• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Super Artificial Intelligence, a naive approach

Beelzebuddy said:
Almost everything. Let's start with the first sentence:



No there are not. There are either zero or lots and lots of them, depending on whether by "brain based learning" you mean algorithms that operate in the same manner as the brain, or algorithms that people say operate in the same manner as the brain.

If you mean the former, there are none, although the Human Brain Project (the current offshoot of Markram's work, he's since dived back into biology) gets half credit for trying to do it right instead of half-assing it long enough to put out a press release.

But if you meant people who half-ass it enough to put out a press release about how their system is "inspired by" the human brain, you'll have a line out the door. And all of them are crap.

Deep learning is in neither of these categories. Any comparisons to brain function are red herrings generated by people who don't understand why it is what it is or how it does what it do.


For the peanut gallery, the book PGJ keeps referencing is online here, and appears to lack all the woo that keeps cropping up in his posts. A relevant snippet from the introduction:
Click to expand...

It is ironical that you read that statement from the DeepLEarningBook, but still wrote the garbage before the quote you cited.


Here are some other snippets you conveniently appeared to avoid:

(I)
Deep Learning Book:
(1)
The original Cognitron (Fukushima, 1975) introduced a more complicated version that was highly inspired by our knowledge of brain function
Click to expand...


(2)
Also, while neuroscience has successfully inspired several neural network architectures, we do not yet know enough about biological learning for neuroscience to offer much guidance for the learning algorithms we use to train these architectures
Click to expand...




(II)
My prior quote via reply #421:


ProgrammingGodJordan said:
(1) Deep learning(a) This ---------->began <---------- by using hints from neuroscience.
(b) However, deep learning benefits from paradigms outside of neuroscience.
(c) This is because deep learning experts don't know enough about the brain, to get much performance.
Click to expand...




FOOTNOTE:
So, from the deep learning book or otherwise, we observe that there are two main classes of brain based learning, Deep learning, (which is loosely based on brain), and Computational Neuroscience, (which attempts to strictly adhere to brain structure)

I had long mentioned this via reply #421.
 
Last edited:
Does anyone else think that this thread could have been so much more interesting if it had been "A General Discussion of Artificial intelligence" rather than "humans gods computers 10^15 =10^18 neural networks reference 1985 left brain right brain deep learning I-know-better-than-you-twerps"?
 
MikeG said:
Does anyone else think that this thread could have been so much more interesting if it had been "A General Discussion of Artificial intelligence" rather than "humans gods computers 10^15 =10^18 neural networks reference 1985 left brain right brain deep learning I-know-better-than-you-twerps"?
Click to expand...

The thread could probably be interesting to others here, perhaps if they were a different type of crowd...


Anyway, I don't expect many crowds to be intrigued by my Supermanifold hypothesis:
RA3GJle.png
 
Last edited:
Euclideon? Is that the name of a mecha I've never heard of, or are you referring to the Australian software company?
 
Myriad said:
Euclideon? Is that the name of a mecha I've never heard of, or are you referring to the Australian software company?
Click to expand...

Thanks, typo purged.

I tend to typo "euclideon" (graphics engine) for euclidean (geometry)

[IMGw=999]http://i.imgur.com/aq4QKZ2.png[/IMGw]

"Euclideon" is an "unlimited" detail graphics engine:

Euclideon main video: https://www.youtube.com/watch?v=00gAbgBu8R4

Another euclideon demon: https://www.youtube.com/watch?v=iVsyB938ovY

My comment on euclideon a year ago: https://www.youtube.com/watch?v=_5hg9VfbyYg
 
Last edited:
wea said:
[qimg]http://i.imgur.com/RA3GJle.png[/qimg]

Only 4 "some" in the sentence? Aren't you risking too much precision in your math language? :rolleyes:
Click to expand...


This is deep learning language.
Deep learning may hint at math paradigms such as manifolds.

Anyway, alot of data is encoded in those small lines.
 
Last edited:
You must log in or register to reply here.

Back
Top Bottom