I work in the Learning and Performance Support Systems program at the National Research Council, a multi-year effort to develop personal learning technology and learning analytics. I am one of the originators of the Massive Open Online Course, write about online and networked learning, have authored learning management and content syndication software, and am the author of the widely read e-learning newsletter OLDaily.

SUBSCRIBE TO OLDAILY DONATE TO DOWNES.CA

~ ~ Vision Statement ~ ~

Twitter EDU

Icon

David Truss has published an eBook on using Twitter in education. According to the blurb, "If you follow along and tweet as you learn, this book will make your entry into Twitter much easier, and enjoyable! It has best practice tips, tricks and explanations that will assist you in building a great network much faster than you could do on your own." The book can be downloaded for free (and there are no signup forms or anything, so I'm already a fan). Available in ePub only. Via Alistair Creelman., who writes, "the real key to success with Twitter is engagement. To get something out you have to contribute. If you show that you provide useful information, ideas and tips then people will follow you."

Today: 136 Total: 254 David Truss, 2018/01/19 [Direct Link]

No, machines can’t read better than humans

This is a counter to a post published earlier this week suggesting that robots can now read better than humans. Technically, the headline wasn't wrong, but the problem lies in the test. "The test is actually a dataset, compiled by a group of Stanford university computer scientists," explains the author. "It’s called the Stanford Question Answering Dataset (or SQuAD for short)." It poses a set of questions based on a set of Wikipedia articles. But this sort of test is easy, says James Vincent. It's just a pattern matching test, and doesn't require inference or comprehension. 

Today: 133 Total: 236 James Vincent, The Verge, 2018/01/19 [Direct Link]

MOOC Trends in 2017: Content Paywalls

Icon

Class Central has been running a series of articles on MOOCs in review in 2017. It reads to me like a description of "how MOOCs stopped being MOOCs in 2017". The articles cover content paywalls, MOOC monetization, and corporate learning. But the series also attests to the continued strength of MOOCs in general. "To date, over 800 universities around the world have launched at least one MOOC. The total number of MOOCs that have been announced stands at 9,400, up from 6,850 last year." That adds up to about 78 million students. 

Today: 138 Total: 243 Dhawal Shah, Class Central, 2018/01/19 [Direct Link]

Innovations for Educators: IBM’s Teacher Advisor

Icon

The premise of this story is that instead of replacing teachers AI might help teachers. As an example, the author uses IBM’s Teacher Advisor. "Teacher Advisor is a free online resource that helps elementary school teachers plan math lessons," writes Luis Flores. "It houses a library of K-5 math open education resources (OERs)—including activities, lesson plans, and supporting materials—from organizations such as EngageNY, [Massachusetts]CPALMS, and UnboundEd." Additionally, "The AI technology in Teacher Advisor allows teachers to search for resources and create lesson plans in a fraction of the time it would take them to do it alone." 

Today: 168 Total: 290 Luis Flores, Christensen Institute, 2018/01/19 [Direct Link]

2018 Map of the Complexity Sciences

Icon

This is a great visualization of the major streams of thought in the field of complexity theory. I like the way it shows the links between the different strands, and also that ti is an interactive graphic - click on an area and be taken to the relevant Wikipedia page. From my perspective it seems that the more recent topics signal an end-game for the field. In the 2010s we see 'applied complexity', 'complexity policy & evaluation', and 'mixed methods'. Via Dave Snowdon.

Today: 64 Total: 380 Brian Castellani, 2018/01/18 [Direct Link]

Adaptive Learning: The Premise, Promise, and Pitfalls

Icon

This is a review paper intended to "explain what adaptive systems are and what kinds of data they require,... to categorize the main use cases and possibilities of adaptive systems [and] to outline the current limitations and concerns surrounding adaptive systems." In two paragraphs it deftly summarizes the landscape, listing new companies (Acrobatiq, Knewton, CogBooks, Cerego, Realizeit, LoudCloud, Smart Sparrow) as well as the work of publishers, LMS companies and universities. The article lists a number of studies showing effect sizes nearly matching that of 1-to-1 tutoring. But it also references a number of studies where "the results were decidedly mixed." And it describes three potential pitfalls: discrimination and labeling of students, creating consequential feedback loops; nNarrow constraints of knowledge, knowing, and learning; and questions around transparency, availability, and security of data. This isn't a long paper, but it's well-written and informative.

Today: 56 Total: 336 Petr Johanes, Larry Lagerstrom, American Society for Engineering Education, 2018/01/18 [Direct Link]

[All the Posts]
Are not two sparrows sold for a penny? Yet not one of them will fall to the ground outside your Father’s care.