The Multimodel, Metadatadriven Approach to Content and Layout Adaptation

1
The Multi-model, Metadata-driven Approach to
Content and Layout Adaptation

• Owen.Conlan_at_cs.tcd.ie
• Knowledge and Data Engineering Group (KDEG)
• Trinity College, Dublin

2
Overview

• Adaptive Hypermedia Systems and Services
• Methods of Adaptivity
• Metadata for Representing Adaptivity
• Multi-Model, Metadata Driven Approach to Adaptive
  Hypermedia Services
• Narrative, Architecture
• Adaptive Layout
• Layout Model
• Multiple Adaptive Engines

3
Adaptive Hypermedia Systems

• What are the components of a typical AHS?
• A User model (may be individual or stereotypical)
• A mechanism to produce personalized content
• Why are AHSs difficult to maintain?
• The content and the rules that govern how that
  content is personalized are usually intertwined
• This makes it difficult to
• Add/Modify new content
• Change the structure of the content
• Use only a sub-section of the content

4
Adaptive Hypermedia Systems
AHS Engine
Personalized Content
5
User, Device, Environment, etc.
Context Modelling
Context Information
6
Methods of Adaptivity

• Adaptive Presentation
• Personalization of content delivered
• Adaptive Navigation
• Dynamically generated navigation and paths
• Historical Adaptation
• Time context
• Structural Adaptation
• Spatial representations

7
(No Transcript)
8
Multi-model, Metadata Driven Approach

• Metadata to describe Adaptive Resources
• Multi-model
• Two versions of the approach
• 3 Models Content, Learner and Narrative (PLS)
• N Models At least one Narrative, the rest are
  metadata based (APeLS)

9
Metadata for describing Adaptive Resources 1

• Developed as part of EASEL (IST Project 10051)
• Educator Access to Services in the Electronic
  Landscape
• Appropriate Descriptive Metadata to facilitate
  discovery and reuse of Adaptive Electronic
  Learning Objects
• Extension of IEEE LOM and IMS LRM

10
Metadata for describing Adaptive Resources 2

• Current specifications dont facilitate the
  description of Adaptive Resources
• Full Adaptive Hypermedia Systems
• Reusable Adaptive Components
• As part of EASEL the IMS Learning Resource
  Metadata v1.2 was extended to facilitate the
  complex nature of Adaptive Learning Resources

11
XML Metadata Representation
ltadaptivitygt ltadaptivitytype namecompetencies.r
equired refgt ltset typeallgt
ltcandidategt ltlangstring
langengtFunctions.Conceptlt/langstringgt
ltlangstring langdegtFunktionen.Konzeptlt/langstri
nggt lt/candidategt ltcandidategt ...
lt/candidategt lt/setgt lt/adaptivitytypegt lt/adapti
vitygt
12
Basic Schema View for Adaptivity

• adaptivitytype
• nameltlangstringgt
• refltURIgt?
• set?
• typeone-or-moreall...
• set
• candidate
• langstring

13
Multi-Model, Metadata Driven Approach

• The Multi-model, Metadata Driven approach
  separates the models used in adaptation (e.g.
  Narrative, Learner and Content) from each other
• Provides a generic run-time engine for
  interpreting Narratives and reconciling models to
  produce an adaptation effect.

14
Simple 3 Model Architecture
Narrative Models
Content
Learner Models
15
Multi-model Approach Requirements

• Separate
• User Model
• Pertinent information that the system can use to
  personalize to the users preferences
• Content Model
• Describes the individual pieces of content
• Narrative Model
• Describes how the content can be
  structured/sequenced for different needs
• Other Models
• Device, Environment, Layout etc.
• Provide appropriate alternative candidates
• Provide an abstraction layer and selection
  criteria

16
Multi-model Approach Narrative 1

• The Narrative Model is
• The Embodiment of a Domain Experts Knowledge
• Represented in Jess (Expert System Shell for
  Java)
• Responsible for assembling the personalized
  course
• The Narrative can access any metadata in the
  repositories
• Narrative is described at a conceptual level,
  i.e. it does not refer directly to learning
  content.

17
Multi-model Approach Narrative 2

• There may be multiple Narrative Models for a
  single course
• There is a Candidate Narrative Repository
• Each Narrative also has associated metadata
• A Narrative may be comprised of sub-narratives

18
Multi-model Approach - Candidates

• What are candidates?
• Elements that fulfil the same role
• Pieces of content that cover the same material
• Narratives that produce courses from the same
  content body
• but achieve that role differently
• The content candidates may be textual, graphical
  or interactive
• Narrative candidates may support different
  approaches to learning

19
Candidate Content Groups

• A Content Candidate is a pagelet and its
  associated metadata
• A Candidate Content Group contains Candidates
  that fulfil the same learning objective, but are
  implemented differently
• The Narrative can refer to Groups rather than
  individual pieces of content
• Most appropriate Candidate selected at runtime by
  looking at the Learner model

20
Multi-model Approach Abstraction and Selection

• Abstraction
• Narratives are built using concept names rather
  than content identifiers
• Enables the service to use the most appropriate
  candidate
• Selection
• There criteria used to select a candidate from a
  group of potential candidates are based upon
• The candidates metadata
• The learners metadata

21
A Generic Architecture

• The Adaptive Hypermedia Service is designed to
  facilitate multiple tiers
• Each tier can achieve one (or more) axes of
  adaptivity
• Facilitated by metadata
• Supported by an extensible AI mechanisms

22
Adaptive Hypermedia Service APeLS Architecture
Learner Modeler
Learner Metadata Repository
Learner Input
Adaptive Engine
Transform
Content Metadata Repository
Rules Engine
Candidate Selector
Personalized Course Model (XML)
Candidate Content Groups
Personalized Course Content
Candidate Narrative Groups
Narrative Metadata Repository
Content Repository
Narrative Repository
23
What about Layout?
Context Information
Layout Strategy
Stylesheet Elements
Stylesheet Elements
Learner Models
Tailored Layout Model
Layout
24
Adaptive Layout
Learner Modeler
Learner Metadata Repository
Learner Input
Adaptive Engine
XSLT Transform
Content Metadata Repository
Rules Engine
Candidate Selector
Personalized Course Model (XML)
Candidate Content Groups
Personalized Course Content
Candidate Narrative Groups
Narrative Metadata Repository
Tailored Layout Model
Content Repository
Narrative Repository
25
Multiple Adaptive Services(APeLS II)
Metadata
26
Summary

• Adaptive Hypermedia Services can deliver
  information personalised for the users needs
• They can also tailor delivery towards environment
  and device (Context)
• Personalization and Adaptation may be facilitated
  by appropriate metadata
• The tiers of the multi-model, metadata approach
  may be used to implement different axes of
  adaptivity

27
Thank You!

• Owen.Conlan_at_cs.tcd.ie
• Knowledge and Data Engineering
• Group (KDEG)
• Trinity College, Dublin

www.iclass.info
http//kdeg.cs.tcd.ie
www.m-zones.org