- Stefan continued his presentation starting from experiment 4 [ppt]
- Some observations:
- Bakhtin’s genres might be useful to label certain observed actions in
- As it is pointed in Fatos’s statistical analysis as well, there seems to be an inverse relationship between the amount of social discussion and problem solving discussion.
- Roughly speaking there are 3 types of groups we observed so far:
- 1: One person knows how to do something, and leads the group accordingly.
- 2: Nobody knows what to do, they wander around together.
- 3: Everyone has one piece of the solution somehow, and they try to put together pieces.
- A relevant question at this moment is to see how we should systemize moderation given these types of groups. Each group would need different sort of moderation.
- Could it be possible to have a Vygotskian style moderation based on what the group has done so far, e.g. here is something you should think about doing next (content based moderation).
- Process intervention is another way of moderation (i.e. moderating the group without commenting on the content of discussion, but rather commenting on the process of collaboration)
- We chose to structure the process of interaction instead of providing intelligent tutoring based on content
- A discussion about learning followed that:
- Learning is a socially oriented process. Classical school setting sort of counters this notion of learning. We are interested in finding ways for favoring social aspects (by using computer support).
- What is learning?
- The concept of learning will hopefully emerge from our analysis
- Collaborative learning is different from the common notion of individual learning. Learning happens in brief moments during interaction through a process called building knowing, or meaning making.
- Some of the things that needs to be supported in the system: the ability to construct meaningful text and artifacts in response to the problems we give them to stimulate their thinking.
- What individuals take away is less important, what is being produced together is more important
- The main question: Can we support those rare moments to occur more frequently?
- Learning math is about being able to articulate insights; that’s what usually happens during chat
- There is an important difference between the moment of production of meaning and location of where it gets produced.
- Does that happen publicly, or does that happen in the heads of individuals?
- Where knowledge is located: ideas from individual cognition and group (or distributed) cognition will be relevant for a discussion about this topic.
- Speech genres
- Speech genres can be used at multiple levels
- Exposition vs exploration can be two genres to consider as a starting point
- Methods people use to accomplish different things can be labeled with genres
- e.g. particular ways of saying things; the ways native speakers say things on certain occasions.
- Can we recognize these in terms of genres?
- One way to find out could be browsing through all our data to find all examples of a particular genre, and see whether there are standard ways of saying things
- Try to define negotiation, misunderstanding moments algorithmically. Them let the computer go through and identify some candidate moments, as an index
- How to segment the chat into phases
- Operationalize different ways people mark the transition moments, then this can be computerized as well
- This sort of analysis is more feasible as opposed to more ambitious natural language understanding methods.
- Moderation will be automated in the future. These tools will be very useful for designing good agents. Methods for giving good feedback to the members of a group about their performance is currently a hot topic.
- Software design ideas
- Murat [ppt]
- Nan [ppt]
- We should organize a local workshop for discussing software design ideas
- Dr. Math extension
- The idea is to interact with a student to see his current level and advice accordingly
- IRB concerns (if we want to do it as part of VMT)
- Try doing things collaboratively by encouraging kids to help each other
![[Home]](/gifs/MFlisting.gif)