Mar

22

The Atomization of Education

Filed Under Big Data, Blog, Cengage, eBook, Education, eLearning, Industry Analysis, internet, news media, Pearson, Publishing, Uncategorized, Workflow | Leave a Comment

When lofty, intellectual figures like Mike Shatzkin (http://www.idealog.com/blog/atomization-publishing-as-a-function-rather-than-an-industry/) quote one’s sayings of many years ago there is real danger of a sudden rush of blood to the head. Thankfully, his purpose is more to mark the entry of Google as a publisher than to say “Worlock told you so”, but he really started me rethinking this turf in the process. You see, “atomization” is the growing experience of all markets. Nature, this week, described the work of a computational biologist at a drug company, searching 23,000 articles in one text mining enquiry” to pick out hundreds of proteins that could relieve a mouse model of multiple sclerosis”. This is not the world of articles and journals in which most journal publishers think they are still living. The practice of law is a workflow now designed around procedural requirements fed by precedents. Readers here will have heard almost ad nauseam of the collapse of newspaper and magazine business models, replaced online and in the device in your hand with filtered references and the ability to call for more. Artificial intelligence and computer written information ( Narrative Science) will take an ever larger role in content creation, and this content will be increasingly read by machines that protect us from the full onslaught of the information-based networked society that we have created. And that machine-generated information will be created in atoms from the very beginning. Addressable, metadata-identifiable atoms. We will collect them, review them, put them into different orders and create, from these objects, the information structures of the future – and some of them we shall probably call “books” just because we cannot think of a better name.

Amongst my reading this week has been a pair of reports from Eduventures (www.eduventures.com) on Predictive Analytics and Adaptive Learning. While they do not take us very far they do very adequately describe the present and I am delighted to see subjects like this being covered in contexts where educational administrators may get to read them. The educational mold probably broke about a decade ago, and much of what now happens in education reflects the dim and distant echoes of the works of true scholars like Seymour Pappert and Marvin Minsky at MIT, or Alan Kay at Zerox and Apple. But for all that long decade of talking about learning objects and SCORM, of learning journeys and personalized learning, where are we now? Still talking about “the migration to the electronic textbook”!

And why? Publishers say teachers demand texts, Teachers say students demand them. Everyone says parents insist on them, only please make them digital (easier to carry, cheaper to buy). And in every part of the developed world you hear the low moan of “falling standards, education is not gripping or immersive, kids are now exam monkies being trained to pass tests etc etc”. The inescapable conclusion is that our society is in denial in the education space so maybe our thinking should be turning towards what we do to change the fundamentals. And here the Eduventures work carries seeds of hope. The report which looks at adaptive learning focusses on “developmental education (for European readers, this is yet another euphemism for remedial work with less able learners). And this is critical in all of our societies: only where most traditional techniques failed completely will we seem to trust ourselves to something else. So where students must play catch-up we can go to people like Pearson Learning Solutions, who by investment or partnership have now put together a considerable hand of potential plays, from SmartThinking in the US to TutorVista in India covering the individualized instruction side of the deal, while the work with Knewton, mentioned here before, moves Pearson centrally into the service-led domain centred on creating course material for specific students with known and diagnosed problems, and making it adapt with them. Only the complete atomization of learning materials into objects with defined learning outcomes will allow these solutions to succeed, and publishers to survive. And yet, this outcome is still far from the minds of the biggest “textbook” publishers in the line behind Pearson.

Who else does good work in adaptive learning, then? Eduventures single out Carnegie Learning, now part of Apollo (the University of Phoenix, not the investor). Here is one of the future patterns: owners of schools and distribution systems, like Pearson and Apollo, custom to their own needs while the textbook players just melt in the heat. Another quoted player is Edmentum (PLATO plus Archipelago), now refinanced and with a strong bias to the developmental education field. Meanwhile a quite different and disruptive set of players are atomizing at the teacher level. Beyond the scope of these reports, note how rapidly systems are developing all over the world to network successful lessons, resources and techniques. When I was a textbook publisher our mantra, in times when the only tech was Monotype, was that we justified our existence by reflecting through our authors the best teaching practice that we could research and locate. Now every teacher can do that for themselves: TSL Education claims a network outreach to 47 million teachers globally from its UK and US sites, while teacherspayteachers.com publicizes a teacher in North Carolina who has earnt $1 million from selling learning resources online. Atomization pays!

But the revolution comes full turn when you apply predictive analytics. Education is a live Big Data environment, with huge caches of material concerning each learner, and increasingly, each learner’s reaction to each learning process. At the moment, these reports note, the focus of predictive analytics is finding out who is likely to fail and trying to help them in time. With 25% of US college learners dropping out before the end, this is a very expensive problem which needs to be solved. Predictive Analytics can be demonstrated to improve retention, both by improving selection and by diagnosing reasons for failure before it is too late. College teaching staff will be worried: failure to learn is also about failure to teach. Typically, the interviews in these reports show that data is siloed, that LMS data does not mix well with other content, that those who use predictive analysis mostly do so in terms of IBM’s SPSS software, and that the use of these analytic techniques was just as prevalent now in retention as in recruitment. Just what one would have expected. But when data analytics becomes an accurate prediction of outcomes, then personalized learning can really begin. Do not be stubbornly publishing textbooks, whether they are digital facsimiles or not, when that golden dawn arrives!

Mar

5

Mosaic Method and SOCMINT

Filed Under B2B, Big Data, Blog, Financial services, Industry Analysis, internet, mobile content, news media, Publishing, Search, semantic web, social media, Uncategorized | 1 Comment

The network makes writing more accessible, in that it reduces the barriers to acquisition while similarly diminishing the challenges to contributions. This is why I am celebrating three new books by old friends already this year. Jim McGinty started it with a powerful drama called “Right to Kill; A Brooklyn Tale” (read my review on Amazon). Myer Kutz followed, turning away from heavyweight tomes on materials science and engineering to contribute “In the Grip”, a clever psychodrama with a twist that gave me real pleasure crossing the Atlantic Order these on Amazon! But the book I will put on the shelf and value long after the first two have become TV scripts and earnt their authors untold millions comes from Alfred Rolington (former CEO of Jane’s Publishing and Oxford Analytica) who will never get rich with “Strategic Intelligence for the Twenty First Century: The Mosaic Method” (Oxford University Press, 2013) but who fills a real gap for professionals in this field. Having always assumed that the intelligence community knew more than the rest of us and been disappointed, I see now why they knew less, and why a new way of viewing strategic intelligence is vitally necessary.

Have you read anything about defense intelligence since the great wave of books about Bletchley Park that suggested that any defense intelligence agency anywhere knew enough about the present, let alone the future, to effectively suggest what might happen next? Alfred quotes Christopher Andrew, the Cambridge historian who is the acknowledged authority here to great effect, and I would add a further rider: Andrew’s big MI5 book is depressing on several levels, but one is the reflection that it prompts that the strategic intelligence operatives employed on both sides of the Atlantic were either not the sharpest knives in the box, or were so constrained by the sclerotic selection of pre-ordained intelligence methodologies available to them that they could only report on the views that they were able to take of the scenery, not the whole panorama.

Alfred’s analysis of current working methods in intelligence services bears reading by anyone who thinks that business intelligence is very much better. Why didn’t we predict the Arab Spring? Because we were not reading the blogs and the social media and the network unrest, not only in the countries concerned, but, more importantly, in the Arabic-speaking world as a whole. The answer here is the Mosaic Method – Big Data analytics for the defense industry – which looks at both historical and personal perspectives over time. Being able to to search vast tracts of data to present the evolving views of representative individuals, to see the intelligence picture through the eyes of the other side or the several concerned parties, becomes a vital extra component alongside the very straight-jacketed and traditional methodologies currently used.

So my surprise here was the relatively unsophisticated nature of much defense intelligence work. Helpful techniques which would drive a predictive analysis approach are already widely deployed in industry. I reflect that Lexis Seisint (clue in the last 3 letters) was originally deployed in the Department of Homeland Securities, and that Thomson Reuters’ ClearForest was passed through the fence by the the Israeli defense Agency to allow it to be exploited commercially. In addition few major Big Data software players – Palantir would be a critical example – do not have a large slug of defense related expenditure in their growth graphs.

So our conclusion must be that while huge amounts of data are gathered and sifted, the ability to construct predictive analysis from them is in its infancy. Alfred remarks that SOCMINT, for social media intelligence, is a relatively new coinage. My first thought was that inadequate intelligence might be conducive to world peace, but on reflection I share Alfred’s view that an overhaul is needed, and one which acknowledges that we are living in a networked world. If McDonald’s can be expected by their shareholders to be able to predict from blogs and social media amidst the firestorm on obesity when the optimum point arrives to launch the salt/fat/sugar free VeggieBurger, then we should as mature nations be able to predict the Arab Spring.

And to do that we need to be watching the right things. I respond very much to Alfred’s suggestion that we are not looking at the right countries – BRICS are important, but it may be yet more important to monitor and know more than we do about Nigeria, Indonesia, the Philippines, Iran (just look at the demographics) or Egypt. Very large countries with high proportions of their populations under the age of 21. They are like Europe in the Fifteenth century. Which leads to the other thing that I like about this book – it at least discusses the primacy of history for prediction. There is a class of intellectual dangerously roaming our universities who seem to believe that history began with Vin Cerf or Bob Kahn or even Tim Berners Lee. The truth is, as I see it, the exact opposite. Because we are moving into the unknown space of a networked society, we need to know more not less about how that society may react to change. Alfred reminds us, in his section on the Dark Web and elsewhere, that we have not yet fully explored what we refer to as the Web. This is the beginning of a story, not an ending.

Meanwhile, the OUP series is to be extended to cover cyber-security and cybercrime and Alfred has been blogging about this on the OUP blog: http://blog.oup.com/2013/02/cyber-attacks/

Here again is a critical area of intelligence, and reading this blog I reflected that by attacking Iranian nuclear installations with trojans the West may, as it has so often, be providing knowledge (in cracking cyber-attacks) which may one day be used against them. Like supplying training and guns to Saddam Hussein? Whatever the outcome, the importance of the subject matter – especially to those of us working on peaceful economic applications – cannot be ignored.

« go back — keep looking »