[ExI] Neural networks score higher than humans in reading and comprehension test

John Clark johnkclark at gmail.com
Wed Jan 17 00:33:36 UTC 2018


On Tue, Jan 16, 2018 at 5:45 PM, Dylan Distasio <interzone at gmail.com> wrote:

​> ​
> I will quote the inventor of Keras which is a heavily used deep learning
> library layered on top of Tensorflow (or other) architecture to sum up my
> views:
>
> "Neural networks" are a sad misnomer. They're neither neural nor even
> networks. They're chains of differentiable, parameterized geometric
> functions,  trained with gradient descent (with gradients obtained via the
> chain rule). A small set of highschool-level ideas put together
>


I look at something marvelous but so complex I can't understand it, ​I then
​break that marvelous thing up into smaller and smaller parts ​until​
eventually I come to a part that is so simple it can easily be understood
even by a child. And that proves I was originally mistaken, when all those
billions of simple parts were working together it couldn't have been
marvelous after all.

I don't know why people make such a big deal about Shakespeare, all he did
was place one ASCII character after another, even a child can do that.

 John K Clark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20180116/e5ffc643/attachment.html>


More information about the extropy-chat mailing list