Adaptive Authoring of Adaptive Hypermedia

1
Adaptive Authoring of Adaptive Hypermedia

• Maurice Hendrix
• Eindhoven University of Technology
• L3S Research Center
• Hannover

2
Outline

• Goal and Motivation
• Problem description
• Initial Environment
• Theoretical Aspects of Development
• Implementation
• Demo
• Conclusions Solutions, Problems further work

3
Goal and Motivation

• We want to make the authoring task easier
• Manual annotation is the bottleneck of authoring
  of adaptive hypermedia
• We want to do this by integrating an authoring
  environment into a semantic desktop environment

4
Problem description plan

• We want to make the authoring task of adaptive
  hypermedia easier by integrating and authoring
  system and a semantic desktop environment
• Find out how to achieve this integration and how
  to (semi-)automatically populate GM DM
• Extend MOT with the necessary functionality

5
Initial Environment (MOT)

• My Online Teacher (MOT) authoring environment
  for adaptive hypermedia
• Beagle semantic desktop environment

6
My Online Teacher (MOT)

• WWW authoring of adaptive hypermedia system using
  5 layer model, most important for us are DM
  (containing concepts) and GM (containing lessons)
• Domain model (DM)
• Goal and constraints model (GM)
• Adaptation model (AM)
• User models (UM)
• Presentation model (PM)

7
MOT hierarchy structure

• Concept maps and lessons are conceptual
  hierarchies
• We can represent them as tree structures

8
MOT CAF

• Mot uses portable CAF XML format CAF format is
  commonly used among other Adaptive Hypermedia
  systems
• CAF describes the lesson in the goal model and
  the uses conceptmaps in the domain model

9
MOT CAF example

• test
• treename
• 
  title
• 
  tree
• 
  
• test\tree\
  title

10
Beagle

• Extended Beagle desktop search tool
• Generates and stores meta-data concerning files,
  emails, publications, bookmarks
• Uses a Lucene full-text index and a Sesame RDF
  store for the meta data
• Both indices are updated upon events (like file
  creation, file adaptation)

11
Beagle meta-data

• REF metadata describing resources like articles
  with attributes like conference, stored_as
• Example

12
Theoretical Aspects Enriching

• The enricher application queries the Sesame
  database store via seRql queries which look like
  this
• SELECT Title
• FROM bplusPublication arttitle bplusTitle
  
• WHERE bplusPublication bplusAttachment
• The results are given a ranking by using one of
  the following methods and are added to the
  concept with the highest ranking

13
Theoretical Aspects Formula

• Where k(x) the bag of keywords in x
• a the publication in Sesame
• c the concept in MOT
• With an extended notion of intersection

14
Theoretical Aspects Extended Algorithm

• IF (card(k(n1) n k(a)) card(k(n2) n k(a))) THEN
• IF (card(k(n2) card(k(n1)) THEN connect to n1
• ELSEIF (card(k(n2) n2
• ELSEIF (card(k(n1) card(k(n2)) THEN
• IF (n1 higherthan n2) THEN connect to n1
• ELSE connect to n2
• ELSEIF (card(k(n1) n k(a)) card(k (n2) n k(a)))
  THEN connect to n1
• ELSEIF (card(k(n1) n k(a)) THEN connect to n2

15
Implementation

• Import / export of CAF XML format in PHP
• Export of lessons in RDF format by XSLT
  transformation of CAF
• Java application which parses CAF XML format,
  searches Sesame RDF store and adds articles to
  the lesson

16
Demo
17
Conclusions work performed

• The task of manual authoring of adaptive
  hypermedia has been made easier by the
  integration of the authoring environment with the
  semantic desktop environment
• By the application of some techniques for
  malleable schemas evolving environments have been
  made possible to some degree

18
Conclusions further work

• Obtaining full flexibility towards evolving
  schemas
• Automatic generation of adaptation rules and more
  advanced labelling and weighting
• More general integration of authoring systems
  with semantic desktop environments