Algorithms and Education

[peter]

I read recently that algorithms have been developed that are capable of marking student examination papers and achieving results that exactly correspond to grades awarded by human markers of the same questions. If this is indeed the case is it not time for some of the major players (think Google, Wikipedia, Facebook et al) to develop an internationally recognised degree awarding online university that would be freely available and free of charge to anyone with internet access. Certainly some subjects would require additional practical seminars to qualify, but there are many that could be entirely conducted online making education to higher levels available to all.
The key to the success of such a venture would be to maintain highest of standards and the establishment of universal recognition of the degrees awarded by the body. Difficult certainly, but by no means impossible - and with the full commitment and cooperation of the huge and influential bodies mentioned, certainly achievable.

[Hashi Lebwohl]

Low-cost university education to anyone in the world with Internet access? The ability to foster a new generation of scientists, engineers, and artists anywhere in the world, whose contributions might be revolutionary?

Are you crazy or something?

[Avatar]

--A

[peter]

Seemed like a good idea at the time.

[Cord Hurn]

The key to the success of such a venture would be to maintain highest of standards and the establishment of universal recognition of the degrees awarded by the body. Difficult certainly, but by no means impossible - and with the full commitment and cooperation of the huge and influential bodies mentioned, certainly achievable.

I agree with you that that would be key to whether such a venture would work or not. It would be great if it could be done, and that we could identify genius and talent wherever it shows up in the world. If done right, the payoff to the world could be enormous.

[Vraith]

It's an interesting idea, peter.
But here's the thing:
By the time you get such a thing designed, approved, and functional---it will be totally out of date.
All of the things that even the greatest teachers can do that are about the content will be more easily achievable, far more effective, and 100% individualized by deep-learning machines within 10-15 years. It COULD be done even more quickly.

There are other aspects to learning and teaching that aren't about content...but once the content is coded and basically free [all you'll need is a laptop or even a fairly cheap smartphone and VR headset. Textbooks are unnecessary---you'd be amazed what we pay for as information that is utterly free, and the AI can find in microseconds], the profession will mostly vanish, or alter beyond recognition.

The same would be true for a whole bunch of professions---but they're richer and more powerful right now, so they'll do everything they can to slow it down. [[I've said before---90% of the legal profession, the financial sector, design work [building, fashion, roads,] and others are ALREADY every bit as obsolete as a burger flipper. In fact most of the financial sector workers/advisors/bankers are LESS necessary, right now, than a good burger flipper. [[and far less necessary than a great teacher...though not for much longer]] I mean that literally.

[peter]

What are deep learning machines V? It's not so much a case of how the learning is done is it, as opposed to the establishment of a recognised and internationally accepted qualification/grading system that can be applied to such 'off-site' learning methods (although a slightly frightening thought occurs to me that even learning may become redundant once AI in the form of quantum computing can outstrip us by performing almost instant research in any form of currently academic endeavour that we now pursue: what purpose then does learning or even humanity serve?)
But yes - the shockwave of the erosion of much of what we currently perform as 'jobs' by the fast approaching tech revolution has yet to hit us, and I suspect when it does there will be a 'luddite' backlash of sizable proportions - but of course by then it will be too late.

[Avatar]

The new path to AI: Deep Learning

--A

[Vraith]

It's not so much a case of how the learning is done is it, as opposed to the establishment of a recognised and internationally accepted qualification/grading system

Well, what you were first talking about is only sorta-related to how the learning is done. The larger issue of how it is done is an entirely different topic, of course.
But my point was by the time you can get agreed standards, agreed required knowledge base, agreed systems of confirmation [preventing cheating/forging/etc], and so on the machines I was talking about will exist---and the whole process will be documented and self-confirming, integrated with the machine/system itself.
People might---probably will---still try to play political/protective/monopolistic/authority games with recognized certifications. But the fact that it is just a scam or power-play will be totally obvious for the shit it is.

[peter]

Well, the professions have traditionally protected their monopoly on their .....errrrr.......professions by the use of an esoteric lexicon available to and understood only those initiated into their ranks. But apparently it is the professional disciplines that lend themselves most readily to replacement by advanced tech and algorithms. In a TED talk I saw a few days ago an AI expert explained that the very repetition that has to be performed by an expert surgeon in order for him to justify the title 'expert' is exactly what automated systems are best able to perform. In a short period all of the diagnosis, prescription and surgury performed by medical practitioners will be performed better, quicker and cheaper by machines than by man. The same applies to the Law, teaching, architecture et al. Apparently it is the traditional middle class ground that will be most affected in the shorter term.

[Hashi Lebwohl]

Here is a recent Science Daily article about using algorithms to analyze music in ways have not been done before.

[peter]

That's going to be a massive step forward in the understanding of and future composition of complex musical pieces. But further, on a higher level will it not be a demonstration of how ultimately it will be computers that will solve the mysteries of the human brain, not humans themselves. We will be watching them, watching us, watching them

[Vraith]

That's going to be a massive step forward in the understanding of and future composition of complex musical pieces. But further, on a higher level will it not be a demonstration of how ultimately it will be computers that will solve the mysteries of the human brain, not humans themselves. We will be watching them, watching us, watching them

Maybe. And maybe not just understanding the past, or analyzing current flows/data, but making the future music/aesthetics [and who knows what else] in general?
After all, aesthetic values DO change. And even our primitive machines DO make decisions that are good, correct in outcome, and which we don't understand...at least not at first and not very quickly.

But I'm more interested in your apparent Jethro Tull quote/analogue.

He's watching me watching you watching him
watching me
I'm watching you watching him watching me
watching

[peter]

That's going to a massive shift in how the world works though isn't it V? Think f it. What we are saying is that for the first time ......well ......ever really, the complete spectrum of advancement, of progress if you like, will be out with and beyond human control, and in all likelihood human understanding. We will neither be party to the direction it takes, nor cognisant of whether it will be to our benefit or not, reduced to the level of non-participants in the world we once assumed to master. By the time we realise this to be the case, it will be the case and way beyond the point where we can do anything to stop it. We will of our own volition and willingly if unknowingly, like TC with his ring to Foul, have ceded control of our destiny to hands other than our own.

(Yes - that was a good spot V, though in fact I was quoting from a low grade UK tv show that used to pull stunts on unsuspecting people and film them for broadcasting and themselves used a paraphrase of the line. Damn I loved Tull! We're you a fan?)

[Vraith]

That's going to a massive shift in how the world works though isn't it V?

(Yes - that was a good spot V, though in fact I was quoting from a low grade UK tv show that used to pull stunts on unsuspecting people and film them for broadcasting and themselves used a paraphrase of the line. Damn I loved Tull! We're you a fan?)

It is a massive shift...but as long as we keep the things from being actually self-serving/self-aware, and check the results/impacts BEFORE we turn solutions loose in the wild [[[yea, we are SO good at looking before we leap. That's why we had a nuclear-energy powered world long before we had a nuclear-fear driven one...hahfuckinghah]]].

And yea, I was a fan up until the album that included that tune. Seen live a couple times. Since that album---I've given all of it a listen, but not much of it was very good.

[peter]

Lucky Man! I missed them live!