The document discusses FROG, a tool for creating rich collaborative learning experiences. FROG allows for embedding configurable collaborative activities that can have complex social structures and data flows between activities. It includes operators that can generate groups, redistribute data, and adapt activities in real-time. This enables live orchestration and analytics dashboards for teachers. The tool is open source and can run collaborative concept mapping, video, and other activities. The presenter invites interested parties to experiment with FROG.

1. Photo by Bill Gracey @ Flickr
FROG:
Embeddable tools for rich collaborative learning
Stian Håklev, CHILI Lab, EPFL
#VFH18, June 14, 2018

3. Key ideas for learning
Collaborative and individual
Synchronous and asynchronous
Rich interactions
With rich content/creating rich content
Multiple representations
Pedagogically/socially useful interfaces
Interaction scripts
Purposeful grouping of students
Teacher orchestration

4. http://axelkrommer.com

5. Thanks to Matthias Giger in Romanshorn for the recommendation

6. Google Form
Etherpad
Confluence Wiki
PhD thesis (2016): From Seminar to Lecture to MOOC:
Scripting and Orchestration at Scale

7. Design Study 2:
MOOC
6 week course on
inquiry and technology
for in-service teachers
Weekly themes:
• Inquiry and student-centred
pedagogy
• Designing inquiry activities
and assessments
• Collaborative learning
• Handheld/mobile devices
• Knowledge co-construction
and student-contributed
content
• Inquiry enactment

8. Crowdsourcing
teaching resources
Taking advantage of
existing domain
knowledge
!8

9. Crowdsourcing
teaching resources
Taking advantage of
existing domain
knowledge
!8

12. Collaborative Workbench

13. Collaborative Workbench

14. Collaborative Workbench

15. Collaborative Workbench

16. Problems
• All technology is custom written for this MOOC,
very hard to reuse, very high barrier for groups that
want to experiment with rich collaborative learning
scenarios (many other possible ways of doing it)
• Captured data much richer than traditional MOOC
data (because space of possible activities much
larger), but no easy way of connecting learning
data back to pedagogical scripts, activities etc.

17. Class
Team
Individual
Write summary
Debriefing lecture
Argumentation
Discuss the map
a1
a2
a3
a4
a5
Questionnaire
Orchestration Graphs

18. Class
Team
Individual
Write summary
Debriefing lecture
Argumentation
Discuss the map
a1
a2
a3
a4
a5
Questionnaire
Orchestration Graphs

19. Class
Team
Individual a1
Questionnaire
Discuss the map
a2
Orchestration Graphs

20. Class
Team
Individual a1
Questionnaire
Discuss the map
Argumentation
a3
a2
Orchestration Graphs

21. Class
Team
Individual
Debriefing lecture
Argumentation
Discuss the map
a1
a2
a3
a4
Questionnaire
Orchestration Graphs

22. Class
Team
Individual
Write summary
Debriefing lecture
Argumentation
Discuss the map
a1
a2
a3
a4
a5
Questionnaire
Operator1
Operator2
Operator4
Operator3
Operator5
Aggregation Distribution Social BackOffice
(A) Listing (D) Broadcasting (S) Group formation (B) Grading
(A) Classifying (D) User selection (S) Class split (B) Feedback
(A) Sorting (D) Sampling (S) Role assignment (B) Anti-plagiarism
(A) Synthesizing (D) Splitting (S) Role rotation (B) Rendering
(A) Visualizing (D) Conflicting (S) Group rotation (B) Translating
(D) Adapting (S) Dropout management (B) Summarizing
(S) Anonymization (B) Converting
(B) Updating
Workflow Operators

23. What policies should cities adopt
towards Uber? A jigsaw collaborative
learning script
In this example scenario, nine students engage in
an exploratory discussion around the policy issues
faced by cities in the new economy. The goal is
for students to get exposed to a wide variety of
arguments and conflicting interests, and develop
critical thinking, argumentation, synthesis, and
creativity.
1. An operator (o1) takes the class list, and
generates groups of 3, distributing expert
roles among the students (taxi drivers,
policy experts, consumer advocates)
2. Experts (e.g. all taxi drivers) read an article
related to their expertise and discuss
relevant ideas (a1)
3. Mixed groups bring their expertise together
and brainstorm problems, ranking them (a2)
4. An operator (o2) aggregates the problems
from the different groups and sends it to a3
5. The whole class sees the top problems, and
collaboratively sort them into four different
categories (by clicking and dragging) (a3)
6. Mixed groups try to come up with solutions
to the problems in a3 (a4)
7. An operator (o3) creates a list of all the
suggested solutions, and creates a list of the
highest ranked ones
8. The list of solutions is displayed, and a class
discussion follows (a5)

26. • Configurable, collaborative (live-synced) activities
• Complex social structure generated at runtime
• Flow of data between different groupings, and types of activities
• Semantically meaningful dashboards for different activity types
• Live orchestration actions (changing activities, pausing)
What you just saw

27. • Demo

28. Social operator
generates or changes social structures (roles, groups)

29. Product operator
generates or changes activity data
Aggregation
Transformation
Distribution
External input
Redistribute Change formatFilter

30. Product operator
generates or changes activity data
Aggregation
Transformation
Distribution
External input
Combining operators
Redistribute Change formatFilter

31. Bridging research and practice
Ishari Amarasinghe, 2016

32. Algorithms for inference

33. Algorithms for inference
visualization/dashboard

34. Algorithms for inference
visualization/dashboard
redistributing student data

35. Algorithms for inference
visualization/dashboard
redistributing student data
regrouping students

36. Algorithms for inference
visualization/dashboard
redistributing student data
regrouping students
adapting activities per
student/group

41. Interoperability?

42. Concept map
activity
data in
concepts from a previous
activity (a form), or even an
operator (text mining
concepts from a text
students wrote)
streaming learning
analytics
teacher can monitor learning
as it happens with
dashboard, etc.
orchestration
teacher can intervene while
activity is running (pausing,
modifying groups, skipping
to next activity ahead of time)
data out
the resulting concept map
in a format that can be
reused/analyzed by
other activities/operators
(json)
social/collaborative
collaborative activities with
complex social structures
configurable
flexible activities which can
easily be configured by
instructor before the class.
unified configuration
interface
Rich embeddable activities

43. xAPI (TinCan API)
Common recipes for video, quiz,
forum, self-regulated learning, etc.

44. Open source
Pluggable content types
Easy to author
Easy to embed

45. Open source
Pluggable content types
Easy to author
Easy to embed
Streaming xAPI statements
to browser

47. Graasp, platform for inquiry learning and knowledge sharing
Part of large EU projects (Go-Lab, NextLab, GoGa)

53. Invitation
FROG is open source, and can be downloaded and run in 10 minutes
(but not ready for prime time yet)
Under heavy development, interface, usability, stability will much improve
Interested in doing experiments with us?
• Experimenting with integrating activities or algorithms?
• Running FROG graphs in university classrooms/virtual courses?
• FROG templates / pre-loaded graphs with Open Educational Resources?
• Workshop at #oercamp18 in Hamburg
Stian Håklev, stian.haklev@epfl.ch
https://github.com/chili-epfl/FROG