OLDaily

Scant Evidence of Power Laws Found in Real-World Networks
Erica Klarreich, Quanta, 2021/05/06

This article is from a couple of years ago, but I had never heard of Quanta then and so had missed it. The key argument is in the headline (and in the title of the related research (14 page PDF), 'Scale-free networks are rare'. For those who don't recall the discussions from the early 2000s, the idea of a power law is that in a dynamic and still-forming network you get 'preferential attraction', a phenomenon where new nodes link to well-established nodes, making them even more well established. 'The rich get richer', in other words. This, it was argued, is a 'natural' phenomenon of scale-free networks like (say) the internet or world financial markets. I argued against this in a presentation at Northern Voice in 2005 (that had audience members yelling at me). And, as it turns out, "the statistical tests rejected a power law as a plausible description of the network’s structure" (though to be fair Erica Klarreich also caution that in this research "both proponents and critics of the scale-free paradigm see what they already believed to be true").

Web: [Direct Link] [This Post]

Scans peek at brains of highly sensitive people
Sonia Fernandez, Futurity, 2021/05/06

To be honest, I think of fMRI as modern phrenology, trying as it does to understand cognition by the shape of your blood flow. As one person said once, it's like trying to read data on a hard drive by measuring the temperature of the case. So I am sceptical about the evidence provided here, but the hypothesis is intriguing. It's this: "depth of processing is a cardinal feature of high sensitivity." What can that mean, aside from the hocus pocus of heat signatures? Maybe this: people who experience the world more deeply may be more prone to sensory processing sensitivity (SPS), "a characteristic that can result in a variety of behaviors, from emotional outbursts to withdrawal, overwhelm, and procrastination." We would still need to define 'more deeply' - it could mean anything from vividness of conscious perception to the number of neural layers a perceptual experience stimulates. But you would think that maybe there's some connection, maybe?

Web: [Direct Link] [This Post]

60% of School Apps Are Sharing Your Kids' Data With Third Parties
Shoshana Wodinsky, Gizmodo, 2021/05/06

The margin of error in this headline is probably several hundred percent, or so it seems at least, when this is the methodology: "The Me2B team surveyed a few dozen so-called 'utility' apps for school districts - the kind that students and parents download to, say, review their school’s calendar or bussing (sic) schedules - and found roughly 60% of them sharing." I mean, it's not good that any apps are illegitimately sharing data, but this article looks only at 'utility apps', and only a 'few dozen' of the hundreds that are out there, and are estimating data sharing only by looking at the list of software development kits (SDK) used by the apps, not actual sharing that may or may not happen.

Web: [Direct Link] [This Post]

Why AI is the future of Socratic learning methods
Aliki Constantinou, eSchool News, 2021/05/06

I've never really been a fan of the Socratic learning method. Named after the Greek philosopher, it's a method of having the learner discover answers for themselves by asking the right questions. In the actual dialogues, they are for the most part just rhetorical questions. Like: "Socrates: 'Killing is wrong, is it not?' Protagoras: 'It is.'" Etc., for 50 pages or so. Anyhow, in this article the method is rendered as: "to identify inconsistencies and achieve greater understanding of one’s own values, beliefs, and even convictions was to ask the right questions and constantly ask: 'How do I know what I know?'" How does AI help? "AI has the incredible ability to ingest knowledge and learn more with each exchange, allowing those who have met their AI avatar to benefit from instruction that is uniquely catered to their own learning style and current of understanding." OK, I'm just going to call all this the utter nonsense that it is. I don't think the editors at eSchool News are even trying any more.

Web: [Direct Link] [This Post]

Learning analytics in higher education: a preponderance of analytics but very little learning?
Carolina Guzmán-Valenzuela, Carolina Gómez-González, Andrés Rojas-Murphy Tagle, Alejandro Lorca-Vyhmeister, International Journal of Educational Technology in Higher Education, 2021/05/06

So what do you get from learning analytics (LA), exactly? We think it means more learning (whatever that means) but it's not clear that the research bears this out. According to the authors, "the extent to which the LA data generated, gathered, and analysed actually corresponds to learning remains unclear." They distinguish between "a practice-based community led by management units within higher education institutions and an academic community whose object of research study is LA as such." Each finds its own value in LA, but that value is different for each. But that is what raises the question of 'what is learning', and "an undue emphasis on metrics and quantification in research on LA legitimises" doesn't make the answer clearer.

Web: [Direct Link] [This Post]