intelligent

1
OPEN-INTELLIGENT-EFFECTIVE Barbara
Kieslinger kieslinger_at_zsi.at Centre for Social
Innovation
2
Corporate Learning Management....state of the art
OPEN
3
Current drawbacks

• Lack of transparency of knowledge offerings
• Increased search costs
• Personalization is site-based (e.g. everything
  you buy at Amazon.com) vs. process-based
  personali-zation (e.g. gain know-how on Web
  Services)
• No decision support

OPEN
4
Corporate Learning Management...future
Publishing House
Magazins,Journals
TommorrowsKnowledge Worker
OnlineBookstore
OPEN
Books
OnlineMarketplace
Smart Space for Learning
PLA
Courses
ToolWeb Browser PersonalLearning Assistant
(PLA)
Intranet
Tutorials
Intranet
Best PracticeStudies
5
RequirementsSystem interface framework
OPEN
6
First Prototype
OPEN
7
Issues

• BrokerUBP-centred approach less flexible ?
  administrative and performance bottleneck
• Adoption of common schema by provider increases
  implementation costs ? local mapping instead of
  enforced schema
• Lack of interoperability on schema level ?
  mapping of search results required
• High implementation effort for plugging in new
  node? sound interface definition required
• Performance of Peer-to-Peer Network?
  Identification of performance bottlenecks

OPEN
8
Requirements for a query interface

• The query interface must be designed to query
  heterogeneous metadata schemas.
• The query interface must not require a specific
  query language.
• The query interface must build upon existing
  standards of the W3C such as XML, RDF, and SOAP.
• The query interface must not require a specific
  network architecture.
• The query interface must be based on a
  light-weight design, which allows for a fast
  implementation.
• First public draft http//nm.wu-wien.ac.at/e-lea
  rning/interoperability/query.pdf
• Submission to CEN/ISSS Standardization
  Workshop(Work Item on Interoperability of
  Repositories for Learningby Erik Duval, Simos
  Retalis, and Bernd Simon)

OPEN
9
Next Milestone
OPEN
10
What is a Smart Space?

• A Smart Space for Learning is a system, which
  aims to manage the distribution and consumption
  of Learning Services via a Personal Learning
  Assistant.
• Space depicts a network of educational nodes,
  designed for the provision of a heterogeneous set
  of learning services. Smart refers to the use
  AI techniques (e.g. reasoning) and shall ensure
  an optimized selection of learning services.

INTELLIGENT
11
Who is smart?

• PLA, the personal learning assistant, performs
  the search for suitable learning services based
  on the learner's individual profile, processes
  the selected services and supports the evaluation
  of the learner.
• Learners with different background
• Learners with different goals, preferences, aims
• Learners with different learning styles

INTELLIGENT
12
The Situation

• Distributed content/services
• Distributed standard based metadata descriptions
  about
• Content/Services
• Relationships between the content/services
• Learner
• Logic Programs
• Query
• Add restrictions to queries
• Enhance metadata

P2P
INTELLIGENT
13
Summary of Personalisation Approach of ELENA

• Service/Content metadata seen as some constraints
  on use for learning services (LS)
• Learning service metadata are retrieved according
  to matching between learner profile and LS
  metadata
• Rules determine how adaptation is performed based
  on the matching learner profile and LS metadata
• Standards LOM and Dublin Core

INTELLIGENT
14
Learner Profiling

• Why To be able to customize (adapt) information
  to specific person
• Different user characteristics can support
  personalization
• Current State
• IEEE PAPI
• IEEE Public and Private Information (PAPI) for
  learner (IEEE P1484.2/D7, 2000-11-28)
• IMS LIP
• IMS Learner Information Profile (v1.0, 2001-3-9)
• Which way to go? (standard analysis, needs
  analysis)

INTELLIGENT
15
Resulting Profile

• ..... after comparison
• Study performance
• Performance, portfolio, certification
• Identification
• Calendar
• HRP
• Job, Title, Department
• Other user features
• Preference, Goal, Interest

INTELLIGENT
16
PLA Search Interface
INTELLIGENT
17
PLA Results Interface
INTELLIGENT
18
What we need Smart Spaces for?

• We want to assist in the process
• of making training management
• more effective for the
• corporate world, by
• introducing Smart Spaces for
• Learning.

EFFECTIVE
19
Interviews

• Interview procedure -- 2nd
• round
• 1.Introduction to the study
• 2.Pre-questionnaire
• 3.Successful / Failed Cases
• 4.Scenario validations
• combined with
• questionnaire
• 5.Claims analysis for
• different artefacts and
• features
• 6.Wrap-up and conclusion

EFFECTIVE
20
Findings 1/2

• Companies spend resources on life-long learning
  (1-1,5 of EPIDA)
• Older, larger companies still have centralised
  behaviours, deciding employee training at boards
• Employees build perceived trust with customers
  by getting certified degrees
• Employees have annual meeting with HR manager.
• Companies assign their training budget on an
  employee-basis or on a department-basis.
• Personnel departments aim to maintain training
  accounts per employee in order to get the
  training partly refunded when the employee leaves
  the company within a specific time period.

EFFECTIVE
21
Finding 2/2

• Companies choose learning services from
  recognised companies/institutions
• Companies allow departmenthead/employees to
  choose learning services
• Employees some times take
• 'wrong decisions when it
• comes to the selection of
• learning services

EFFECTIVE
22
Reflections on a Smart Space

• A corporate demand for a well-managed central
  market place for learning services does exist.
  (Google for learning services).
• Well-managed means that providers are carefully
  selected, references to providers do exist, and
  the portfolio of the learning service offer
  exceeds a significant size.
• Learning service markets are local. Building a
  single European corporate portal for learning
  services does not make sense. Language and
  location are too big barriers. à Most effective
  filtering attributes period when learning
  service is intended to be consumed, anguage and
  location
• E-Learning is hardly accepted by the
  corporations. Sometimes failed projects do exist.
  Hence, the learning service portfolio also needs
  to include presential courses.

EFFECTIVE
23
User centred approach

• Involve users e.g. companies with training
  management
• in design
• trials
• and surveys

EFFECTIVE
Smart Space for Learning
24

• Thank you!
• www.elena-project.org