Semantic Web and PeertoPeer Technologies for Distributed Learning Repositories

1
Semantic Web and Peer-to-Peer Technologies for
Distributed Learning Repositories

• Wolfgang Nejdl et al
• L3S Learning Lab
• University of Hannover
• Stanford University
• Contributors too numerous to mention here

2
Overview

• PADLR Personalized Access to Distributed
  Learning Repositories
• Project Context / Research Aspects
• Learning Objects and Metadata
• Edutella - Main Idea RDF P2P
• Resource Description Framework (RDF)
• Peer-to-Peer (P2P)
• Edutella Query Service / RDF Query Exchange
  Language RDF-QEL-i
• Edutella Infrastructure and Services
• Annotation and Modification (Ontomat, RDF-MEL)
• Peer-to-Peer (P2P) and Efficient Routing
• Personalization in an Open Environment
  (extending conventional adaptive hypermedia
  research)
• ELENA Ambient Intelligence for Learning Spaces

3
PADLR / Edutella Context

• Personalized Access to Distributed Learning
  Repositories (www.learninglab.de/english/projects/
  padlr.html)

• Most important (CS) modules
• Peer-to-Peer Infrastructure (incl. Clients and
  Providers)
• Courseware Watchdog and Metadata Extraction
• Personalization and Personalized Queries based on
  Metadata
• Web-(Learning-) Services
• PADLR Participants
• (Stanford, Hannover,
• Karlsruhe, Stockholm, Uppsala)
• Edutella Participants
• (Vienna, Berlin, Darmstadt,
• etc.)

4
PADLR / Edutella Some Research Aspects

• Some Aspects in Organizing and Using Distributed
  Learning Materials
• Exchanging learning materials based on
  educational metadata standards (LOM, SCORM)
  part 1
• Accessing distributed learning materials in a
  world-wide peer-to-peer network (Edutella) part
  2
• Personalization functionalities in a distributed
  environment part 3
• Automatic metadata extraction, classification and
  resource discovery
• Integration in commercial LMS
• Web-Services for Learning

5
Learning Object Standards / Metadata

• Definition A Learning Object is any entity,
  digital or non-digital, that may be used for
  learning, education, and training.
• Dublin Core (DC) 15 Metadaten Elemente
  dublincore.org
• title, creator, subject, description, publisher,
  contributor, rights
• Simple, flexible, extendible
• IEEE Draft Standard for Learning Objects Metadata
  (LOM)(IEEE Learning Technology Standards
  Committee (LTSC) ltsc.ieee.org
• General (title, description, keywords)
  introduces taxonomies / ontologies for dcsubject
• Lifecycle (development and current state)
• Technical (Information on media, type, size,
  software requirements)
• Educational (pedagogical characteristics
  interactivity level, typical learning time,
  semantic density)
• IMS Content Packaging / IMS CP (part of SCORM
  www.adlnet.org )

6
Overview

• PADLR Personalized Access to Distributed
  Learning Repositories
• Project Context / Research Aspects
• Learning Objects and Metadata
• Edutella - Main Idea RDF P2P
• Resource Description Framework (RDF)
• Peer-to-Peer (P2P)
• Edutella Query Service / RDF Query Exchange
  Language RDF-QEL-i
• Edutella Infrastructure and Services
• Annotation and Modification (Ontomat, RDF-MEL)
• Peer-to-Peer (P2P) and Efficient Routing
• Personalization in an Open Environment
  (extending conventional adaptive hypermedia
  research)
• ELENA Ambient Intelligence for Learning Spaces

7
RDF Describing Resources on the Web

• Resource Description Framework

• Key representation language in the Semantic Web

• URI-based identifying mechanism to describe
  distributed resources, and state relationships
  between these resources

• RDF graph model makes it easy to integrate a
  number of other formats for recording
  information, e.g. tables/tuples in RDB, simple
  assertions in formal logic, etc.,

• RDFS offers a mechanism to define specific RDF
  vocabularies, enabling schema / ontology level
  meta-data mapping.

8
RDF/RDFS What is missing?

• RDF/RDFS describe distributed resources on the
  Web and to describe the vocabulary (properties)
  and constraints (classes, domain/range) for these
  descriptions
• How do I query these distributed sources ???
• How do my queries reach their destination?
• How do I ensure they are understood at their
  destination?

9
Edutella Distributed RDF-Metadata in a P2P
Network

• Edutella connects highly heterogenous peers
  (heterogenous in uptime, performance, storage
  size, functionality, number of users...), which
  describe their resources as RDF metadata
• Distributed nature of Edutella services (e.g.
  repository storage) is transparent to Edutella
  clients
• Means to get there specification and
  implementation of a set of Edutella services

10
Edutella Goal and Approach

• Specify and implement a RDF-based meta-data
  infrastructure for P2P networks
• Developed as part of the open source peer-to-peer
  project JXTA
• edutella.jxta.org
• gt 50 contributors from various institutions

11
Background JXTA

• Project JXTA (www.jxta.org)

• An open source programming platform to enable
  P2P services and applications
• Interoperability, Platform independence, Ubiquity

12
Edutella JXTA Layers
JXTA service layer

• JXTA-enabled Edutella Services

• Edutella Query Service query Edutella provider
  peers and retrieve query results

• Edutella Annotation Service annotate meta-data
  stored in repositories

• Edutella Update and Replication Service update
  and replicate (meta-) data

• other services (mediation, mapping, etc.)

13
RDF-QEL RDF Query Exchange Language

• RDF-based Query Exchange Language (RDF-QEL)

14
Edutella Query Service Architecture
15
Edutella Query Class Model (ECDM)

• Details see Nejdl et al EDUTELLA A P2P
  Networking Infrastructure Based on RDF, WWW 2002

16
Concept Browser
4 Edutella Consumers have been implemented
Conzilla (Query GUI) Implementer KTH, Stockholm
17
SWEBOK Demonstrator
Ontology Query (Query GUI) Implementer Stanford
18
Annotation in Ontomat
Ontomat (Annotation tool) Implementer AIFB,
Karlsruhe

• Details see Nejdl, Wolf, Staab, Tane EDUTELLA
  Searching and Annotating Resources within an
  RDF-Based P2P Network, SWWS_at_WWW 2002

19
Edutella Annotation / Replication / Update

• Transmit Update and Replication Requests within
  the Edutella Network see Nejdl, Siberski, Simon,
  Tane Towards a Modification Exchange Language
  for Distributed RDF Repositories, ISWC 2002

20
Overview

• PADLR Personalized Access to Distributed
  Learning Repositories
• Project Context / Research Aspects
• Learning Objects and Metadata
• Edutella Main Idea RDF P2P
• Resource Description Framework (RDF)
• Peer-to-Peer (P2P)
• Edutella Query Service / RDF Query Exchange
  Language RDF-QEL-i
• Edutella Infrastructure and Services
• Annotation and Modification (Ontomat, RDF-MEL)
• Peer-to-Peer (P2P) and Efficient Routing
• Personalization in an Open Environment
  (extending conventional adaptive hypermedia
  research)
• ELENA Ambient Intelligence for Learning Spaces

21
P2P and Efficient Routing

• How do peer-to-peer networks scale?
• Requirements
• Symmetric topology (every node is a root)
• Low network diameter (small worlds property,
  should be O(log n))
• Limited node degrees (number of peer-connections
  from a node, should be O(log n))
• Load balancing of traffic
• Efficient broadcast (receive broadcast messages
  only once)
• Adaptable to dynamic number of peers

22
HyperCuP Peer-to-Peer Topology

• Details see Schlosser, Sintek, Decker, Nejdl
  HyperCuP Shaping Up Peer-to-Peer Networks,
  AP2PC 2002

23
How to Personalize Open Hypermedia

• Historically, adaptive Hypermedia Systems have
  been closed systems using
• Content level adaptation
• Link level adaptation
• Content Level Adaptation
• Modifies the content of pages in the hypermedia
  system based on a user profile
• Link Level Adaptation
• Modifies the link structure in the hypermedia
  system based on a user profile

24
(No Transcript)
25
Personalization in Open Hypermedia

• Extending adaptive functionalities (especially
  link level adaptation) to distributed
  environments
• The open hypermedia system consists of the
  learning objects / pages present in the P2P
  network
• RDF metadata describe properties of these pages
• dcsubject describes the content / topics of
  these pages
• Values of dcsubject (can be multi-valued) are
  topics from taxonomies / ontologies to describe
  domains of discourse / topic hierarchies
• Express link level adaptation and annotation as
  personalized queries over RDF metadata in a
  distributed environment (RDF-QEL)
• Use Learner Profiles (e.g. IEEE P1484.2) to
  support personalized learning spaces, learner
  information is used as a parameter to these
  personalized queries
• First ideas Henze, Nejdl Knowledge Modeling
  for Open Adaptive Hypermedia, AH 2002, see also
  the paper by Bailey et al

26
Example Metadata and Queries on P2P Network

• Using TRIPLE notation, see TRIPLE papers from
  Decker/Sintek, this example based on the example
  in Stefan Decker, Michael Sintek, Wolfgang
  Nejdl TRIPLE A Logic for Reasoning with
  Parametrized Views over Semi-Structured Data
• _at_semdocuments semd_01_01 dctitle -gt
  TRIPLE---A Query, Inference, and Transformation
  Language for the Semantic Web
  dccreator -gt Michael Sintek dccreator gt
  Stefan Decker dcsubject -gt RDF
• dcsubject -gt triples .....

27
Peer1 Educational Resources

• _at_l3slearningobjects l3sL3SP1 rdftype
  -gt l3sPDF dctitle -gt Introduction to
  Requirements Engineering
• dccreator -gt _ vCardfn -gt Friedrich
  Steimann vCardorg -gt University of
  Hannover lom_clstaxonomy -gt
  swebokswebok
• dcsubject -gt swebokReqDocument
  dcqisPartOf -gt l3sL3.
• l3sL3dctitle -gt Requirements Engineering.
  ...
• FORALL P1, P2 P1 dcqhasPart -gt P2 lt-
  P2 dcqisPartOf -gt P1.FORALL LO, SuperTopic,
  SubTopic LO dcsubject -gt SuperTopic
  lt- LOdcsubject -gt SubTopic AND
  SuperTopiclom_clssubtopic -gt
  SubTopic_at_swebokswebok.

28
Peer2 SWEBOK Ontology, and Peer3 Student
Profiles

• _at_swebokswebok
• swebokReqSpecification lom_clssub_topic -gt
  swebokReqDocument lomclssub_topic -gt
  swebokDocQuality.
• ....
• _at_l3sstudents
• l3sstudent1
• l3sname -gt Ingo Brunkhorst
• l3sknows -gt _ dcsubject -gt
  swebokReqDocument l3srating -gt 1
  _ dcsubject -gt swebokDocQuality l3sratin
  g -gt 0.5
• ...

29
Peer4 Query for Unknown Learning Objects

• FORALL LOModel, StudentModel, Topic, UserName
• _at_Unknown_LOs (LOModel, StudentModel, Topic,
  UserName)
• FORALL LO, C, T, R LO dccreator -gt C
• dctitle -gt T
• l3srating -gt R
• lt- LO dccreator -gt C
• dctitle -gt T
• dcsubject -gt Topic_at_LOModel
• AND _ l3sname -gt UserName
• l3sknows -gt _ dcsubject -gt Topic
• l3srating -gt R_at_StudentModel AND R lt
  0.8

30
ELENA Creating Personalized Learning Spaces

• Details see EU/IST-Project Description,
  www.elena-project.org

31
Summary and Outlook

• P2P and RDF are a perfect match for providing
  distributed and heterogeneous information
  services -gt EDUTELLA
• Accessing and managing learning objects is an
  important application area for such a distributed
  and metadata-based infrastructure -gt PADLR and
  other projects
• Intelligent (and Distributed) Services can be
  built on top of that Personalization (Hannover),
  Annotation (Karlsruhe), Mediation (Uppsala), etc.
• Use P2P and RDF to integrate resources not only
  on the Web, but also in distributed environments
  involving learning management systems (ELENA) and
  digital libraries (TIB)