Tuesday, April 07, 2020

Ng & Koller – Academic and corporate MOOCs…

Ng and Koller were two Stanford academics who set up Coursera. They claim that Coursera could never have happened within Stanford, as it needed to be an independent company to raise capital and focus on its MOOC mission. Their role in the development of an early large MOOC provider, as well as their research on what we can learn from MOOCs, has made them significant players in online learning.

MOOC phenomenon

The MOOC acronym was created by Dave Cormier in 2008, a label he attached to a specific source, the course Connectivism and Connective Knowledge by Stephen Downes and George Siemens. This was described by Siemens and Downes as a connectivist MOOC. We then had a hiatus. Thereafter the MOOC phenomenon was massively amplified by the Khan Academy, who was using his method in 2004, but quit his job in 2009 to get the Khan Academy going, with its millions of video views and free courses. Sebastian Thrun (who publicly acknowledged his debt to Khan) and Peter Norvig of Google, were then significant, as their AI course, in the fall of 2011, caught the imagination and boosted the phenomenon even further. Stanford University also played an important role, as it produced a number of courses and entrepreneurs, especially Ng and Koller.


Ng is an AI specialist in machine learning and deep learning, using neural networks trained using deep learningalgorithms. Koller’s area of research is AI in robotics biomedicine. They launched Coursera, a for-profit company,  in 2012. Coursera partners with major Universities to deliver courses across a range of subjects, providing a consistent platform for the delivery of online courses, with video, interactions, various forms of assessment and social communication and learning.
It has increased its courses to include what it calls ‘Specializations’. These are bundles of courses that teach skills in a specific subject. They also deliver degrees and courses for workforce training, for businesses and government organizations.
From 2017 Coursera offered Master's degrees, first in Innovation and Entrepreneurship from HEC Paris and Accounting from the University of Illinois. They have added Computer Science in Data Science and Business Administration, from the University of Illinois. Many other degrees have been added since.
In Coursera MOOCs, there are various levels of assessment. In addition to multiple-choice questions, there is the automatic grading of short open answer input, math expressions, models, even software assignments (seeing real outputs). Ng and Koller claim that peer grading correlates well with teacher grades. They also claim that self-grading correlates even better. 
Their model is to partner with Universities and other organsations across the world to deliver high-level courses. They currently have over 200 partners in over 50 countries and over 4000 courses.

MOOCs influence

MOOCs have been criticized for high levels of drop out and the fact that it is mostly well-educated graduates taking the courses. Others counter that to compare MOOCs to traditional 18-year-old undergraduate courses is a category mistake and that drop-out is less important than the huge number of people that drop-in. It is also natural, that in the early stages of MOOC marketing, that the early adopters would be the well-educated, as they are the group that know about the phenomenon and want to try it out. MOOC providers have started to move away from purely academic learning, with a move towards practical courses on IT, business, health and other vocational subjects.
MOOCs can be taken on their own but they can also be seen as part of university courses. However, their lasting influence may be the way they made every higher education institution at least consider the strategic option of online learning. They have also furthered research into what works online, in terms of video length, content, assessment and social components.

MOOC movement

Ng and Koller see MOOCs as serving a role in overcoming scarcity of higher education in the developing world but even in the developed world the rising costs of tuition is making education unaffordable. Andrew Ng teaches his machine learning course to 400 students at Stanford, it would have taken him 250 years to teach the numbers that turned up to his machine learning MOOC. 
It is this combination of the possible with the necessary that, she thinks, has made MOOCs fly. They broke the back of the one hour lecture, provide alternative resources and avoid the one-size-fits-all model. Around all courses, communities form, within forums, on social media but they also self-assemble with meet-ups. Personalisation is also important. She quotes the Bloom 2 Sigma paper, and claims that MOOCs are now pushing towards personalized delivery. MOOCs are also being produced in many languages, moving beyond the institutional model, which is that everything is taught in one language.

Research in learning

The data collected is a huge research opportunity, giving data opportunities way beyond any individual institution. Poor questions and weaknesses in course design can be identified, optimal video lengths and other optimal instructional and support techniques can be studied and finessed. There has been a wave of innovation within MOOCs as well as beyond MOOCs to online learning in general.


Ng and Koller’s Coursera was neither the first, nor the only, MOOC provider but it was hugely significant in expanding the movement and delivering the majority of MOOCs worldwide. They were also significant in exploring the monetization of MOOCs through recruitment, certification and other sources. Ng and Koller are part of a new breed of educationalists that come from AI backgrounds. As part of the revival of interest in AI, they turned their attention to education, with notable success and influence.

No comments: