Visual Browsing and Editing of TM-based Learning Repositories

1
Visual Browsing and Editing of TM-based Learning
Repositories

• 1Boriana Ditcheva 2Darina Dicheva
• 1Columbia University, NY
• 2Winston-Salem State University, WS
• USA

2
OR

• The Long Way to a Usable TM Editing
  Interface Are we halfway through?
• The Case of TM4L

3
Introduction

• TMs provide a bridge between the domains of
  knowledge representation and information
  management and serve as the skeleton of
  ontology-aware applications, including digital
  learning repositories
• The lack of convenient tools that allow authors
  to directly enter, modify, index, and query
  resources in ontology-aware digital repositories
  is the major obstacle to their deployment

4
TM4L

• An environment for building, maintaining, and
  using ontology-aware e-learning repositories
• a free tool http//compsci.wssu.edu/iis/nsdl/
• interface in 8 languages
• Users
• authors, with limited/no background of ontologies
  
• learners, seeking information support in their
  learning
• Functionality
• supports 3 interaction tasks editing, browsing
  and querying

5
The Problem

• Ontology-aware collections are typically large
  and complex networks
• ontology engineers lack domain expertise
• domain expert (resource authors) lack ontology
  engineering skills to build them
• Most of existing tools assume users experience
  in structuring and classifying resources (based
  on specific representation models)

6
Scenario 1 Boriana wants to create a Java TM

• Boriana develops ontology-driven portals
• ------------------------
• Relationship
• Supeclass-Subclass

• Java Ontology
• Expressions
• Arithmetic Expressions
• Logical Expressions
• Relational Expressions
• Java Data Types
• Primitive Type
• Reference Type
• Class Type
• Array Type

7
(No Transcript)
8
TM4L Interface V. 1

• TM4L Editor interface is a typical tree
  rendering, with the left pane showing the tree
  and the right pane showing the properties
  (facets) of each selected node. The nodes of the
  tree are topics and the edges denote the default
  binary superclass-subclass association
• Topic attributes, resources, topic parents and
  relations are displayed in separate panels
• Four additional relationship types are
  predefined part-whole, class-instance,
  similar-to and related-to
• By offering this minimal set of five predefined
  relation types we support TM4L authors that
  experience difficulties in articulating and
  naming relationships

9
Scenario 2 Christo wants to create a Prolog TM

• Prolog
• 1 Presentation of Prolog
• 1.1 What is a Prolog program
• 1.1.1 The Program
• 1.1.2 The Query
• 2 Facts
• 3 Variables and Unification
• 4 Rules
• 5 Backtracking
• 6 Recursion

• Christo teaches an AI course
• ------------------------
• Relationships
• Whole-Part
• Related

10
(No Transcript)
11
Partonomy

• Whole-Part relationship
• Has an important explanatory role in e-learning
  context
• Explaining what a learning unit is about often
  involves describing its parts and how are they
  composed
• For example, in structuring learning material on
  Programming Languages in terms of its components,
  the learning units describing the syntax,
  semantics and pragmatics are part of the
  Programming Languages unit and not subclasses of
  it
• By emphasizing the compositional structure, the
  partonomy is closer to the approach normally used
  by instructors/authors for representing learning
  content

12
TM4L Interface V. 2

• The same five predefined relationship types
  superclass-subclass, part-whole, class-instance,
  similar-to and related-to.
• However, three different views on the topic map
  are supported Taxonomy (superclass-subclass),
  Partonomy (part-whole) and Topic Typing
  (class-instance).
• The user chooses which view to work in and can
  change it at any time
• partonomy is the default
• The nodes of the tree are topics and the edges
  denote the default binary part-of or the
  relation chosen by the user (superclass-subclass
  or class-instance)

13
(No Transcript)
14
(No Transcript)
15
Scenario 3 Darina wants to create a Photos TM

• Photos
• Where
• What kind of trip
• When
• Who on the photo
• What on the photo
• Additional info

• Darina travels (a lot) and makes a lot of photos
• ------------------------
• Relationships
• Trip-Photo
• Whole-Part

16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
Ontology visualization

• Visualization in the form of graphs helps a user
  comprehend and analyze information better
• A number of graphical interfaces available
• Protégé TGVizTab, Jamballaya, and OntoViz
• In the area of Topic Maps Ontopias Vizigator,
  TMNav, ThinkGraph, the LIP6 visualization tool
• Most of the GUIs offer only visualization,
  browsing, and search capabilities
• Some interfaces also provide editing
  functionality, e.g. IsaViz , OntoViz, ThinkGraph
  
• Usually no support for synchronized textual
  editing

33
TM4L Visual Editing Summary

• Functionality
• Create a new topic
• Rename topic
• Delete topic along with its occurrences and
  associations
• Connect topics with an existing association
• Connect topics with a new association (incl. asso
  type defining)
• Delete association
• Synchronized with the textual editing pane

34
Conclusion

• Automatic topic map construction is not yet an
  alternative to the manual educational TM design
• Automatic TM acquisition an expensive operation
  with limited reliability
• Even if initially accurate and complete, TMs need
  modifications / adaptations reflecting their
  evolution
• To foster the development of Topic Maps-based
  e-learning applications, authors need adequate
  authoring/editing environments
• Use of a balanced cooperative modeling
  construction approach
• TM4L proposes browsing and editing in one view
  authoring environment

35
Conclusion (contd)

• TM4L supports three interaction tasks editing,
  browsing and querying
• Undergoing work on querying
• Tolog queries
• Template-based queries
• Relations queries
• Resource queries
• Context queries
• Paper at the Semantic Web User Interaction
  workshop at ISWC 2006, Athens, USA, November 6th,
  2006
• Much more research needed for designing elegant
  and efficient interfaces that provide multiple
  views on a collection and offer users different
  perspectives on a selected entity