Thursday, July 23, 2009

Full Immersion 2025 - A vision of the future

Wow.

A casual Diigo link to the Colorado Virtual Academy this afternoon had me interested. I'm always on the lookout for virtual learning environments that match my ideas of the way to the future, and this link just kept ticking the boxes.

First off I noticed "An individualized approach to learning". Now that term has been bandied about a lot recently without really any substance to it, but paired with a virtual environment it is possible and a main goal of my VR dream .

A bit further down the page and another cathphrase pricked my ears: "An ideal fit for accelerated learners". Having completed an assignment last term on how to manage gifted and talented students in the current curiculum and also the implications of accelerated learning in a VR environment, it seemed likely that we might be on the same wavelength.

Digging into the accelerated learners section confirmed my suspicion: "The K-8 curriculum is mastery-based, enabling students to advance when they're ready—not when the rest of the class is." Wow. First time I've seen that type of thought expressely used in an established learning institution. I'm excited!

Looking deeper into the K-8 curriculum I find another component of my model: "And with more than 700 lessons per subject, he can dive deeply into areas of interest." Yes! A curriculum exploded into bite size chunks. No real indication of a meshed list of prerequisite skills, but nothing to the contrary either.
Opening up the individual courses it starts to sound more like an ordinary school approach. I got excited that "From helping younger students make the link between the concrete and the abstract to introducing older students to Algebra" might mean having an intrinsic peer mentoring approach, but upon reading it again it's really just showing the breadth of the whole course.

A phrase appeared a couple of times that didn't really make much sense: "Big Ideas + Consecutive Down Payments + Practice = Mastery". Found out a little later on that the downpayments are the stepping stones to the big ideas. Makes more sense now and shows the benifits of an adaptive curriculum.

Through tea I was really excited on how close it matched my model. The closest by far. I dug up a blog of Bror Saxberg for my google reader and the first article caught my eye: "A vision of the future", where he expands on his participation in a published article on how 10-year olds will learn in 2025. This article is amazing! It's like reading something I would have written myself. Our visions of a future educational environment through VR are almost identical!

I really like the implementation of the helper AI (yoda). I'd dumped most of the AI elements from my model as I wanted it to be as believable as possible to the Uni lecturers I was presenting the model to without having to explain the intricacies of tracking AIs, shadow AIs, Tutor AIs and helper AIs (which was a big part of the vision in my early days; I guess that's because AI was a big hobby for me back then). I still think that a yoda helper AI is still further off than 2025, but then again it depends on the scope of something like Project Natal. Having AI's able to learn standard answers from real teachers and tutors would then be able to fill in forums / FAQ sessions in the absence of a real tutor. I had always envisaged helper AIs saying things like "You answer is most likely ..." to give the relative confidence to the learner on the veracity of the answer given, and being able to show the pieces of information that made that assumption if questioned.

For me the big part for AI's to play is through "shadowing". For every student you have a shadow AI that attempts to emulate the thought processes of the student. It tries to predict what questions the student will get right, what they get wrong, etc. With those "aha" moments in learning where the patterns and processes finally click into place, there will be a big positive spike in the error between the shadow AI and the student. This can be used to trigger the next phase of learning. The shadow AI can also be used for system wide testing, where it can be interrogated and compared to other student's AIs for help in choosing the next course to take. Bror's vision seems very close, but relying more on the student's actual results to make the system learn how to educate that student better.

The role of educators in both systems seems very much inline. I think that through the Uni course I'm concentrating more on allowing different learning styles to be catered for through the variety of courses on offer to achieve the same skillset, whereas Bror focuses more on deep database analysis.

Deep database analysis is one area where Bror's model seems more advanced than what I had imagined. His use of data mining techniques to shape the curriculum seems more solid than the shadow AI approach. Both systems would compliment each other though; shadow AIs would just give more data for analysis.

So pleased that I found this link and Bror's ideas. I'm over the moon that someone has demonstrated what I'd been contemplating for a long time. Well, not everything's implemented yet, but with someone at the helm with such a closely aligned vision I can see exactly where the boat is going...
I want on ...