Learning to Smell: Using Deep Learning to Predict the Olfactory Properties of Molecules
Alexander B Wiltschko,
Google AI Blog,
2019/11/13
How does our knowledge of smells work? According to this work with machine learning, graphs. "Since molecules are analogous to graphs, with atoms forming the vertices and bonds forming the edges, Graph Neural Networks (GNNs) are the natural model of choice for their understanding.... (a) single vector, representing the entire molecule, can then be passed into a fully connected network as a learned molecular featurization. This network outputs a prediction for odor descriptors, as provided by perfume experts."
Web: [Direct Link] [This Post]
Learning through Play? Why We Should Play It Safe
Paul A. Kirschner, Mirjam Neelen,
3-Star Learning Experiences,
2019/11/13
Citing David Geary's An Evolutionarily Informed Education Science (see also this overview of teleological biology, Paul Kirschner and Mirjam Neelen argue that while "children are naturally motivated to learn through play, experimentation, and discovery" this might not be appropriate for "cultural or biologically secondary knowledge" - that is, the knowledge we teach in schools. "It would be wonderful if we could acquire biologically secondary knowledge with the same ease, but unfortunately that’s not so. For example, skills such as reading, spelling, writing, doing maths, etcetera, take conscious effort and involve the limits of our working memory." I don't think there's any actual evidence to be found in evolution (or through the misuse of terms like 'folk psychology') to support the idea that there are these two distinct types of knowledge (the first of which would be positively Lamarckian), but there you have it. Image: Ashman's Taxonomy.
Web: [Direct Link] [This Post]
Announcing the Essential Skills Playbook
Leah Belsky,
Coursera Blog,
2019/11/13
This booklet is intended to tell readers "where and how can you discern the critical skills your workers need to succeed in an age of digital transformation." The link provided currently results in a 404, so I can't verify that one way or another. But what caught my eye were some very odd "findings" listed in the blog post. For example, in finance, "several skills are essential to finance, cropping up in multiple functions—SQL, Python language, big data, and business analytics." These are the essential skills in finance? Or for professional services: "data-related skills including data-based decision making, blockchain, and machine learning are must-haves in professional services." It's hard to believe these lists are the result of any actual analysis, much less any consultation with the real world.
Web: [Direct Link] [This Post]
How to Get More Rural Students Ready for College? Start With Broadband.
Goldie Blumenstyk,
Chronicle of Higher Education,
2019/11/13
Goldie Blumenstyk takes the report Why Rural Matters, 2018-2019: The Time Is Now from the Rural School and Community Trust as a starting point for her discussion of lagging achievement by rural students. Rural students tend to be from poorer families, they don't have easy access to resources like libraries and broadband internet, and even if they do succeed, it's not clear that there are employment opportunities using their new skills back home. Not all of this can be addressed with broadband, but a lot can be - if I had decent broadband in the small town where I like I could much more easily forge a high-tech career here, without having to drive into the city every day. The same is even more true the more rural you get. Image: Third Way, Broadband for All.
Web: [Direct Link] [This Post]
New Blockchain Student Recruitment Platform Manages Credentials, Transcripts and More
Rhea Kelly,
Campus Technology,
2019/11/13
I personally think that innovative uses of blockchain in education ought to consist of more than just taking existing database content and putting it onto a blockchain, but the bar does not seem to have risen that far yet. Thus we are treated to these database contents being placed onto a blockchain and called innovation.
Web: [Direct Link] [This Post]
This newsletter is sent only at the request of subscribers. If you would like to unsubscribe, Click here.
Know a friend who might enjoy this newsletter? Feel free to forward OLDaily to your colleagues. If you received this issue from a friend and would like a free subscription of your own, you can join our mailing list. Click here to subscribe.
Copyright 2019 Stephen Downes Contact: stephen@downes.ca
This work is licensed under a Creative Commons License.