2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

22
gefördert durch das Kompetenzzentrenprogramm www.know-center.at © Know-Center 2011 Computational Support for Work-integrated Learning November 11, 2011 Stefanie Lindstaedt

description

2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

Transcript of 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

Page 1: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

gefördert durch das Kompetenzzentrenprogramm

www.know-center.at

© Know-Center 2011

Computational Support for Work-integrated Learning

November 11, 2011

Stefanie Lindstaedt

Page 2: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

2

Background

Knowledge Management Institute

Competence Center for Knowledge Technologies

Interdisciplinary team of about 45 researchers

Bridge between science and industry

> 400 applied research projects over last 10 years

Page 3: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

3

Power Tools for the Brain

Page 4: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

4

3. Invent

5. Learn1. Discover

2. Visualize 4. Mature

Power Tools for the Brain

Page 5: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

5

Overview

Challenges for Competence Development

Work-integrated Learning

Spectrum of Learning Guidance

Software Framework for Work-integrated Learning Support (APOSDLE)

Application Domains

Evaluation Results

Page 6: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

6

Challenges for Competence Development

Increasing time to competence

7 years for a new engineer entering the corporation [Shell]

Increasing unpredictability

Increasing fluctuation (25% of workers have been with their current employer for less than 1 year)

The amount of new technical information is doubling every 2 years

Half of what technical degree students learn in their first year of study will be outdated by their third year of study

Page 7: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

7

Solutions?

US$ 70 billion spent yearly on formal training [Haskell, 1998]

Seminars and Courses

Learning Management Systems

Web-based Tutors

LifeLongLearning Programs

Why formal training does not suffice

<< 30% learning transfer from formal training to professional workplace

[Robinson, 2003]

Page 8: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

8

Myths about Learning

MYTH 1: ‘Stock-piling’ learning is possiblefalse it is hard to predict which knowledge will be needed in 2 years

MYTH 2: Separation of working and learning activities is possiblefalse they are deeply intertwined [Eraut, 2007]

MYTH 3: Separation of ‘technical content’ and application skills (incl. social skills) is possiblefalse learning happens on tightly intertwined learning trajectories [Hirsh, 2007]

‘A shift from training to learning is sorely needed.’ [CIPD, professional body for trainers and HR managers in the UK, 2004]

Continuous learning at work

Page 9: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

9

Courses▪ pre-described goals▪ structured topic ▪ ask teacher▪ general▪ strong learning guidance

Daily Work▪ short term goals▪ spontaneous search ▪ ask colleagues ▪ work context▪ no learning guidance

Work-integrated Learning▪ learning goals deduced from work context▪ intelligent information delivery based on

work context & competencies▪ using & extending organization wide knowledge▪ dynamic building of learning groups▪ varying degrees of learning guidance

Knowledge Work

Page 10: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

10

Key Distinctions: Learning Perspective

Real Time

Real Work Environment

Real Content

Work-integrated Learning

Page 11: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

11

Courses▪ pre-described goals▪ structured topic ▪ ask teacher▪ general▪ strong learning guidance

Daily Work▪ short term goals▪ spontaneous search ▪ ask colleagues ▪ work context▪ no learning guidance

Knowledge Work

Assisting knowledge workers in advancing their knowledge and skills directly during

their real work tasks

Recommender

Supporting Work-Integrated Learning

Descriptivelearningguidance

Prescriptivelearningguidance

Page 12: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

12

Varying Degrees of Learning Guidance

Suggest Artifacts: contextualized recommendation of knowledge artifacts

Suggest People: contextualized recommendation of people

Browse: navigational support and visualization based on underlying knowledge structures

Learning Hints: expose relations to surrounding topics

Shared Collections: share artifacts, insights, and links to people

Collaboration Wizard: scripted support for collaboration

Triggering Reflection: visualize competences in user profile

Learning Paths: learning resources ordered according to prerequisite relations

GUIDANCE

Page 13: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

13

Key Distinctions: Technical Perspective

Application of semantic technologies together with ‘soft computing’ approaches

Automatic discovery of user context based on user interactions

Automatic inference of user competencies based on task executions

Automatic extraction and mapping of semantic structures based on analysis of backend systems

Automatic identification of similarities based on text, multi-media data and semantic analysis

Automatic maintenance of similarity measures and user profiles based on usage data and user feedback

Page 14: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

14

Robert, Innovation Consultant

… Kick-Off Meeting

User Model

Tasks: • kick-off meeting organized (21) • project proposal developed (57)

Knowledge:• customer-relationship skills (++) • kick-off methods (+)

Creativity Workshop

KnowledgeGap Analysis Recommendations

• Learning Opportunities• Snippets• Documents• Experts & Peers• Learning Hints• Learning Paths•

Adaptive Sytem for WIL Support

Page 15: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

15

Software Framework for WIL Support

Sensors Functions embedded in Tools

LearningGoal

Model

ProcessModel

Domain Model

(Ontology)

OrganizationalIT-Infrastructure

HybridKnowledgeServices

Clients

Users

User ModelUser Context

UICO

Domain independent

Recommendations

Context aware

Utilizing knowledge resourcesfrom within organization

Adaptive

Varying degrees oflearning guidance

Page 16: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

16

Software Framework for WIL Support

APOSDLE: Advanced Process-Oriented Self-DirectedLearning Environment (www.aposdle.org)

Offers support on varying degrees of learning guidance for learning activities within work and learning processes

Domain independent: semantic models, repurposing content from organizational memory for learning (text & multi-media), and employees as learning peers and “tutors”

Considerably reduced modelling efforts for creating domain-specific installations

Embedded within computational work environment

Stable and good level of usability

Page 17: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

17

September 30, 2010 / 17

Client

Page 18: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

18

Application Domains

Simulation ofElectromagnetic Effects

on Aircraft

Intellectual Property RightsManagement

Requirements Engineering

Process

Page 19: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

19

Summative Evaluation: Methodology

⇒Goal: Evaluation of WIL environment effect on learning at work

⇒Setting: workplaces at 3 enterprises for a duration of 3 months

⇒Domains: electromagnetic simulation (EADS), innovationmanagement (ISN), intellectual property rights management(CCI)

⇒Data: User diaries, log data, on-site visits and interviews, exit questionnaire

00,5

11,5

22,5

33,5

44,5

5

Learning material

relevant to current task

Aware of learning material

Existing knowledge improved

High quality learning material provided

APOSDLE learning helped task completion

Learning time planned and

managed

Experts accurately sorted by relevance

Exit questionnaire results (5 = Strongly agree, 0 = Strongly disagree)

Mean

SD

Page 20: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

20

Summative Evaluation: Results

Users clearly favored features with low learning guidance

learners extensively used the different functionalities to browse and search the knowledge structures, followed the provided content suggestions, and collected relevant learning content within collections

Most features with strong learning guidance were rarely used

Hints, notes, and learning paths were only sporadically used and mainly to explore their functionalities

Users did not use Reflection Support to reflect on their activities but rather to examine the environment’s perception of their usage

Are learning support measures derived from instructional theories which are focusing on formal learning contexts not very relevant for learning at work?

Page 21: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

21

Summative Evaluation: Results

WIL support improves task completion ability and supports learning

Domain: Very effective for highly specialized, stable domains in which knowledge is captured in working documents; less effective in broad, customer-driven domains in which knowledge is mainly share in person

Experience Level: Most effective for learners and inexperienced employees in ‘information seeking mode’

Page 22: 2011 11 11 (uc3m) emadrid slindstaedt kmi tug computational support for work-integrated learning

© Know-Center 2011

22

iAPOSDLEwww.APOSDLE.org

Thanks for your Attention!

see also APOSDLE video on YouTube