OLDaily

The Changing Face of TAFE
Reg Johnson, Cisco Blogs, 2021/06/30

How has the pandemic impacted Australia's Technical and Further Education (TAFE) system? Cisco and Optus commissioned a report, which is now available: Changing the Design of TAFE for a New Normal (18 page PDF). Not surprisingly, the report argues that the pandemic "has created a permanent change in the way TAFE campuses are designed and operated." Also not surprisingly, given the source, we read that "industry partnerships will be more critical than ever." Interestingly, the report suggests a rise in 'blue tech skills' - "occupations and skills that are technology-intensive but require a sub-degree qualification." There's a lot more in the report (which reads like a series of infographics), including automated buildings, digital-first strategies, and new program areas.

Web: [Direct Link] [This Post]

The big deal this week in online ed
Goldie Blumenstyk, Chronicle of Higher Education, 2021/06/30

Goldie Blumenstyk raises some of the many questions raised by this week's $800 million sale of EdX. Among the concerns: how will the money be spent - because EdX was a non-profit, it can't be kept as profit. "We know precious little about what the new nonprofit plans to do with this gargantuan infusion of cash," she writes. Also, what do we do about 2U "gaining access to a vast database of potential learners?.. Talk about a marketing channel." How will 2U use EdX? "Although 2U said it would operate edX within the company as a “public-benefit entity,” I can’t tell you exactly what’s been promised because that isn’t public."

Web: [Direct Link] [This Post]

An introduction to object detection with deep learning
Ben Dickson, TechTalks, 2021/06/30

This post might be too technical for some, but if you're willing to wade in, it's not too technical, and gives you an appreciation for one particular aspect of artificial intelligence: object detection. One major type of AI used for this is the the convolutional neural network (CNN). These are composed of convolutional 'layers', each of which extracts a specific feature from (say) an image; these are then 'pooled' to reduce the set of feature maps and zero in on an object. A more recent type of AI is the the region-based convolutional neural network (R-CNN), and even more recently, Fast R-CNN and Faster R-CNN which bring region selection and feature selection into a single model. Finally, there's “You Only Look Once” (YOLO), which "instead of extracting features and regions separately, performs everything in a single pass through a single network."

Web: [Direct Link] [This Post]

Building A Skills System Of Record: EdCast Releases Skills Studio
Josh Bersin, Bersin Associates, 2021/06/30

Josh Bersin Associates continue their very good work looking at developments in the skills and competencies taxonomy industry. "Every new recruiting tool (Eightfold, Avature, iCims) builds a skills taxonomy around job postings. Every learning tool (EdCast, Degreed, Cornerstone, Fuse, Docebo) builds a skills database to index, find, and recommend content. And all the talent marketplace platforms (Gloat, Hitch, Fuel50, and others) build skills databases." This post looks at EdCast, and divides the offering into three products: a skills graph, a skills studio "that lets you view, edit, and manage skills and job hierarchy", and analytics and reporting tools. There's a lot of AI behind the scenes, "using the EdCast Skills Studio to 'train' and 'untrain' the taxonomy... This, by the way, is what Microsoft Viva Topics does for documents and business processes. It monitors activity and continuously improves it’s 'topic index' engine."

Web: [Direct Link] [This Post]

We Should Test AI the Way the FDA Tests Medicines
Carissa Véliz, Harvard Business Review, 2021/06/30

This is a grossly misleading headline in Harvard Business Review. AI is already tested - and retested, and tested again - for accuracy, bias, and a host of other factors. What author Carissa Véliz is actually saying is that AI should be regulated the way we regulate medicine. We don't get to this until eight paragraphs in, but we get to it: "An agency similar to the Food and Drug Administration could be created to make sure algorithms have been tested enough to be used on the public." Now I don't have an issue with regulating AI, for exactly the sorts of reasons cited in this article. But let's call it what it is - regulation. And if we're going to regulate, let's ensure that the testing and regulation is proportional to AI can produce, rather than simply copied from another discipline, and let's ensure that the focus is on preventing harm in specific applications.

Web: [Direct Link] [This Post]

Education Research and ‘Negative Capability.’
Carl Hendrick, Chronotope, 2021/06/30

Aaron Davis links to this article from 2014 and I think it is well worth sharing in 2021. In it, Carl Hendrick writes about the need to be sceptical about education research as applied to the classroom. He introduces us to the idea from Keats that "what quality... Shakespeare possessed so enormously (is) Negative Capability, that is, when a man is capable of being in uncertainties, mysteries, doubts, without any irritable reaching after fact and reason." And this, he writes, applies in education as well. "Instead of passively accepting the ‘stone tablets’ of research we should be engaged in a constant dialogue with research, questioning it, challenging dogmatism, teasing out relevance to our own context and our own individual problems in a sort of ‘detached attachment.’

Web: [Direct Link] [This Post]