What about Re-usable Adaptive Education

What about Re-usable Adaptive Education

by Sarunas Jampolskis -
Number of replies: 9

The new chalenge in e-learning is Re-usable Adaptive Educational e-Content.

Is it possible to adopt moodle to these needs? approve.gif

There is nothing impossible for moodle? But could we discuss about it, I am interested in participating on doing such work (teoreticaly and programing thoughtful.gif).

At first i thought that to make such adaptive content would be too complicated, but now I think it is possible, for example: student takes an exam at the very begining of theme/week. Then, according to his rezults, moodle presents him a adopted to his knoledge paper (only information he didn't answer in the quiz). So, he doesn't need to reed all the document, he gets just what he needs. I mean all the theoretical and practical things (of the theme/week) here, not just feedback. cool.gif

Average of ratings: -
In reply to Sarunas Jampolskis

Re: What about Re-usable Adaptive Education

by Martin Dougiamas -
Picture of Core developers Picture of Documentation writers Picture of Moodle HQ Picture of Particularly helpful Moodlers Picture of Plugin developers Picture of Testers
OK, well, I don't want to discourage you here, but I'll take this opportunity to spill a few thoughts on your first statement here:
The new challenge in e-learning is Re-usable Adaptive Educational e-Content.

I'm very familiar with this "angle" on online learning (intelligent tutoring and so forth), and while there's a place for it, I think it can be dangerous to get too excited about it.

It's a kind of dream for computer scientists (and I used to be one) that if we just try hard enough we can build a completely automatic teacher, a system that can adapt to a student and give them exactly the information they need. "Why, then we could set up a course for thousands of busy little students, working away at their homework like rats in a maze, and not have to lift a finger!"

It's not really that new - the research goes back for decades.

The trouble is that this technology has a looooong way to go before it can even get close to an ordinary teacher. If you've ever had a good teacher who cared about your learning I think you know what I mean. There's always a slight excitement to receiving personally-directed communications that will be noticeably missing from any automatic spoon-feeding loop.

I would rather see work progressing on two fronts:
  1. improving the teacher's tools to make it as easy and transparent as possible for a teacher to interact with their class and ... well, teach.
  2. improving the student's tools to help give students skills in communication, critical reflection and research, so they are able to decide for themselves what they need and seek it out (like the real world). I don't think there's anything wrong with learning how to navigate content.


Now I realise I'm probably going off on a philosophical tangent here, Sarunas - you just got me started. smile

Back to your idea, one thing you can do already in this vein is to put links to content in the feedback of quiz questions. A modest improvement that many might find useful would be to add an interface to formalise this a bit: a popup menu of course resources, for example, on the question editing page.

Is that the sort of thing you were thinking of? Perhaps you can explain further about the vision you had in mind and how the details might work.
Average of ratings: -
In reply to Martin Dougiamas

Re: What about Re-usable Adaptive Education:AIML?

by Michael Penney -

What about AIML and 'ai' chat bots that use it? To my mind these seem more like natural language search engines than an 'intelligent' tutor. But the beauty seems to be that when you have an instructor with a large body of work, once it is formatted in AIML (or an XML schema that can be converted/translated to AIML), students could then interact with the work using the chatbot?

Of course, the Alice bot and her clones can 'learn' new content in AIML format as well as be edited on the fly by conversing with them. Seems to me this could server a couple purposes in e-edu:

As mentioned above, natural language search/interaction with a large body of work

An AIML bot could act as an initiator of interaction in an chat-room, as well as act to keep the discussion more or less focused on the topic at hand.

Further, the chat logs from such interactions could serve as an ongoing informal assessment tool for the instructor to see where his/her students are at with the curricular material at a given time.

I agree that the idea of an actually concious or 'intelligent' tutor is far off if ever realizable, but it seems to me that the AIML approach is more like distributed 'intelligence', in that the decisions requiring intelligence are made by human beings, while the computer just keeps track of them and responds when a similar question is asked/idea is raised that has been part of the discussion before.

Anyway, interesting to talk about. Some links on AIML for folks that are interested.

 

Average of ratings: -
In reply to Michael Penney

Re: What about Re-usable Adaptive Education:AIML?

by Ger Tielemans -
  • Chess is a little game, played on a 8x8 playfield
  • the game has a small set of rules
  • their are only 6 roles in the game
  • all the chessplayers obey to the rules
  • Until now pure AI-chess did never win the game, only the ugly brute-force machines like Deep Blue..

Now about Life and Life-education:

  • more rules
  • much bigger playfield
  • too many roles in Life
  • lots of players neglect the rules

But thanks, I will study your reference-website with pleasure, hopfully picking up some new ideas..
...if I only had more time.

Average of ratings: -
In reply to Martin Dougiamas

Re: What about Re-usable Adaptive Education

by Sarunas Jampolskis -

I am not trying to dream about super intelligent system that can change teacher dead.gif.

The problem is: when I want to build a very good course, I have to put a lot of information in it. And if there is a lot of information, students dont like it (it could even frighten them).

The salvation could be a course (or even a theme) starting quiz, where system could make a research what things do respondent need to learn. I am trying to make a course of 30 themes and every theme divide into very small parts - could be feedback of the every quiz question (every small part of a course refer to one or more questions).

 

I am thinking, it is possible biggrin.gif:

  1. make a usual course.
  2. divide every course theme into very small portions (and when you try to think questions, they usualy refers very stictly to one or other portion of material).
  3. make quizes with refering questions.
  4. it could be a feature that joins these all small portions into biger groups (standart reading)
  5. after the quiz, this feature could generate a theme of the needed (wrong answers, less than 70% or something...) to learn material acording to quiz results.

And you still need teacher!!! blush.gif Just the material is more adopted to students needs and student could learn faster.

Average of ratings: -
In reply to Sarunas Jampolskis

Re: What about Re-usable Adaptive Education

by Patrik Nilsson -

What you say sounds interesting and I think I understand.
We call it validation here. Its so the students can build on to there allready learned skills without having to read again about what they allready knows. The student might not have grades from an education but have life expirience from a work. And then you use a test to validate that knowlige and then make a course based on what the person doesn't know and needs to add on to there knowlige to pass the course. It would still be the same with a teacher and all but some students needs only to study for a week. while others needs to study 10 weeks.
Right ?

Average of ratings: Useful (1)
In reply to Sarunas Jampolskis

Re: What about Re-usable Adaptive Education

by Ger Tielemans -

There are some places where AI is very succesful. To give an example: A Dutch professor teaches informatics to law-schools-studentS in Rotterdam. He let students write an essay about a subject.

Students give their assignments to a computer and get automatic a grade. (You must train the machine for every separate subject, by feeding it a sorted list of 20 (?) essays about that specific subject. The machine also detects if students copy each others work.

(I doubt if it is useful for doing a more closed exercise like making a translation . I will do several tests next year with this tool.)

I like your spirit and your belief in the instructional design tradition. Would be great to have a resource center where students can do exercises after they fail a quick-scan-test on that subject: debriefing of the test should have links to resources in your Moodl as Martin proposes. With this approach you work on knowledge reproduction level.  I like your approach of first the test and then reading the resources. The more traditional approach is first the reading then the test..

Next problem is to bring students in learning situations where they can transfer this new knowledge to slightly different situations. Problem centered approaches in combination with groupwork are good environments for that... and there Moodle comes in with my favorit: the personal journal option. 

Average of ratings: -
In reply to Martin Dougiamas

Re: What about Re-usable Adaptive Education

by Richard Treves -

The trouble is that this technology has a looooong way to go before it can even get close to an ordinary teacher. If you've ever had a good teacher who cared about your learning I think you know what I mean. There's always a slight excitement to receiving personally-directed communications that will be noticeably missing from any automatic spoon-feeding loop.

I have heard this situation being referred to as the '2 sigma problem', that is, if you train someone to do a task one on one their performance will be on average 2 standard deviations (of a class of learniners) better than if you did it one on many.  The goal of CBL is therefore to produce teaching communication software so sophisticated it passes the Turing test (people cannot distinguish it from a person) at which point your improvement in training goes up '2 sigma'.  But that assumes students are motivated to do well on a piece of software.

I would go further than Martin here in the importance of the person in the system, IMHO it is a central human feature to wish to impress people, be that a tutor or another student.  Advanced CBL will only work as a carrot if a tutor is following your progress or you can turn to a fellow student and compare scores.  Or if your boss sees your grade and gives you a raise wink (which is training not higher or school education).

Of course it is entirely possible to teach yourself with no interaction with other people by just using a good text book.  IMHO any teaching technique that doesnt involve people is JUST a glorified text book.

This is not a chriticism of RuAE, I just like the topic area and writing or talking about it is my learning style smilesmilesmilesmile

Average of ratings: -
In reply to Sarunas Jampolskis

Re: What about Re-usable Adaptive Education

by Timothy Takemoto -

I am really interesting in adaptive education too. There is a recently a big wave of this in Computer Aided Language Learning with increased interest in (really dull?) learning techniques such as the use of flash cards and flash card software such asthis very highly rated piece of equipment.

In a sense flash card packs are the easiest version of AE in that one can separate the flash cards into two piles - ones you got right and ones you got wrong - put the ones you got right to one side, shuffle the ones that you got wrong and then take the ones that you got wrong again. This is what VTRAIN does (above) and it is said to work very well.

Better still, and this might one day be possible, since questions could be grouped into categories, it would be GREAT if, when a student gets a question in a particular category wrong, they would be faced with more and more questions in that category.

I am not sure how the links idea would work that Martin suggests below. Have you tried implementing it or any other adaptive method?

Communication is great. But try getting a lot of Japanese to communicate. I try all the time. But sometimes others tools help. Horses for courses.

MFFITOP

Tim

 

Average of ratings: -