moderately to highly ambiguous u Formal Languages u invented u rigidly defined u little ambiguity © 2001 -2005 Franz J. Kurfess Knowledge Exchange 19

Natural Language u formal description is very difficult u sometimes u mostly non-systematic, inconsistent, ambiguous used for human communication u easy on humans u tough on computers u context is critical u situation, beliefs, goals © 2001 -2005 Franz J. Kurfess Knowledge Exchange 20

Formal Languages u symbols u terminal symbols finite set of basic words u not: alphabet, characters u u non-terminal u symbols intermediate structures composed of terminal or non-terminal symbols u strings u sequences of symbols u phrases u sub-strings © 2001 -2005 Franz J. Kurfess grouping important parts of a string Knowledge Exchange 21

Formal Languages Cont. u sentences u allowable strings in a language u composed from phrases u grammar u rules describing correct sentences u often captured as rewrite rules in BNF notation u lexicon u list of allowable vocabulary words © 2001 -2005 Franz J. Kurfess Knowledge Exchange 22

Communication Models u encoded message model ua definite proposition of the speaker is encoded into signs which are transmitted to the listener u the listener tries to decode the signs to retrieve the original proposition u errors are consequences of transmission problems u situated language model u the intended meaning of a message depends on the signals as well as the situation in which they are exchanged u mis-interpretation may lead to additional problems © 2001 -2005 Franz J. Kurfess Knowledge Exchange 23

Communication Types u telepathic communication u speaker and listener have a shared internal representation u communication through Tell/Ask directives u language-based communication u speaker performs actions that produce signs which other agents can perceive and interpret u communication language is different from the internal representation u more complex involves several mappings v language needs to be generated, encoded, transmitted, decoded, and interpreted v © 2001 -2005 Franz J. Kurfess Knowledge Exchange 24

Telepathic Communication © 2001 -2005 Franz J. Kurfess [Russell & Norvig 1995] Knowledge Exchange 25

Language-Based Communication © 2001 -2005 Franz J. Kurfess [Russell & Norvig 1995] Knowledge Exchange 26

Communication Steps: Speaker u intention u decision about producing a speech act u generation u conversion of the information to be transferred into the chosen language u synthesis u actions that produce the generated signs © 2001 -2005 Franz J. Kurfess Knowledge Exchange 27

Communication Steps: Listener u perception u reception v of the signs produced by the speaker speech recognition, lip reading, character recognition u analysis syntactic interpretation (parsing) v semantic interpretation v u disambiguation v selection of the most probable intended meaning u incorporation v the selected interpretation is added to the existing world model as additional piece of evidence © 2001 -2005 Franz J. Kurfess Knowledge Exchange 28

Communication Example © 2001 -2005 Franz J. Kurfess [Russell & Norvig 1995] Knowledge Exchange 29

Distribution of Knowledge u Digital Repositories Libraries u Distribution Models © 2001 -2005 Franz J. Kurfess Knowledge Exchange 30

Knowledge Repositories u persistent u internal v storage of digital documents representation in the original format loss-less transformation may be acceptable u transparent internal organization u multiple presentation methods for various users and usage methods u multiple access methods u according to users’ needs and capabilities © 2001 -2005 Franz J. Kurfess Knowledge Exchange 31

Digital Libraries u collections of documents and artifacts stored and accessed via computers u remotely accessible through networks u enhanced functionality compared with paper-based libraries u access methods u organization principles u duplication u implementation © 2001 -2005 Franz J. Kurfess and usage unclear Knowledge Exchange 32

Vision. . . “… a network of learning environments and resources for Science, Mathematics, Engineering and Technology education, will ultimately meet the needs of students and teachers at all levels—K-12, undergraduate, and lifelong learning—in both individual and collaborative settings. ” NSF © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 33

NEEDS u National Berkeley Stanford Northern Arizona UMBC Virginia Tech digital library emerging from the Synthesis Coalition (1990 - 1999) u Integrated database of multimedia courseware u Multilevel courseware evaluation system u Community © 2001 -2005 Franz J. Kurfess of Engineering Educators [NEEDS 2000] Knowledge Exchange 34

www. needs. org - Contents © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 35

www. needs. org - Communities u Community - a group of individuals who engage in communication through a common bond. u Faculty prefer to learn where they are a member of a community that is built upon shared values and interests regarding teaching, learning, and pedagogy u GOAL: Provide faculty a means to learn from one another unconstrained by barriers of time, distance, technology, and geography. © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 36

Towards A National SMETE Digital Library. . . Should NSF Establish and Fund a National Science, Mathematics, Engineering and Technology Education Digital Library? u April 1996 - NSF Committee Meeting (LIBUSE) • u August 1997 National Research Council • u Digital National Library for SME&T Education Workshop July 1998 National Science Foundation u u “Towards a National Library for Undergraduate Science Education Resources in Science, Mathematics, Engineering and Technology” SMETE-Lib Workshop January 1999 National Science Foundation u Digital Libraries and Education Workshop © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Funding through Digital Libraries Initiative – Phase 2 And NSDL Programs Knowledge Exchange 37

www. needs. org Delivery n Classrooms n Instructional Labs n Small Study Groups n Residences n Libraries n Anywhere Database Development n Courseware Studios n Instructional Labs n Faculty Offices & Residences n Libraries n Anywhere NEEDS is the foundation for the National SMETE Digital Library at www. smete. org. © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 38

NEEDS Aspects u Digital u u Library of Multimedia Engineering Courseware Bibliographic records with downloadable courseware Multimedia elements - downloadable movies, images, and text u Multilevel u u u Courseware Evaluation System Peer Review of Courseware Premier Award for Excellence in Engineering Education Courseware Expanding Services and Features © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 39

How does NEEDS help users “re-use” learning materials? u Provides mechanisms to help user locate materials u Uses standardized descriptions (metadata) to describe resources u Provides mechanisms to help users evaluate the “quality of materials” u Developed u upon an extendable platform to: Support multiple uses Integrate new services and features Integrate research © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 40

Quality Review of Courseware on the NEEDS Database u Establish credibility of NEEDS as a source of Quality educational material u Enhance recognition of scholarly and creative effort of courseware developers u u Peer/Expert Review of Courseware Premier Award for Excellence in Engineering Education Courseware © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 41

The Premier Award for Excellence in Engineering Education Courseware u. A national competition to identify and reward the authors of high-quality, non-commercial courseware designed to enhance engineering education. u u. A The Premier Award is about the entire experience of using the courseware by learners, not just the courseware itself dissemination system to distribute the Premier Courseware (via CD’s and presentation at engineering education conferences).

Premier Courseware of 1997 -1999 u u u Virtual Disk Drive Design Studio Drill Dissection and Bicycle Dissection Mars Navigator u u u Della Steam Plant u Engineering Graphics MDSolids u Cracking Dams Structural Engineering Visual Encyclopedia - UNH 6, 700 CD-ROMs Distributed For more info or to receive copies go to http: //www. needs. org/engineering/premier/ © 2001 -2005 Franz J. Kurfess Knowledge Exchange 43

Needs Assessment with Members of the Math, Science and Engineering Community Purpose: To understand the math, science and engineering communities of educators and examine their needs in order to design services and structures to support users from multiple communities. Research Questions: u What services, features & programs are integral to success? u What do users expect with regards to quality of the holdings? u Who makes up the SMETE digital library community? American Association of Physics Teachers, American Mathematical Society, American Association for the Advancement of Science, members of the NSF Chemistry Consortia and the NSF Engineering Education Coalitions © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 44

Translating Findings into Services & Features Quality u System to rapidly identify the quality of holding u Place to comment about a learning object or regarding something of interest to the community u Reviewers should include experts in pedagogy and content © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 45

Translating Findings into Services & Features Community u Embedded structures for developing and maintaining communication links u Developing community should be on par with building content u Build on discipline based communities to establish connection to a broader community Content u Useful content and community interaction ensures user participation as authors, reviewers, adapters/adopters, and consumers © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 46

Prototype: www. smete. org Berkeley UCOP Stanford Northern Arizona John Wiley Math Forum ENC UMBC Virginia Tech Prototype Goals (1998 -2001) Develop a Prototype National SMETE Digital Library u test interoperability of federated searches/shared service with partners u expand requirements analysis to include K– 12 u develop criteria and standards to assess the impact of learning objects across disciplines u implement community feedback systems, evaluate [NEEDS 2000]

Building a National SMETE Digital Library at www. smete. org u Searching for learning resources u Cataloging (adding) learning resources u Standards, u Evaluating IEEE and IMS the quality of learning resources u “User” reviews u “Expert” reviews u Forming a community of users in SMETE u PKAL workshops and seminars u Research on adapters © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 48

Systems Development u Expanding www. smete. org/NEEDS platform u Continuing to participate in the development of IEEE/IMS Learning Object Metadata Standards u Adopting emerging IEEE standards u Expanding user comments u Implementing discussion systems u Implementing customized Total Collection user profiles Engineering 58% Chemistry 21% u Expanding Collections Physics 14% u Expanding into Chemistry, Math 5% Physics, and Mathematics Other 2% © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 49

Collaborating with Partners u Working with University of California Office of the President - (10 campuses) u Identifying courseware under development systemwide u Developing TLT@UC Website to showcase teaching and learning with technology at the University of California © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 50

Challenges Toward the Future. . . u Continuing to understand support changing user needs u Improving ability to encapsulate the instructional intent and use of materials u Metadata u Supporting standards and cataloging practice communities of use and practice u pedagogy u content © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 51

Challenges Toward the Future. . . u Continuing to understand support changing user needs u Improving ability to encapsulate the instructional intent and use of materials u Metadata u Supporting standards and cataloging practice communities of use and practice u pedagogy u Content u Which allows for a user to find what they’re really looking for as well as personalization of content. © 2001 -2005 Franz J. Kurfess [NEEDS 2000] Knowledge Exchange 52

www. smete. or g NACME Web. CT Bio. QUEST John Wiley Math Forum Berkeley ENC UMBC AAAS UCOP ILT Utah State PKAL Stanford DLESE Virginia Tech Cisco SRI Northern GDL Eduprise Arizona MAA Sun TI NSDL Goals (2000 - ) u Develop u u the National SMETE Digital Library Provide seamless access to services and resources Create a dynamic learning community that promotes and supports SMET education in the 21 st century u Expand Partnerships u Expand Services and Community [NEEDS 2000]

UNC Digital Library Project Digital Services Grants (In-Kind) REQUEST FOR PROPOSALS (RFP) Spring 2001 Agenda Overview Q&A for RFP Digital Library Background © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 54

Overview u Current and Future Project Services • Current Structure of the DLP • Prototype Interface for DLP © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 55

Services for Curators u u u Improved access to collections Support for a variety of media types 24 x 7 storage services Standards-based metadata schema Capability to support discipline-specific vocabularies Mini-grant program for digitizing high-impact collections © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 56

Services for Instructors u u u Flexible rights management (limit access to students in a course, departments, etc. ) Improved access to other supporting materials Presentation options u u u Web-based slide show Side-by-side object comparison Integration with desktop applications Support for annotations and other associated data Documentation, training © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 57

Services for Scholars u u Common web interface to many digital collections Customizable “My Digital Library” user interface Multiple search options (keyword, browsing, etc. ) Compliance with guidelines for disability access © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 58

Structure of the DLP INTERFACE DATABASE MIDDLEWARE OBJECT STORAGE © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 59 OTHER PROJECTS

Structure of the DLP INTERFACE DATABASE MIDDLEWARE • Metadata Directory DATABASE • User Management OBJECT STORAGE • Schema Repository © 2001 -2005 Franz J. Kurfess • Crosswalk Repository • Controlled Vocabularies [UNC Digital Library 2000] Knowledge Exchange 60 OTHER PROJECTS

Structure of the DLP • Metadata Transformer • Authentication and Authorization DATABASE MIDDLEWARE • Discovery Tools • Object Manipulation INTERFACE MIDDLEWARE • Presentation Tools • Class Rolls OBJECT STORAGE © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 61 OTHER PROJECTS

Structure of the DLP INTERFACE DATABASE MIDDLEWARE INTERFACE • My Folders • Administration, OBJECT Display, and Search STORAGE • The Buzz; What’s New? © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] Knowledge Exchange 62 OTHER PROJECTS

Prototype Interface My Folders A personal space where your uploaded objects as well as links to other objects can be organized. This space will also show the classes and projects you are currently involved in, in the capacity of instructor and/or student. The Buzz / What’s New? This space will include recent additions to the digital library, particularly those related to the visitor’s established areas of interest. The Buzz will specifically display the most accessed objects overall, within my field, etc. This space will also ultimately contain other information, such as live news feeds, both local © 2001 -2005 Franz J. Kurfess [UNC Digital Library 2000] and world-wide. Admin, Display, Search ADMIN: Allows the digital library visitor to publish his own material for others to use as well as arrange the contents of his My Folders space. DISPLAY: Allows for different ways of showing off an individual object or group of objects within one’s space. Some common “displays” here would include slideshows and thumbnails. SEARCH: Allows for basic keyword searching at any point in time, as well as access to more extensive search capabilities. Knowledge Exchange 63

Digital Library in-a-Box © 2001 -2005 Franz J. Kurfess [Sweeney & Kurfess 1998] Knowledge Exchange 64

The EXPECT Project at USC/ISI (I) u EXPECT: architecture to develop knowledge based systems that uses highly declarative representations u [Swartout and Gil, KAW-95], [Gil and Melz, AAAI-96] u http: //www. isi. edu/expect u Application focus u knowledge-rich planning tools, often integrated within a larger architecture u Research focus u knowledge acquisition tools that exploit declarative representations to help end users augment KBs u exploiting declarative representations to mediate agent communication [Blythe&Gil&Chalupsky&Mac. Gregor 00] © 2001 -2005 Franz J. Kurfess [Gil 2000] Knowledge Exchange 65

The What-to-How Spectrum [Feigenbaum 96] The history of software development Software translates “WHAT” to “HOW” © 2001 -2005 Franz J. Kurfess [Gil 2000] Knowledge Exchange 66

The Good News => The Bad News u As we develop more advanced and higher-level languages, we still require formalisms and logic u Complete information u Consistent statements u Executable/compilable u Requires putting knowledge “in a straightjacket” u Represent v captures only the result of the modeling process u Represent v knowledge once it is digested only a subset of the overall knowledge i. e. , whatever fits what the formalism requires u Represent only what there is time to represent v the rest is assumed © 2001 -2005 Franz J. Kurfess unnecessary [Gil 2000] Knowledge Exchange 67

Connections Are Missing! u “There is a real conflict between the logician’s goal and the educator’s. The logician wants to minimize the variety of ideas, and doesn’t mind a long thin path. The educator (rightly) wants to make the paths short and doesn’t mind - in fact, prefers - connections to many other ideas. ” [Minsky 1969] u Knowledge turns up piecemeal u Compartmentalized and seldom rejoined u Only truly necessary views/opinions are represented u Only truly necessary connections are represented u As long as the information is there, the system should figure it out u No value to redundancy/alternatives © 2001 -2005 Franz J. Kurfess [Gil 2000] Knowledge Exchange 68

The Need for Knowledge Mobility: HPKB Challenge Problem - Year 1 LOOM HPKB Upper Ontology BS Ontologies (military units, vehicles, etc) KB of Eng. Actions COA Generation Problem Solver © 2001 -2005 Franz J. Kurfess COA Ontology Eng. Techniques Ontology Features and Terrain (Stub) Ontology Eng. Equpt. Ontology Temporal Estimation/ Assessment Problem Solver Methods for Time Estimation EXPECT [Gil 2000] Knowledge Exchange 69

The Need for Knowledge Mobility: HPKB Challenge Problem - Year 2 © 2001 -2005 Franz J. Kurfess [Gil 2000] Knowledge Exchange 70

The Need for Knowledge Mobility: JFACC ISTI Architecture Plan JTF Plan Server MASTERMIND Refined plan (CORBA) Plan Expanded Plan Critiques Knowledge Base Server (LOOM) © 2001 -2005 Franz J. Kurfess Web-Based Planner CJTF Objectives (Manual Entry) JMTK Map Visualization Plan SDA INSPECT-II [Gil 2000] CPET/APAT Targeting (Automated) Knowledge Exchange 71

The Need for Knowledge Mobility: ARPI MAPVi. S MPA Planning Agents Meta Planning-Cell Manager (PRS) Advisable Planner (SIPE-2) Scheduler (OPIS) Planning-Cell Manager (PRS) Process Monitor APAT GUI Plan Viewer MPA Agent Evaluation Toolkit ACS VISAGE Beddown Critic Plan Server Cue: ACT 2 (TEST (ready unit 1)) ACT 1 Cue: Answer query Annotations Inspect Triggers © 2001 -2005 Franz J. Kurfess [Gil 2000] Knowledge Exchange 72

The Need for Knowledge Mobility u Reuse and integration [Valente et al. , IEEE IS 99] u Ontology/theory manipulation operations [Knight&Luk, AAAI-94] [Hovy et al. 94] [Mac. Gregor& Chalupsky 99] u Aligning, u Morphing merging, slicing, etc. [Chalupsky, KR-2000] u Transformation-based knowledge morphing u Repeated mergers and imports are the norm u Ontology-based translation u Mismatches resolved on a case by case basis, on demand u Hub and spokes model © 2001 -2005 J. Kurfessof v. Franz n-degrees Exchange 73 2000] indirection[Gil if agents have n-degrees. Knowledge of “ontological

What is Needed: Resilient Hyper Knowledge Bases (RHKB) u Extend our view of what a KB contains u Support a variety of knowledge media/formats/representations u Include alternative renderings of knowledge u Everything should remain part of the knowledge base Initial sources: introductory texts, expert hints, explanations, dialogues, comments, examples, exceptions, … v Intermediate analysis and models of those sources as they are digested v Final representations v u Emphasis on knowledge connectivity u Among progressively more digested renderings of the same knowledge © 2001 -2005 Franz J. Kurfess [Gil 2000] Knowledge Exchange 74

Post-Test © 2001 -2005 Franz J. Kurfess Knowledge Exchange 76

Important Concepts and Terms u u u agent knowledge representation linguistics © 2001 -2005 Franz J. Kurfess u natural language Knowledge Exchange 78

Summary Knowledge Exchange © 2001 -2005 Franz J. Kurfess Knowledge Exchange 79

© 2001 -2005 Franz J. Kurfess Knowledge Exchange 80