08 Apr 2010
Image via Wikipedia
What happens when the Harvard Graduate School of Education and the Harvard Business School get together 150 people for 2.5 days to listen to 53 presentations from the folks who are working to reinvent global education in the context of 'Preparing Children and Youth for an Interdependent World'? The product of this high quality idea churn might take a few years to come down the pike, but the near term result of two days in conversation with this group has me feeling inspired, awed, and...well...educated.
I was honored to have the opportunity to sit on a panel entitled 'Promoting Tolerance and Understanding Around the World.' The panel was chaired by the great Henry Chow, former CEO of IBM China and a current Advanced Leadership Fellow at Harvard (who spoke about his interest in innovating new approaches to traditional Chinese medicine with roughly the same enthusiasm and energy you would expect from a 23 year old entrepreneur).
In terms of Meedan's contribution's to the event, I was--as I ever am--the one to stand up and talk about language as the inclusion/access to knowledge issue. I then bridged to discussion of social networks as facilitating, beyond access to knowledge, 'access to each other.' This is the promise of global education, namely, the ability for a diverse network of teachers and learners across the world to enable access to understanding.
Here are some notes from my talk- notes which were not read, but perhaps might resemble in some rough form some of what I actually said.
"It is an honor to be here for this gathering and I thank Fernando for inviting me to join
Today I would like to discuss the tangible, technical obstacles to global distributed learning, particularly those obstacles related to linguistic boundaries. I would also like to talk about changes that we see in the social web, implications that are fundamental not only to the range/scope of education and educational collaborations, but also to the form of these collaborations; to say, the evolution in the way that we create and share knowledge.
The observations I will share have been gained over the course of a five year effort to develop the Meedan project. Meedan is the Arabic word for a town square, with Meedan.net we are enabling virtual town squares for cross language Arabic-English social networking and knowledge exchange.
Taking a lead from Peter Copen's remarks yesterday--we do have a vision for Meedan, it is a big, irrational vision - and irrational completely; in the sense of being a vision that is technically, organizationally and culturally impossible. We only exist today because we have chosen to meet the charge with an equally irrational persistence.
The vision for Meedan from the start was the vision for social network driven translation communities where the common interest in access to information in a journalistic or educational setting would incentivize these groups to collaborate via the web to augment machine translation (MT) processing. We realized that MT was generally quite poor (in those days it was plain terrible), but supposed that the data we generated would create a powerful feedback loop to improve the quality of our MT engine.
In our first implementation, a media sharing site, news.meedan.net, this network of translators works concert with a distributed community of young journalists to weave together a global conversation that represents, and sometimes bridges, the many many narratives that generally surround a breaking news story.
I will briefly discuss some of the challenging elements of a cross-language exchange. I am known as one of the people who speaks with reporters about the real world application of Machine Translation technologies. When the first circuit board was etched and engineers began to pass complex queries off to uncomplaining machines, human quality automated translation was said to be near-term achievable. My organization has been engaged with Salim Roukos and his team at IBM research working on Arabic/English MT for the past five years. From the outset of our engagement with IBM we forwarded a wikipedia model for translation, allowing a distributed network of humans to collaborate, improving and revising the MT output. With this model we have been able to modestly scale the world's first truly bi-lingual media sharing site. Our translators process about 300,000 words per month through our system (data that is used, by the way to improve the system performance). Translation itself becomes a social pursuit on Meedan. While anecdotal estimates right now have MT and translation memory systems bringing about 20-30% saving in translation efficiency, the fact that we are creating a feedback loop has me reasonably optimistic that we will see disruptive progress in MT over the coming five to ten years.
We are taking this model also into interfaith scholarship, with a pilot project under development for Cambridge University and Al Azhar to network Arabic and English speaking religious scholars via distributed online professional translators. This inter-faith project will provide bi-lingual evidence that dialogue and scholarship can happen across language communities. Due to the sensitivity of these dialogues, we are using trained and vetted human translators to process these scholar's dialogues.
Meedan for Education
What does the social web and translation technology mean for global educational engagement. A couple trends we should keep in mind:
The transition from the information age to the annotation age. In the evolution of catchphrases, we move from 'content (and in the educational domain substitute 'curricula') is king' to 'conversation is king.' This is not the end of the classics and the fall of the empire of formal education, rather it is the ascendence to publishing of the real spark of learning, namely, the interaction that has happened in every great classroom since we started writing on cave walls. To say, the technologies that are redefining how we interact on the web seed forms of education that are more social, conversational, and creative.
The blurring of boundaries between formal and informal learning networks. Social networking enables persistent connections. The promise of global education is not simply enabling what we have called 'access to understanding' - the more profound pedagogical fall-out is the opportunity for the student in such settings to trade roles as the de facto teacher.
So, what does the platform for Global collaborative education look like? There is an opportunity to provide bundled social learning solutions to groups like iEARN (whose Ed Gragert is here in today) and we hope to engage a discussion with a network of private, non-governmental, and public sector stakeholders to explore and pilot these technologies in the context of classroom to classroom and teacher to teacher engagement. Building a globally networked learning community seems to us to be a fairly important piece of the broader work of wiring the world for understanding and tolerance. We at Meedan are pleased to be a part of the global conversation about the technology that might support this vision."