Presentation of Rebo, the reflection guidance chatbot. Evaluation of agent and planned incorporation of new features: active help seeking & reflection on pre-defined learning goals.
Collaboration with training workshop
Upload of practical work tasks -> trigger reflection
Finding learner identity
Incorporating reflection as deliberate learning strategy
Lifelong professional learning
5-13 interactions per apprentice (8)
8 interactions each (4-11)
Damerau-Levenshtein string distance
RJ: when there is a problem: how could an agent fix this, assure/improve reflection on given level?
Would’ve expected more impact of motivational factors:
Workshop 1: highly motivated, face2face, first week of training
Workshop 2: lockdown light, online
Workshop 3 hard lockdown, online, facemasks, right before holidays
Pretty strong statement…
RJ interactions: picked 8 comparable ones: same task, same day
Ad fake: precision and other keywords; dictionaries
Ratte tool, university of Regensburg
Amel will tell more about implementation
Tested work in progress in various workshops
2 field studies with apprentices in years 2&3, 4 weeks, 15 apprentices
Field study 1: 73 interactions (5 each)
Now: 12 weeks, 17 apprentices in year 1
2 examples with neg. summative comments and next interaction was a lot more engaged (not done coding yet)
Make apprentices think about goals, relate their daily working and learning experience with goals of their training
Apprentice sees legal basis of their education, (trainers and apprentices have common goal…)
Soft skills could be everywhere