1. Adaptive and Intelligent Collaborative
Learning Support systems (AICLS)
Magnisalis Ioannis
Aristotle University of Thessaloniki, Greece 2011
2. Presentation flow
– Background & Work up to now
– Research Directions & Future Roadmap
– Publications
Aristotle University of Thessaloniki, Greece 2011
3. Background
• Adaptive systems
• Intelligent systems
• CSCL
– Theory
(Pedagogic perspective)
• Classification scheme
– Focus on Target of Intervention
• Group formation (Pre-task)
• Knowledge domain support (In-task)
• Peer Interaction (In-task)
• Assessment (Post-Task) - FUTURE
Aristotle University of Thessaloniki, Greece 2011
4. Work up to now 1/8
• Standards and widespread technologies
– IMS-LD, QTI, Moodle, LAMS, Webconference
tools,BPEL, Semantic Web & Ontologies …
• Adaptation patterns modeling (IRMO design
specification)
• Architecture for AICLS systems (MAPIS)
• Case studies
– Pre-task (group formation)
– In-task (Moodle forum): Mirroring & Meta-
cognitive level
Aristotle University of Thessaloniki, Greece 2011
5. Work up to now 2/8
Adaptation Pattern Design Specification (IRMO)
db or manifest
MODELLED ENTITIES
Interaction On Screen
Analysis Representation
INPUT RULES OUTPUT
ADAPTAT I O N PAT T E R N
• During design:
– Define monitored parameters (e.g. from interaction analysis tools)
– Rules (the adaptation model of the pattern) are hard-wired to the pattern
– Define Output (form, content, etc.)
Aristotle University of Thessaloniki, Greece 2011
6. Work up to now 3/8
During design…
• Linking APs to Script Design Activity Flow
Script Representation
………………………….
AP: ‘Advance the SCRIPT PHASE:
Advanced’ Individual Study
………………………….
6 / 37
Aristotle University of Thessaloniki, Greece 2011
7. Work up to now 4/8
Running AP
Advanced Study: The learner is
Output of the Adaptation Pattern: The activity of prompted to study the
“Advanced_Study” is presented to an advanced learner. In advanced material and answer
contrast his partner – novice learner – is guided to normal some relevant questions.
study (not shown in this screen capture)
SLED screenshot of the adapted user interface
for the advanced learner
according to the implemented adaptation pattern
7 / 37
Aristotle University of Thessaloniki, Greece 2011
8. Work up to now 5/8
MAPIS Architecture
(i.e. Mediating Adaptation
Patterns & Intelligent
Services)
Problem: adaptive LDs
with external
tools & use IMS-
LD and EA with as
little as possible
intervention
Solution: SOA
architecture with
Web services as
main constituent
Requirements:
a) Interoperability,
b) extensibility
Aristotle University of Thessaloniki, Greece 2011
9. Work up to now 6/8
Proof of concept: ‘Group Heterogeneity’ adaptation
as an example case study/scenario
According to the IRMO specification this adaptation is described as
follows:
• Input: the outcome of a prior knowledge questionnaire which is
used as a measure of learners’ expertise,
• Model: Prior domain knowledge of each learner & Mean of Prior
domain knowledge of all participants & number of groups &
number of participants,
• Rule: IF Group work is needed THEN provide new groups of mild
heterogeneity (complex rule which entails calculations of a)
number of groups, b) best distribution within them),
• Output: Form New Groups mildly heterogeneous according to
prior domain knowledge.
Aristotle University of Thessaloniki, Greece 2011
10. Work up to now 7/8
IMS-LD is interconnected with
GF component and 1
Moodle
Pre-task adaptation (AP)
3
2
Aristotle University of Thessaloniki, Greece 2011
11. Work up to now 8/8
IMS-LD is interconnected
with Moodle forum 1
In-task adaptation (AP)
2
3
Mirroring
vs Meta-cognitive
Aristotle University of Thessaloniki, Greece 2011
12. Research Directions 1/3
• Implement courses in
real environments
– Educational
– Workplace
• Investigate use of
various tools
– Synchronous (e.g.
Web-conferencing)
– Asynchronous (e.g.
Wikis)
• Interconnect all these
in an AICLS system
under a framework A Webconference system as a candidate to
incorporating design & external system to communicate with IMS-LD
architecture players
suggestions
Aristotle University of Thessaloniki, Greece 2011
13. Research Directions 2/3
Re-Course screenshot Adaptation Pattern: ‘Advance the Advanced’
………………………….
Monitored Parameter(s)
How is the Advanced
Add parameter
Learner Defined?
SCRIPT PHASE:
MaxIndividualof
number Study 3
learners in Group
Resources
candidate tools to work upon are ………………………….
What is Adapted Activity
Re-Course Editor and
Webcollage Other
introduce adaptation patterns as A possible interface of a s/w component
components/ services/tools in facilitating AP application
the form of a toolbox
Aristotle University of Thessaloniki, Greece 2011
14. Research Directions 3/3
As a next step we are already working in designing a complete course
in IMS-LD and Moodle providing adaptive support at three
distinct levels in a pyramid script:
• Pre-task adaptation: for example a questionnaire in Moodle to
activate specific learning activities
• In-task adaptation: providing hints and careful interventions in
discussion within a Moodle forum according to participation
levels monitored in Moodle and set into IMS-LD.
• Post-task adaptation: Assessment of the CSCL process from the
students can provide ratings for the hints introduced during in-
task adaptation. The system should not use hints rated as not
helpful in a next run.
– Target is a system that evolves by its use (LEGO system)
– Assessment that is adaptive itself
– Built of folksonomies instead of Ontologies
Aristotle University of Thessaloniki, Greece 2011
15. Publications
• S. Demetriadis, I. Magnisalis and A. Karakostas, “Adaptation Patterns in Systems for
Collaborative Learning and the Role of the Learning Design Specification”, Scripted vs.
Free CS collaboration: Alternatives and paths for adaptable and flexible CS scripted
collaboration Workshop in CSCL2009, Rhodes, 2009, pp. 43-47.
• I. Magnisalis, S. Demetriadis, “Modeling adaptation patterns with IMS-LD specification: a
case study as a proof of concept implementation", International Conference
on Intelligent Networking and Collaborative Systems (INCoS 2009), Barcelona, 2009.
• Ioannis D. Magnisalis, Stavros N. Demetriadis, Andreas S. Pomportsis, “Implementing
Adaptive Techniques in Systems for Collaborative Learning by extending IMS-LD
capabilities", International Conference on Intelligent Networking and Collaborative
Systems (INCoS 2010), Thessaloniki, 2010 (accepted).
• I. Magnisalis, S. Demetriadis, “Modeling adaptation patterns in the context of
collaborative learning: case studies of IMS-LD based implementation", Technology-
Enhanced Systems and Adaptation Methods for Collaborative Learning Support, (under
revision).
• Magnisalis, Ioannis; Demetriadis, Stavros; Karakostas, Anastasios; , "Adaptive and
Intelligent Systems for Collaborative Learning Support: A Review of the Field," Learning
Technologies, IEEE Transactions on , vol.4, no.1, pp.5-20, Jan. 2011
doi: 10.1109/TLT.2011.2
• D. Meimaridou, I. Magnisalis, S. Demetriadis, A. Pomportsis, “Web conferencing to
support blended learning in the school context: a case study in a Second Chance School ”,
ICICTE 2011, (under review).
Aristotle University of Thessaloniki, Greece 2011
16. Technological background extensions and interests
• Java, Javascript, PHP, MySQL
• Web services, BPEL, Semantic Web
• Ontologies, Rule-based systems
• Annotation, Rating systems
• Wikis, Forum, Chat, web-conferencing
tools
Aristotle University of Thessaloniki, Greece 2011
17. Thank you for your attention!
Questions?
Contact:
E-mail: imagnisa@csd.auth.gr
Department of Informatics, AUTH: http://www.csd.auth.gr
Multimedia lab: http://mlab.csd.auth.gr/
Aristotle University of Thessaloniki, Greece 2011
18. Common
• SOFCLES project
• LEADFLOW4LD
• GSI
• CLFPs and APs
19. Scenario
• GF
• IA
• Weights of hinttype
• Adaptive behavior updated every time is run
• Two types of implementation: Client (php/java)
or BPEL based data of complex scenario e.g. IA &
GF: role reallocation or IA from various tools in an
activity (e.g. drawing & forum & chat)
• Linked data org
21. Semantics over MAPIS (SMAPIS)
Annotation
with:
OWL,
OWL-S,
RDF,
BPEL,
MPEG-21
Catalog of Tools/services
with semantics.
Wookie ++ Input, Output
of Widgets (IRMO)
More than WADL, WSDL
http://code.google.com/apis/explorer/#_s=translate&_v=v2&_m=translations.list&q=good%20morning&target=EL
http://code.google.com/apis/ajax/playground/?exp=libraries#translate
http://code.google.com/intl/el-GR/apis/discovery/
Aristotle University of Thessaloniki, Greece 2011
23. Scenarios…
• Orchestration Design: Use in Dicsuss2 (small groups) in pyramid
script a tool with affordances…(IA indicators, SNA, text based with
upload capability etc.)
• Orchestration conducting: A forum selected at design time is not
available. Shall I use a chat with same affordances?
• EEE: A learner with low participation and rating(s) in a Web 2.0
(Moodle forum) is given the role of coordinator with extra
material in next activity of Second Life (SL) and given tools in
another activity of Augmented Reality/Virtuality (in class/SL he is
permitted to use specific material that others in group cannot) -
Kinect use in SL.
• CompleX WSs (=BPEL) and annotated with semantics: Based on
IRMO, input of WS2(IA indicators) is output of WS1, and output of
BPEL complex process is an overall assessment of a learner in
various group activities.
Aristotle University of Thessaloniki, Greece 2011
24. Orchestration Layered model
SMAPIS (LinkedData
paradigm)
Learning Flow (OWL) Meta-Adaptation
Data flow (BPEL, OWL-S)
Adaptation
IRMO, MAPIS
Pedagogy (CLFP, AP) Design – Script
Technology – Communication
Aristotle University of Thessaloniki, Greece 2011
25. Challenges of Ubiquitous…
• 1. The “Accidentally” Smart Environment
• 2. Impromptu (i.e. NO) Interoperability
• 3. No Systems Administrator
• 4. Social Implications of Aware Technologies
• 5. Reliability
– Example: No 3G/wifi connectivity – No problem, use
your phone to take a photo, use Lights to attract
attention
Aristotle University of Thessaloniki, Greece 2011
26. So What’s the Problem?
• BPEL: Description of
how Web Services are
composed. Limitations:
No IOPEs, Allows
execution of a manually
constructed
composition
• UDDI: Directory Service
for Web Services.
Limitations: keyword
searches, limited
capability search
Aristotle University of Thessaloniki, Greece 2011
27. Process Model in OWL-S
• Process
– Potentially interpretable description of service provider’s
behavior
– Specifies service interaction protocol
• Tells service user how and when to interact (read/write messages)
– Specifies abstract messages: ontological type of information
transmitted
• Used for: Service invocation, planning/composition,
interoperation, monitoring
• All processes have: Inputs, outputs, preconditions and effects
• Composite processes deal with:
– Control flow
– Data flow
Aristotle University of Thessaloniki, Greece 2011
28. SWS tasks
Aristotle University of Thessaloniki, Greece 2011
29. Tackling Semantic Interoperability…
Lack of Semantic Interoperability is a major hurdle for
• Discovery: Different terms used for advertisements and requests
• Invocation: Different specs for messages and WS interface
• Understanding: Interpreting the results returned by the Web
service
• Composing Services: Reconciling LD goals with goals of the WS
• Negotiating contracts & communications: Different terminology
and protocols used
Aristotle University of Thessaloniki, Greece 2011