Deep learning

    This post by Stephen Downes took me to this piece on the adoption of interactive touchscreens. I won't repeat much (if anything) of what was said there, but led (via a tweet) to this diagram on deep learning created by Impact Wales. What a chain of links! There is so much 'curation' of content going on here.

    But my point was that the digram is very useful, such that I wanted to keep a record of it.

    Promoting deep learning

    Promoting deep learning

    But to return to the topic of the chain of links, there is so much 'curation' of content going on here. This fact prompts two thoughts, one about old technology and one about new technology. First, RSS feeds. Stephen also pointed to this piece on Who killed Google Reader (the excellent RSS reader that met an untimely, and some would say unwarranted, demise). RSS with Reader was meant to "become the world’s best collaborative and intelligent web content delivery service". In other words, it was meant to enable the sort of curation that led me to the deep learning image.

    Second, and much more contemporary, this sort of chaining is something we're seeing AI (in the form of chaptGPT, Bard, etc) working towards. For example, Google is now talking about when you ask it a question (which is what most of us do, I think, when we use Google), it will present a synthesized answer from all the links it has; i.e., a paragraph that answers the question. One can then ask Google to 'unpack' the answer, in which case it will provide the links to the main sources it used to create the answer. As an aside, I'm not pretending that's unproblematic, I'm just parking for issues for now.

    Either way, this ability to chain content together seems like a form of deep learning to me.

    If you webmention this page, please let me know the URL of your page.

    BTW: Your webmention won't show up until I next "build" my site.

    Word count: 400 (about 2 minutes)


    Updated: 2 Jul '23 09:37

    Author: Peter Smith


    Section: blog

    Kind: page

    Bundle type: leaf

    Source: blog/2023/07/02/deep-learning/