Concept Maps for Knowledge Capture and Sharing

1
Concept Maps for Knowledge Capture and Sharing

• Yiwen Zhang
• Presentation on Ph.D. Seminar
• May 3, 2002

2
Outline

• Knowledge Learning
• Concept Maps
• Concept Maps in our project
• Research questions and suggested approaches

3
Knowledge Learning
Personal experiences
summarize
Supporting tools?
Learning evaluation?
Individual knowledge
Learned by people
Knowledge in various representations
Contribute to
Get from
Store in
Will it cause Misunderstanding?
Knowledge repository
Whats the efficient way to share?

• Individual knowledge refers to knowledge
  possessed in the mind of an individual.
  (individual/collective knowledge scheme (Nonaka,
  1994)).
• Collective knowledge is both a composite of
  individual knowledge and the social process that
  leads to shared understanding by articulating and
  exchanging information among individuals.(Krippen
  dorff, 1975).

4
What Kind of Knowledge Representation do We Need?

• There is a need for a way of helping people to
  visualize knowledge and maintain and develop a
  common visualization and representation
• Requirement
• Consistent with learning process
• Understandable
• A plus form a bridge between human knowledge and
  machine knowledge

1. John L. Gordon, Creating Knowledge maps by
exploiting dependent relationships, Knowledge
Based Systems, Vol13 (April 2000)
5
Knowledge Representation

• Rules (E.H. Shortliffe 1976)
• If attribute A1 has value V1 and attribute A2 has
  value V2, then attribute A3 has value V3
• Frames
• Elery stone
• Specialization of Frame Person
• Date of Birth 300462
• Sex Male
• Nationality British
• Occupation Tailor
• Health (Consult Medical System)

1. John L. Gordon, Creating Knowledge maps by
exploiting dependent relationships, Knowledge
Based Systems, Vol13 (April 2000)
6
Knowledge Representation

• Semantic Network
• Concept Diagrams
• Conceptual Graphs (Sowa 1993)
• Concept Maps (Novak, 1984)
• Other derivatives Knowledge Maps

7
Concept Map
8
Introduction

• Concept mapping is a graphical representation of
  knowledge.
• Nodes (points or vertices) represent concepts
• Links (arcs or lines) represent the relationships
  between concepts
• In the 1960s, Joseph D. Novak at Cornell
  University began to study the concept mapping
  technique.
• His work was based on the theories of David
  Ausubel (1968)----meaningful learning. Meaningful
  learning involves the assimilation of new
  concepts and propositions into existing cognitive
  structures.

9
Concept Maps for Knowledge Acquisition and Sharing

• Develop an understanding of a body of knowledge.
• Explore new information and relationships.
• Access prior knowledge.
• Gather new knowledge and information.
• Share knowledge and generated information.
• Design structures or processes such as written
  documents, constructions, web sites, web searches
  and multimedia presentations.

10
Questions from Byrons Presentation

• Is it hierarchical?
• Comparison to other graphical representations?
• Which kind of knowledge?
• How do we evaluate maps and learning?

11
Research Directions

• Design and implement concept map tool
• IHMC Cmap Institute for Human and Machine
  Cognition, University of West Florida(Novak)
• WebMap Knowledge Science Institute, University
  of Calgary (Brian R. Gaines), Canada
• Effectiveness in different disciplines
• Eugene V, Aidman Gerald Egan, Academic
  assessment through computerized concept mapping
  validating a method of implicit map
  reconstruction, International Journal of
  Instructional Media, 1998, Vol 25

Mary Jo Carnot, etc, Concept Maps vs. Web Pages
for Information Searching and Browsing
12
Research Directions (continued)

• Evaluation
• Gayle Nicoll, A three-tier system for assessing
  concept map links a methodological study,
  International Journal of science education, 2001
• Collaboration
• Kremer, R. and Gaines, B.R. (1994). Groupware
  concept mapping techniques. Proceedings
  SIGDOC'94 ACM 12th Annual International
  Conference on Systems Documentation.

13
Research Directions (continued)

• Information searching browsing
• Concept Maps vs. Web Pages for Information
  Searching and Browsing, Mary Jo Carnot, et al,
  2000
• Map integration
• Fuzzy Integration and Fuzzy Matching S. W.Chen
  et al. IEEE transactions on systems, man and
  cybernetics-Part B Cybernetics, Vol 31. No. 5,
  October 2001
• Automatic map construction
• A Semi-Automatic Construction Method of Concept
  Map based on Dialog Contents, Hajime Saito et
  al.
• Presentation of concept maps
• Versatile Concept Map Viewing on the Web, Antti
  Karvonen et al

14
Concept maps in our project

• Project description
• Develop tools, algorithms and techniques to
  create learning environments where learners can
  more effectively accumulate and share knowledge
• Concept Map
• Require user to build concept maps
• Use individual maps to form group maps
• Evaluate concept maps
• Use concept maps to organize resources
• Use concept maps to enhance search functions
• Collaboration on building concept maps
• Use concept maps to share knowledge

15
Constructivist Approach A Frame Work for
Integrating Concept Maps in Learning
Learner is actively involved in building on what
he or she already knows to come to a new
understanding of the subject under study
introduce a problem
identify a general area for investigation
Initiation
summarize the topic and prepare to present to the
intended audience
Presentation
Selection
explore information to form a focus
Exploration
Collection
gather information that defines/supports the focus
Formulation
form a personal perspective of the information
encountered
Carol Collier Kuhlthau, Learning in Digital
Libraries an information search process
approach, Libray trends, spring 97
16
GetSmart E-Learning Framework
Get a whole view of the domain
Initiation
Select a starting point
Selection
Read formal material to form basic understanding
Exploration
Collaborative learning
Search to explore further
Discuss/ask/answer
Get suggestion from knowledge database
Formulation
Contribute to build knowledge repository
Build personal concept map
Collection
Presentation
Get evaluation
Merge into group map
Build personal repository on it
Contribute to form Knowledge map
adjust
17
Research Direction 1

• Motivation
• Categorization on relationships to support the
  process of concept mapping
• How to evaluate the map automatically?
• Can we merge individual maps to get expert map?
• Directions
• Can we predefine sets of relationships or concept
  types to facilitate user in building concept
  maps?
• Can group maps merged from individual concept
  maps capture the knowledge more precisely?

18
Differences
Pros expressive, understandable, flexible Cons
poor computational ability
Scott Hamilton Thinking outside the Box at IHMC,
IEEE, Computer, January, 2001
19
Research on Semantic Network Conceptual Graph

• Semantic Network Collins and Quillian (1968)
• Conceptual Graph Sowa, John F.
• (1976) "Conceptual graphs for a database
  interface," IBM Journal of Research and
  Development, vol. 20, no. 4, pp. 336-357.
• Conceptual Structures Information Process in
  Mind, Machine, Reading, Addison-Wesley, MA, 1984
• Sowa, John F. (2000) Knowledge Representation
  Logical, Philosophical, and Computational
  Foundations, Brooks Cole Publishing Co., Pacific
  Grove, CA.
• Automatic Design Research Group, Department of
  Electrical Engineering, University of Virginia
  Polytechnic Institute and State University
• Walling R. Cyre, Knowledge ExtractorA Tool for
  Extracting Knowledge from Text, 1997, Proce
• The Semantic Network Processing System Research
  Group, Department of Computer science and
  Engineering, University of Buffalostart from
  1960s,
• Using Text Processing Techniques to Automatically
  Enrich a Domain Ontology
• Supporting ontology driven document enrichment
  within communities of practice
• Communication of the ACM,  Volume 45 ,  Issue 2
   (February 2002)Special Issue Ontology
  different ways of representing the same concept

20
Related research

• Danseveau and Holley 3-category link
  system(1982)
• Chaining leads to, result in, produces
• Clustering is like, property of, evidence of
• Hierarchical part of, example of
• Collins and Quillian a system of
• relational categories (1979)
• Superset is a , is a member of
• Subset consist of, contains
• Similarity is like, is not like
• Part part of
• Proximity is adjacent to, is next to
• Consequence leads to, influence, cause
• Precedence prior to

• Lambiotte (1989)
• Hierarchy
• Chain
• Cluster
• Procedural
• Influence
• Part

21
Related research - continued

• M. David Merill, Knowledge Objects and Mental
  Models, in IEEE, 2000
• Knowledge structure refers to the
  interrelationships among knowledge components.

• Michael James Young,(National Center for
  Research on Evaluation, Standards, and Student
  Testing) Quantifying the Characteristic of
  Knowledge Structure Representations A
  Lattice-Theoretic Framework, Working Report,
  June, 1998

22
Suggested Approach

• Provide predefined sets of relationships
• Danseveau and Holley 3-category link system
• Provide several structures as templates
• Procedure
• Attribute
• Based on this more formal concept map, use a
  graph theory algorithm to evaluate and merge
  concept maps, and compare to expert maps

23
Experiment Design (ideal)
Dependent Variables
Independent Variables
Independent Variables
Read Map, Refer to readings
Efficiency and precision
Relationships to use
Measurement
Use whatever relationships you can think of
Group 1-1
Time used
Time used
Speed
Build a Concept Map
Group 2-1
Expert evaluation
Use predefined sets with specific relationships
within each set
Test scores
Understanding
Group 2-2
On a given reading
Group 1-2
Test score
Mediators
learning style previous knowledge
Knowledge domains (Date Structure vs. Project
Management) Difficulty levels of knowledge
(undergraduate level vs. graduate level)
Moderators
24
Experiment Design

• Provide the option to use predefined sets of
  relationships and templates
• Measure how students use these from students
  concept maps
• Use an algorithm from conceptual graph research
  to merge individual concept maps
• Evaluate the merged map

25
Other Research Directions -1

• Motivation
• Our system provides powerful search functions,
  how can users searches contribute to a knowledge
  repository?
• According to Liebowitz Beckman (1998), A
  knowledge repository is an online, computer-based
  storehouse of expertise, knowledge, experience,
  and documentation about a particular domain of
  expertise. In creating a knowledge repository,
  knowledge is collected, summarized, and
  integrated across sources.
• Which kind of system design can help users
  contribute to and benefit from a repository?
• What technologies can help to filter, store and
  organize a dynamically increasing knowledge
  repository?

26
Related Research

• Recommendation system
• Content management
• Document clustering
• Usage pattern
• ISO for topic maps

27
Suggested Approach

• Implement flexible database design to support
  content management
• Use clustering technology to categorize
  collections

28
Other Research Directions -2

• Motivation
• We can use group collaboration in building
  concept maps or group projects. Whats the
  relationship between group collaboration and
  performance?
• Which collaboration pattern is the most effective
  and efficient?
• What are the interactions between individual
  leaners and group environments?

29
Related research
Amy Soller, Adaptive Support for Collaborative
Learning on the Internet http//virtcampus.cl-ki.
uni-osnabrueck.de/its-2000/paper/poster5/ws2-poste
r-5.htm
30
Related Research (continue)

• Collaboration patterns among online learners

Chiung-Hui Chiu, etc, Interaction Process in
Network Supported Collaborative Concept Mapping
31
Suggested Approach

• Text mining on group discussion logs
• Identify different collaboration patterns and
  define the characters of learners
• Change the group, trace the behavior of learners
  in different groups