CPE/CSC 580: Knowledge Management

1
CPE/CSC 580 Knowledge Management

• Dr. Franz J. Kurfess
• Computer Science Department
• Cal Poly

2
Course Overview

• Introduction
• Knowledge Processing
• Knowledge Acquisition, Representation and
  Manipulation
• Knowledge Organization
• Classification, Categorization
• Ontologies, Taxonomies, Thesauri
• Knowledge Retrieval
• Information Retrieval
• Knowledge Navigation
• Knowledge Presentation
• Knowledge Visualization

• Knowledge Exchange
• Knowledge Capture, Transfer, and Distribution
• Usage of Knowledge
• Access Patterns, User Feedback
• Knowledge Management Techniques
• Topic Maps, Agents
• Knowledge Management Tools
• Knowledge Management in Organizations

3
Overview Knowledge Management Techniques

• Motivation
• Objectives
• Evaluation Criteria
• Chapter Introduction
• Review of relevant concepts
• Overview new topics
• Terminology
• Topic 1
• Subtopic 1.1
• Subtopic 1.2

• Topic 2
• Subtopic 2.1
• Subtopic 2.2
• Topic 3
• Subtopic 3.1
• Subtopic 3.2
• Important Concepts and Terms
• Chapter Summary

4
Logistics

• Introductions
• Course Materials
• textbook
• handouts
• Web page
• CourseInfo/Blackboard System and Alternatives
• Term Project
• Lab and Homework Assignments
• Exams
• Grading

5
Knowledge Repositories
KPMG 1998
6
KM Infrastructure
7
KM Initiatives
8
Pre-Test
9
Motivation
10
Objectives
11
Evaluation Criteria
12
Corporate Memory (CM)

• definition attempts
• purpose
• concepts
• implementation

13
Definition Attempts Corporate Memory

• explicit, disembodied, persistent representation
  of knowledge and information in an organization
  Van Heijst, van der Spek and Kruizinga 1996
• may include knowledge on products, production
  processes, clients, marketing strategies, plans,
  strategic goals, etc.
• the collective data and knowledge resources of a
  company Nagendra Prasad and Plaza 1996
• may include project experiences, problem-solving
  expertise, design rationale, etc.

Dieng et al. 1999
14
Purpose Corporate Memory

• capitalization of knowledge
• integration of resources and know-how
• cooperation through effective communication and
  active documentation
• the right knowledge to the right person at the
  right time and at the right level

Dieng et al. 1999
15
Links in the Knowledge Chain

• list existing knowledge
• determine required knowledge
• develop new knowledge
• allocate new and existing knowledge
• apply knowledge
• maintain knowledge
• dispose of knowledge

Dieng et al. 1999
16
Corporate Memory Management

• detection of needs
• construction of the corporate memory
• diffusion of the corporate memory
• use of the corporate memory
• evaluation
• maintenance and evolution

Dieng et al. 1999
17
Corporate Memory Management Overview
Dieng et al. 1999
18
Multidisciplinary Perspective on CM

• technological (computer science, information
  technology)
• concentrate on technical and implementation
  aspects
• may neglect requirements and constraints of
  systems in practical use
• organizational (CKO)
• emphasize the role of CM in an organization
• may overlook technological problems, or
  underestimate efforts needed for implementation

Dieng et al. 1999
19
Corporate Memory Techniques
Dieng et al. 1999
20
Corporate Memory Example
Dieng et al. 1999
21
Motivations for Establishing a CM

• avoid knowledge loss
• departure, retirement, change of roles of
  employees
• exploit past experience
• cumulative technical know-how
• successful and failed projects
• utilize collective knowledge for strategic
  purposes
• detection of new opportunities
• reaction to changes
• improve knowledge exchange and communication
• establish venues for sharing information
• improve learning
• integrate knowledge from different areas
• cross-disciplinary knowledge exchange

Dieng et al. 1999
22
Knowledge in Organizations

• explicit knowledge
• specific know-how to design, build, sell and
  support products and services
• tacit knowledge
• individual and collective skills enabling the
  organization to act, adapt, and evolve
• tangible knowledge components
• data, procedures, plans, models, algorithms,
  documents of analysis and synthesis
• intangible knowledge components
• abilities, professional skills, private
  knowledge, organizational culture, history of the
  organization, contexts of decisions, etc.

Dieng et al. 1999
23
Types of Corporate Memories

• technical memory
• know-how of the employees about technical aspects
• organizational memory
• knowledge about the internal structure of an
  organization
• project memories
• lessons and experiences from past projects
• individual memories
• status, know-how, activities, relationships of
  individual employees
• internal vs. external memory
• indicates the source of relevant knowledge and
  information

Dieng et al. 1999
24
CM Needs

• organization is also a knowledge production unit
• not necessarily as primary purpose
• depends on size, type, and organizational scheme
  of the organization
• e.g. distributed network of consultants
• needs of individual users vs. organizational
  needs
• detecting the right needs can be difficult
• target users, domains, tasks, situations,
  knowledge

Dieng et al. 1999
25
Determination of CM Needs

• stakeholder-centered
• influenced by the members of the community of
  people affected by or invested in the system
• requirements analysis
• early involvement of stakeholders is critical and
  feasible
• most stakeholders are internal to the
  organization, and many are motivated
• most solutions are adaptations or evolutions of
  previous systems
• CSCW, KBMS, MIS, ...

Dieng et al. 1999
26
CM Construction

• sources
• non-computational CM
• document-based CM
• knowledge-based CM
• case-based CM
• distributed CM
• project-centered CM
• combinations of several techniques

Dieng et al. 1999
27
Sources

• human sources
• domain experts, experienced specialists, people
  with organizational memories
• physical documents
• printed documents, notes, design artifacts,
  products, tools, etc.
• digital documents
• reports, technical documentation, design
  artifacts, email, case libraries, dictionaries,
  sketches, etc.

Dieng et al. 1999
28
Non-computational CM

• establishment of paper-based knowledge repository
• existing documents
• generation of new documents
• synthesis of knowledge not explicit in reports,
  technical documentation, etc.
• improve strategies and structural aspects of the
  organization
• systematic generation of knowledge in an
  organization
• may be the predecessor to a digital CM

Dieng et al. 1999
29
Document-based CM

• comprises all existing documents in an
  organization
• may be in paper-based or digital form
• organizes the collection in a systematic way
• indexing
• interface to manage documents
• preparation, storage, retrieval, processing,
  evaluation, distribution

Dieng et al. 1999
30
Knowledge-based CM

• based on the elicitation and explicit modeling of
  knowledge from experts
• may use a formal knowledge representation
  framework
• this is often quite expensive
• serves as an assistant to human knowledge
  workers
• different from traditional expert systems
• their goal is the automation of a particular task

Dieng et al. 1999
31
Case-based CM

• utilizes case-based reasoning
• past experiences are collected in a (semi-)formal
  representation mechanism
• allows the comparison of cases
• the assumption is that new problems can often be
  solved by looking up solutions to previous
  problems
• helps with the concentration of expertise around
  specific cases
• continuous evolution of the CM through the
  continuous addition of new cases

Dieng et al. 1999
32
Distributed CM

• emphasis on collaboration and knowledge-sharing
  across traditional boundaries
• geographically distributed persons/groups
• structurally separated entities
• common tasks, domains
• essential for virtual organizations
• teams or people collaborate on-line

Dieng et al. 1999
33
Project-centered CM

• captures the relevant knowledge accumulated while
  working on a project
• discussions, arguments, decisions, compromises,
  etc.
• important aspects
• represent and reconcile perspectives of different
  stakeholders
• changes of priorities in the project
• communication of decision rationales
• recovery of insights and solutions from past
  scenarios
• re-inventing the wheel
• example
• issue-based information system (IBIS) Rittel
  1972

Dieng et al. 1999
34
Combinations of Several Techniques

• informal and formal knowledge representation
  methods
• combination of paper-based and digital documents
• semi-automatic extraction of knowledge
• collaborative construction of community
  knowledge
• integration of existing components
• libraries, data bases, case bases, document
  collections, multi-media collections, etc.

Dieng et al. 1999
35
Diffusion and Use of CM

• diffusion modes
• knowledge attic
• archive that can be consulted when needed
• collection and diffusion are passive
• knowledge sponge
• active collection, passive diffusion
• knowledge publisher
• relevant elements are distributed to users
• passive collection, active distribution
• knowledge pump
• specific roles or methods for collection of
  relevant knowledge
• active collection and active diffusion

Dieng et al. 1999
36
Diffusion via Intranet/Internet

• frequently centered around Web servers
• has some conceptual and technical limitations,
  but substantial benefits
• confidentiality, security, reliability,
  distraction, etc.

Dieng et al. 1999
37
Knowledge and Information Retrieval

• traditional index-based techniques are integrated
  in most approaches to CM
• enhancements through advanced techniques
• ontologies
• collaborative filtering
• intelligent agents

38
Evaluation

• financial perspective
• improve the bottom-line of the organization
• may be difficult to measure
• organizational perspective
• work environment
• employee satisfaction
• technical perspective
• transfer of know-how
• some effects may not be direct consequences of
  the CM, but side-effects of its introduction or
  use

Dieng et al. 1999
39
Maintenance and Evolution

• should be based on the evaluation of the current
  situation
• addition of new knowledge
• removal or modification of obsolete knowledge
• coherence problems
• scalability
• user acceptance
• should become a continuous activity

Dieng et al. 1999
40
Examples of CM Methods

• CYGMA
• REX
• MKSM
• KAMM

Dieng et al. 1999
41
CYGMA

• Cycle de Vie et Gestion des Métiers et des
  Applications, KADE-TEX
• construction of a professional memory in
  manufacturing
• relies on six categories of industrial knowledge
• singular knowledge
• terminological knowledge (dictionary)
• structural knowledge (ontology, factual knowledge
  base)
• behavioral knowledge
• strategic knowledge
• operational knowledge

Dieng et al. 1999
42
REX

• needs analysis and identification
• construction of elementary pieces of experiences
• construction of a computer-based representation
• implementation through a software system

Dieng et al. 1999
43
MKSM

• Method for Knowledge System Management
• systemic-based decision support method
• views knowledge assets as a complex system
• models this complex system through different
  perspectives
• syntactical, semantic, pragmatic
• different components
• information (data processing)
• signification (task modelling)
• context (activity modelling)

Dieng et al. 1999
44
KAMM
Knowledge Associates 2000
45
KAMM Architecture
Knowledge Associates 2000
46
Knowledge Technology Framework

• identifies key KM activities and related
  knowledgeoriented techniques and tools
• personalization
• codification
• discovery
• creation/innovation
• capture/monitor

Milton et al. 1999
47
Knowledge Technology
(Key P"Person, K1"Knowledge 1echnology,
I1"Information 1echnology)
48
Personalization

• sharing knowledge through person-to-person
  contacts
• tools for more effective communication
• email, message boards, chatrooms, personal
  ontologies

Milton et al. 1999
49
Codification

• capturing existing knowledge and placing it in
  repositories
• tools and techniques for knowledge representation
• generic models
• rules, frames, case-based reasoning, ...
• specialized techniques
• task- or domain-specific

Milton et al. 1999
50
Discovery

• searching and retrieving knowledge from
  repositories and data bases
• tools and techniques from information retrieval,
  knowledge-based systems, natural language
  processing
• search engines, ontologies

Milton et al. 1999
51
Creation/innovation

• generation of new knowledge
• tools and techniques from cognitive science,
  psychology
• brainstorming support, creativity assistance
• mainly a human endeavor

Milton et al. 1999
52
Capture/Monitor

• capturing knowledge as people work on their
  normal task
• tools and techniques from Human-Computer
  Interaction, AI
• audit trails, case collections

Milton et al. 1999
53
KM Framework
Macintosh et al. 1999
54
KM Processes
Macintosh et al. 1999
55
PROMOTE Architecture
Karagiannis Telesko, 2000
56
PROMOTE Framework
Karagiannis Telesko, 2000
57
(No Transcript)
58
Organizational Memory Context
Abecker et al. 1998b
59
Context-Sensitive Knowledge Supply
Abecker et al. 1998b
60
Integration of Ontologies
Abecker et al. 1998b
61
Knowledge Task Support
Abecker et al. 1998b
62
Related Research Areas
Abecker et al. 1998b
63
Developing a KnowledgeManagement TechnologyAn
Encompassing View on the Projects of
theKnowledge Management Group at DFKI
Kaiserslautern

• Michael Sintek, Andreas Abecker, Ansgar Bernardi
• German Research Center for Artificial
  Intelligence
• Kaiserslautern, Germany

Abecker et al. 1998b
64
Overview
Development of Knowledge Management technology of
the Knowledge Management Group at DFKI
Kaiserslautern
requirements and approaches to support KM
infrastructures for organizations related
research fields
KnowMore active knowledge supply finished
Know-Net collaboration ongoing
FRODO distribution, framework current
MOTIVE 3D access planned
summary we propose a rich, modular KM middleware
as a solid basis for engineering intranet-based
KM solutions
Abecker et al. 1998b
65
Knowledge is an Important Productivity Factor for
Organizations

• besides labor, capital, and land, knowledge has
  been recognized as an important productivity
  factor
• knowledge is stored in individual brains or
  implicitly encoded and hidden in organizational
  processes, documents, services, and systems

KM is concerned with discovery, acquisition,
creation, dissemination, and utilization of
knowledge.
Abecker et al. 1998b
66
Organizations Have Serious Problems in Managing
Their Corporate Knowledge
Distribution
Discovery
Accessibility
Acquisition
Resources
KnowledgeProblems
Multiple Views
Documentation
Multiple Formats
Awareness
Availability
Various fields of computer science tackle some of
these knowledge problems.
Abecker et al. 1998b
67
Resarch Fields Related to KM

• Groupware, Workflow, CSCW
• collaboration of individuals and departments
• Document management, retrieval, and filtering
  systems
• most of the available abstract, strategic
  knowledge written down in text-based documents
• often advertised as KM solutions
• Artificial Intelligence
• formal ontologies
• data mining
• case bases
• expert systems

We strive for a new quality of knowledge systems
by integrating all these areas.
Abecker et al. 1998b
68
KnowMoreKnowledge Management for Learning
Organizations

• basic research project funded by German
  government
• central idea access to multiple heterogeneous
  knowledge sources
• enabled through comprehensive knowledge
  description using several formal ontologies
  (information, domain, enterprise ontology)
• active information delivery integrated into
  business processes
• explicit representation of context

In KnowMore, knowledge can be viewed as
information linked into the application context.
Abecker et al. 1998b
69
The KnowMore System Architecture
Abecker et al. 1998b
70
Know-NetKnowledge Management with Intranet
Technologies

• funded by the European Commission within the IT
  for learning and training industry program
• integrate groupware functionalities with AI
  methods enabling the handling of knowledge
  objects
• based on Knowledger suite (Lotus Notes
  application from Knowledge Associates) and
  intelligent agents (DFKI)
• intranet- and agent-based knowledge platform
• codification, mapping, sharing, and reuse of
  explicit knowledge in multimedia content
• corporate knowledge ontologies
• intelligent navigation, searching, filtering

In addition to a KnowMore-like knowledge
platform, collaborative aspects play an important
role.
Abecker et al. 1998b
71
Know-Net Collaborative Aspects

• collaborative tools supporting communities of
  practice at the team level to facilitate the
  creation of shared memories and interpretative
  context
• real-time group discussions/meetings
• project-based bulletin boards and forums
• on-line topical conferences with threading
  features and interactive expertise databases
• Know-Net mainly exploits the collaboration and
  coordination technology provided by Lotus Notes
  and add-on products like Sametime

Abecker et al. 1998b
72
The Know-Net Intranet- and Agent-Based System
Architecture
Abecker et al. 1998b
73
FRODOA Scalable OM Framework for Evolutionary
Growth (future work)

• basic research project funded by German
  government, successor project of KnowMore
• KnowMore global set of ontologies, centralized
  inference
• FRODO conjointly use knowledge from several
  independent knowledge sources
• legacy databases
• independently introduced partial OMs based on
  specific ontologies
• external knowledge sources (with own ontologies)

• ontology mapping problem
• communicating and cooperating services

We propose a rich, modular KM middleware as a
solid basis for engineering intranet-based KM
solutions.
Abecker et al. 1998b
74
The FRODO KM Middleware Will Exploit Various
Notions of Agents

• digital reference and acquisition librarians
• know their respective knowledge source and
  organization principles
• know how to effectively access, search, maintain
  the knowledge
• wrappers, mediators, ontologists, knowledge
  brokers
• add intelligent interfaces to legacy systems
• make sources accessible to higher-level
  inferences
• document analysis and information extraction
  specialists
• allow transition between informal and formal
  representations
• task/process agents, knowledge push/pull
  mechanisms
• manage workflow enactment
• realize context-sensitive information supply

Abecker et al. 1998b
75
A Sample Instantiation of the FRODO OM Framework
Abecker et al. 1998b
76
MOTIVEFostering Individual Users Motivation for
Accessing Online Learning Training Resources
(planned)

• will be submitted to the EU 5th framework
• online front-end to electronic learning and
  training (LT) systems
• addresses users motivation important driving
  factor is social interaction
• MOTIVE proposes an environment that wraps LT
  tools and content together with peoples
  interactions
• virtual representation of the LT environment
• workspace with 3D representation of the
  organization and of knowledge assets
• avatars associated to users
• wizard agents with specific roles for promoting
  available material
• support for social processes events
  organization, social places (café) etc.

Abecker et al. 1998b
77
MOTIVE Adds Access to LT OMs Through 3D
Knowledge Portal

• the LT contents is accompanied by a
  KnowMore/FRODO-like knowledge meta-level based
  upon various ontologies
• XML as upcoming standard will be used for this
  knowledge representation task
• a 3D knowledge portal wraps these ontologies to
  provide a highly motivating access to the LT
  resources
• thus, the MOTIVE 3D knowledge access can be
  viewed as an additional, but highly user-friendly
  information retrieval aspect of the general KM
  scenario

In general, 3D spaces can be used to replace
legacy information retrieval, knowledge
acquisition, and workflow frontends of OM
systems.
Abecker et al. 1998b
78
Summary

• In our view, KM technology is a combination of
• distributed, heterogeneous knowledge sources
• various formal ontologies (information, domain,
  enterprise)
• knowledge meta-descriptions
• informal-formal transitions
• workflow, active support, context
• collaboration
• framework, middleware, agents
• user-friendly access through 3D spaces

Abecker et al. 1998b
79
Reference Kearns 00
Dieng et al. 1999
80
Reference Sommerville 01

• Sommerville 01

Sommerville 01
81
Post-Test
82
Evaluation

• Criteria

83
References

• Abecker et al. 1998 Andreas Abecker, Ansgar
  Bernardi, Knut Hinkelmann, Otto Kühn, Michael
  Sintek. Techniques for Organizational Memory
  Systems. Technical Report D-98-02, Deutsches
  Forschungszentrum für Künstliche Intelligenz
  (DFKI), 1998.
• Abecker et al. 1998b Andreas Abecker, Ansgar
  Bernardi, Knut Hinkelmann, Otto Kühn, Michael
  Sintek. Toward a Technology for Organizational
  Memories. IEEE Intelligent Systems, vol. 13,
  no.3, pp. 40-48, 1998.
• Dieng et al. 1999 Rose Dieng, Olivier Corby,
  Alain Giboin and Myriam Ribiere, Methods and
  Tools for Corporate Memory. Int. J.
  Human-Computer Studies, no. 51, pp. 567-598,
  1999.
• Karagiannis Telesko, 2000 Dimitris
  Karagiannis and Rüdiger Telesko. The EU-Project
  PROMOTE A Process-oriented Approach forKnowledge
  Management. Proc. of the Third Int. Conf. on
  Practical Aspects of Knowledge Management
  (PAKM2000) Basel, Switzerland, 30-31 Oct. 2000,
  (U. Reimer, ed.).
• KPMG 1998 KPMG Management Consulting Knowledge
  Management Research Report 1998.
• Macintosh et al 1999 Ann Macintosh, Ian Filby,
  and John Kingston. Knowledge Management
  Techniques - Teaching and Dissemination Concepts.
  Int. J. Human-Computer Studies, no. 51, pp.
  549-566, 1999.
• Milton et al. 1999 Nick Milton, Nigel Shadbolt,
  Hugh Cottam, and Mark Hammersly. Towards a
  Knowledge Technology for Knowledge Management.
  Int. J. Human-Computer Studies, no. 51, pp.
  615-641, 1999
• Sintek et al. 1998 Michael Sintek, Andreas
  Abecker, Ansgar Bernardi. Developong a Knowledge
  Management Technology. Presentation at WET ICE
  KMN 99, Deutsches Forschungszentrum für
  Künstliche Intelligenz (DFKI), 1999
  www.dfki.uni-kl.de/simtek/.

84
Important Concepts and Terms

• natural language processing
• neural network
• predicate logic
• propositional logic
• rational agent
• rationality
• Turing test

• agent
• automated reasoning
• belief network
• cognitive science
• computer science
• hidden Markov model
• intelligence
• knowledge representation
• linguistics
• Lisp
• logic
• machine learning
• microworlds

85
Summary Chapter-Topic
86
(No Transcript)