presentation by Andreas Schmidt at 2nd International Workshop on Learner-Oriented Knowledge Management and KM-Oriented Learning (LOKMOL 06), in conjunction with the First European Conference on Technology-Enhanced Learning (ECTEL 06), Heraklion, Oct. 4-6 2006

1. Simone Braun
Andreas Schmidt
Socially-Aware Informal Learning Support:
Potential and Challenges of
g
the Social Dimension
FZI Research Center for Information Technologies
Information Process Engineering
Karlsruhe, Germany
{Simone.Braun | A d
{Si B Andreas.Schmidt}@fzi.de
S h idt}@f i d
http://www.fzi.de/ipe

2. Motivation: The Social Context Matters!
Paradigm shift in technology-enhanced learning
g gy g
• from formal & organized towards informal, networked
• instead of "content is king", now "context is king" (P. Baumgartner)
• process-embedded learning, learning in context, ...
b dd d l i l i i t t
• new system paradigm: learning support in open environments
instead of closed learning environments
The social dimension
• roles of learner and teacher are changing situatively
• growing importance of social networks (learning is a social activity)
• informal teaching situations: it does matter "who is the other side"
Should learning support not become socially aware as well?
2

3. Outline
Potentials
• why do we want systems to be socially aware?
y y y
• what are showcases for these potentials?
Challenges
• why is it hard?
• what prevents us from large-scale usage?
• what are first steps to cope with the challenges?
3

4. Potentials

5. Social People Finder & Network Management
First set of shortcomings of expert finder applications
• do learners always want to have experts?
• doesn't the social relationship towards the expert matter?
Social factors
• asking for help means admitting a weakness, exposing vulnerability
• subjective relevance matters
o not: th ultimate expert ( bj ti relevance)
t the lti t t (objective l )
Generalizable to partner search (research projects
projects,
cooperative industry projects)
5

6. Socially-Aware Mediation of Communication
Second set of shortcomings of expert finder applications
g p pp
• the side of the expert is not appropriately considered
• annoyance of the expert because of overload
But...
• it is not objective overload and bad timing
• rather: missing consideration of how the informal teacher views the
quality of social relationship
First solution: balance the recommendation of teachers
Second solution: communication mediation
• take into account interests of both sides to avoid annoying forms of
communication
6

7. Example: Context-Aware Mediation of
Communication
Braun & Schmidt, I-KNOW 2006
m O
7

8. Socially-Aware Opinion Sharing & Resource Ranking
Willingness to share your evaluation of what is relevant and good gives
g y g g
away a competetive advantage
Analysis in applied research institutes revealed
• high degree of uneasiness of sharing with anyone
(especially competing institutes)
• but also high degree of willigness to share with individuals to whom you have
good relationship (even in competing institutes)
• trusting the opinion of others also depends on your subjective opinion of the
person
We need
• socially-aware sharing policy
• socially-aware ranking for results (those we trust most)
8

9. Challenges

10. Challenges
1 Social Context is hard to describe
2 Social Context is hard to acquire
3 Social Context is hard to make use of
4 Social Context is a sensitive issue
10

11. What do we need?
1 Social Relationships Ontology
2 Social Relationship Mining
3 Methodological framework for socially-aware
learning support
4 Privacy preservation methods
11

12. 1 Social Relationship Ontology
Describe social relationships
• at different levels of abstraction
• along different criteria
o trust, loyalty, expectancy of reciprocity, reliability …
• from a subjective point of view
Existing Approaches
• several FOAF extensions
• activities in social network analysis
they take an objective (whole-network) point of view
Visualization in sociograms
12

13. 2 Social Relationship Mining
Wide range of social networking approaches
• low level of differentiation
o usually a relationship intensity, co-occurence, or roles
• objective point of view
Better approach: egocentric networks
• mailbox
• address book, buddylists
• interconnecting with social networking platforms (people tagging)
13

14. 3 Methodological Framework
Contextual factors are usually interdependent with respect to
the impact on "relevance"
• how is the effect of a certain quality of relationship
• to which degree does the social relationship influence it
• when do I want social awareness?
o serendipity effects
Empirical studies are needed
• sociology can contribute
• but usually not differentiated enough
14

15. 4 Privacy Preservation
Already the low detail of SNA is p
y problematic
• who is the outsider
• who is the hub
• whole network approaches usually conflict with privacy protection laws
h l t k h ll fli t ith i t ti l
(e.g., in Germany)
With the increasing level of detail and the subjective opinion
g j p
about the quality of relationships:
• revealing asymmetric opinions
• damaging effect of making explicit
Principle: only for the individual and adaptation to the
individual, no global analysis
, g y
Problem: socially-aware applications can reveal the quality of
relationships indirectly (via their adapted behavior)
15

16. Summary
Social awareness of learning support seems to be the next
frontier of context awareness
• potential to provide more (subjectively!) relevant recommendations
• but poses even more challenges, especially with respect to the
sensitivity of the data
• social relationship data will be damaging beyond one s own
one‘s
environment
Next steps
• test ideas within scientific collaboration processes
16

17. http://www.im wissensnetz.de
http://www.im-wissensnetz.de
Andreas Schmidt
d s d
Team Manager „Knowledge and Learning“
FZI Research Center for Information Technologies
Information Process Engineering
Andreas.Schmidt@fzi.de
http://fzi.de/ipe
17