Machine Learning

I am still thinking what it means when learning of subjects such as Mathematics, Algebra etc. is based on machines – the machine learning as elaborated by David Hu here.

It seems that the machine learning software could capture all the “learning” and “assessment” progress and results for any students who logged into the system.

Relating to the learning analytics that is available in analysing the individual student’s progress and units completed, I am thinking about:

(a) To what extent would such learning analytics be useful in monitoring the student’s progress?

(b) What about student’s confidentiality and privacy in learning?  Would there be a need of consent statement given by the students for the disclosing of such data to any one involved (including the teacher) in the education process?

Finally, would we want our students be taught and graded by machines?  What are the implications if such machine learning is used to replace teachers’ teaching in K-12?


In MOOCs, more is less, and less is more (Part 2)

In this part 2, I would focus on x MOOC.

More is less in Instructivist x MOOCs:

The more courses that are offered through the x MOOCs, the less difference there are between the formal courses offered by the universities and those by the x MOOCS.  However, there are then more paradoxes surfacing from such x MOOCs, in terms of the measurement and assessment of mastery used by the universities.  In this post on contradiction-facing-moocs-and-their-university-sponsors:

Subject mastery in a MOOC environment may be a necessary but not yet sufficient condition for “mastery,” at least in certain galaxies of higher education. In fact, perhaps the mastery we are ultimately hoping for from the range of galaxies in the higher education universe is more than the ability to answer 50 questions correctly. Instead, our ultimate goal is to develop a capacity to convert the implications of those answers to new questions, new ideas, and new inventions — dynamic sources of impact. Developing and supporting this dynamic capacity may not scale in the same way that MOOC education can. Read more: Inside Higher Ed

This prompted me to ask: Are subject mastery in a x MOOC not yet sufficient condition for mastery?  If that is the case, more mastery in x MOOC may not be equivalent to a degree obtained from a University as mentioned above. Indeed, as mentioned here on what-moocs-are-missing-truly-transform-higher-education:

A college education is much more than mere knowledge transfer. It is a rite of passage and an important part of personal development and the maturation process. As universities work to assure that result, online courses will no doubt be part of the mix. Read more: Inside Higher Ed

In nearly all xMOOCs, the caveat is: Adopt a push approach in education, as education is based on institutions’ provision.  “Sounds true.” In this interview with Anant Agarwal:

I think that online learning can really transform education in quality, scale and efficiency. Learning can be much better. Students have told us they’ve been much more engaged, they could learn at their own pace, and overall they felt the experience was much higher quality than a regular bricks-and-mortar class. In scale, we were able to support 154,000 students from around the world with online technologies. Finally, in efficiency, we were able to have the same staff support a large number of students, so learning can become very efficient.

x MOOC is about learning efficiency, with more output (number of students who enrolled into the course), and less or same input (with same or reduced number of staff or resources).  So, education could also be delivered to any one in the world if they could access it, or any corner in the world, for those who need it.

What do those blended models look like?  As Professor Anant Agarwal says:

In a blended model, you do what is called flipping the classroom. Flipping involves having students do video [lecture] sequences and some concept exercises at home before they come to class. Then, you can ask the students to come into the class for interactive sessions where they can sit down and have discussions, ask questions, do interactive laboratories, solve problems. It kind of reverses what is done today and it can be very effective.

x MOOCs are based on the flipping the classroom model:

xMOOC is based on the teaching model where the teacher teaches, and the students learn and consume the knowledge from the course, like watching the videos, or reading a book, an artifact, and be assessed on what has been taught or covered in the videos.  The main differences between off-line and online approaches seem to lie with the machine grading and feedback, in the responses to computer generated quizzes or test, and that the students would respond and repeat the learning until they have achieved content mastery.   That is STILL based on the instructivist approach – which is based on behavioral/cognitivist learning theory, where the learners master the content, probably with the transfer of knowledge from one person or a number of persons (the professor(s)) or the machines (robot or virtual teacher), or information source to that of the learner.

If there are more x MOOCs, would less teachers be needed?

It seems that we could reduce the number of teachers required when more x MOOCs are offered, as these courses would become a duplication of the mainstream courses in lots of universities.  Why not asking students to enroll into the x MOOCs, and so we would not need so many teachers to run the SAME courses in local universities or institutions?

Instead, the local universities could focus their efforts to run more “assessment based tests or examinations” to validate the students who have passed the x MOOCs.  Is it a more economical, effective and efficient way to offer education? The more connectivity there are with the x MOOCs, the less formal teaching is required for local institutions, as such teaching could be done by the professors in x MOOCs. Is that a paradox or ideology?

If there are  more MOOCs as alternative education and learning platform, would there be less number of institutions needed to run the courses?

x MOOCs provide more platforms for any one who would like to learn in particular courses, so far if they could master the content of the course.  Does it mean less institutions would be needed for the running of the SAME courses?

Would x MOOCs take down branch campus?

In this “Will MOOCs take down branch campus – we don’t think so here. Designing New Learning Environment is another course from Stanford University.  Here the questions look familiar, and I think could be interesting to explore.  Into the future with MOOC gives you more choices, though you have to think about what it means to be educated.

Is MOOC massive-open-online-course-a-threat-or-opportunity-to-universities? Let’s see (my comments in italics):

  1. Extend:  The MOOCs are “disruptive extensions” rather than threats to traditional universities.  MOOCs will not replace the existing campus-based education model, but represent a huge opportunity to create a completely new, and much larger market that universities couldn’t serve previously due to the physical limitations of the campus.  Yes, that is the reality now, but would that become a threat to other universities?  It’s too early to know.
  2. Capture Value:  Different business models will emerge for MOOCs in the future. The unbundling of learning, credentials, social interaction, facility access, and assessment will make it possible for institutions to establish new business models.  This would also challenge the ‘traditional model of education’, and so not every institution would “survive” or cope with the “revolutionary” business model.  Those who have difficulties in embracing the new models would likely “fail”.
  3. Continue:   Professional education will become even more critical in the future, and MOOCs provide a great means.   Knowledge and even some skills have a shorter shelf life these days.  I think this depends on what sort of education that we are looking for.  See my previous post relating to c MOOCs.
  4. Customize:  MOOCs enable you to optimize the mix various educational offerings irrespective of location, time, age, and provider.  For example, it makes more sense to combine a German language program (delivered by a teacher in Germany) and American history program (delivered by an expert in the USA). Sounds good to have such varieties of education programs.  More customization would lead to better learning, in theory.
  5. Be There:  openHPI is planning to turn their virtual community into a real-world interaction experience.  Learners will pair up locally in coffee shops or conference rooms, study the online courses together and have face-to-face discussions. Hopefully, that is what MOOC is all about, opening doors to different learners, and to different parts of the world.

Now, even the small OOC is emerging, as posted here, and a start up of A Revolution University here tyler-cowen-does-a-mooc.

Is MOOC inside or outside the box?

Finally, is xMOOC one of the most discussed topics so far in 2012? I reckon yes, as evidenced in post like this.  Such conversations and opinions are the heart of MOOCs.   I will continue to explore the more is less and less is more in Part 3.

Postscript: Refer to this post moocs-could-hurt-smaller-and-for-profit-colleges-moodys-report-says