SPEL Tribalization_ E-learning and the Semantic Web by wpr1947

VIEWS: 13 PAGES: 32

									SPEL: Tribalization, E-learning
   and the Semantic Web

   Arthur Stutt, Cleci Martins,
        John Domingue
              Why SPEL is needed


•   Increased use of e-learning
    BUT e-course components don’t always have
    a full range of tools
    IN PARTICULAR often no tools for
    communication or collaboration
•   Increased availability of learning
    resources
    BUT no means of searching the Web for
    these in a manner which takes account of
    learner’s needs, institutional requirements,
    the nature of learning tasks etc.
                    SPEL



•   Acts as an enriched medium for the
    presentation of the materials forming an
    e-learning course
•   Searches for alternative learning
    materials to supplement the primary
    course materials
•   Provides facilities for communication
    and collaboration among learners
            Interaction Scenario


• Learner chooses course
• Interacts with primary materials using
  guides (for essay writing, argument etc.)
• Selects from alternative material
  available locally or on semantic web
• Initiates discussion with other members
  of a learning community
• Collaborates on task (assignment, model
  etc.)
             Previous work at KMi



• Alice
  Generic personalization interface for e-
  commerce sites
  NOT educational
• OntoWeb Edu
  Web portal with repository of links to
  educational resources
  No Semantic Search
  No Personalization
Tribalization


                             Alice
Personalization




                  SPEL


 Learning
  Objects                OntoWeb Edu

Semantic Web
               Personalization



• Content filtering
  Tracks and make inferences about what
  consumers have purchased in the past in
  order to find new products which might be of
  interest
• Collaborative filtering
  Tries to make inferences about what is
  relevant for a particular consumer based on
  some measure of similarity with other
  consumers and the products they have bought
Personalization and the Movies
         Minority report - Lessons



• Sophisticated systems can focus on
  individuals
• Can offer relevant choices
BUT
• Can be intrusive
• Can get it wrong (John Anderton’s new
  eyes)
    Solution: Converging on E-Tribes For
                  Learning


• Virtual Communities
• Theories about postmodern consumer
• Availability of Tools
• Collaborative/Constructivist theories of
  learning
• Theories about the development of the
  human brain - social intelligence
             Kozinets’ E-Tribes



• Kozinets introduces notion of e-tribes or
  virtual communities of consumption
• An e-tribe is a virtual community whose
  focus is some particular product (Barbie
  dolls) or some consumption activity (wine
  lovers)
• Also interested in knowing about the
  product, acquiring the cultural norms of
  the community etc.
• Structured
• Proactive
• Members Consume as a community
           Benefits of E-Tribes



• Not invasive
• Members have an emotional commitment
  to some product or activity
• Easier to know an e-tribe
• Easier to track
E-Tribes - Members
E-Tribes - Interactions
          The components of e-tribes



•   Members
•   Roles
•   Places
•   Tasks
•   Standards
•   Events
•   Evolution
    The Glue for SPEL - an Ontology of
               Communities


• Ideas from Kozinets and e-commerce
• Already available ontologies
• Ideas from other areas of computer
  science e.g., multi-agent systems
• Cybersociology
  etc.
Semantic Web Community
Generic
Community
Ontology
                Stages in evolution



• Virtual villages
  Highly fragmented; containing multiple, small
  subcommunities
• Concentrated constellations
  Concentration of core communities; development of
  affiliate relationships with niche communities
• Cosmic coalitions
  Core communities aggregate across complementary
  core topic areas
• Integrated infomediaries
  Communities and coalitions evolve into agents for
  members
                           Hagel and Armstrong, Net Gain, 1997
          Need for Dynamic Ontology



•   Changes in roles
•   New sub-groups
•   New topics
•   New goals
•   New standards
•   Groups split away …
       Use of Clustering to dyamically update
                 Learner Ontology

            PERSONALIZATION SCHEME




                                     Student
 Student                             preferences
                                                                            Analysis of
Interface                                                    STUDENTS’
                                                             CLUSTERS
                                                                            clusters
                                     Browsing
                                     data
                                                             Machine
                                                             learning
                                                             If…then..%..           Course
                                                                                  developers/
                                                                                  Institution
                                                                                   Interface




                                                   Student
                                                   data/KB
SPEL - Resource finder

      Query:
      (and (member-of ?learner ?community)
               (has-topic ?community ?topic)
               (has-focus ?resource ?focus)
               (is-sub-topic-of ?topic ?focus))
                         Guides


• Guides are software agents which
  enhance the interaction between
  machines and learners
• Guides were produced at Apple in the late
  1980s as an interface for an educational
  hypermedia database depicting early
  American history
• In SPEL guides have three roles:
      Learners indicate preferences by selecting guides
      Guides will act as a customized means of
      presenting material
      Guides will act as facilitators of communication
      among members of learning communities
    Guide-based Collaboration - Feedback
                 modeling


• Ontology of model components
• PSM for model construction
• Tools for interacting with model
• Discussion space focused on how model
  should be built
• Reference material
BuddySpace
                  Summary


• Gaps - Helps deal with inadequacies in
  material and lack of community dimension
• Easy to Use - Provides a friendly and
  consistent e-learning environment
• Personalization - Tailors resources to student
  and community interests
• Communication - Facilitates communication
  among students with the emphasis on learning
  activities rather than on unfocussed chat
• Collaboration - Encourages students to work
  on learning tasks as a part of a group rather
  than on their own
                SPEL in future



•   Course sequencing
•   Course configuration
•   Guides for argumentation
•   Combined guides

								
To top