anecdata - I once trained char-rnn on 100s of 1000s of autogenerated parody postmodernist articles http://www.elsewhere.org/journal/pomo/ it was pretty good and even numbered the footnotes but it never got the idea that the footnotes needed to be ordered...
Math is hard--for neural networks. Particularly multiplication. New benchmarks, problematic performance. = the perils of trying to learn things that might more sensibly be built in. https://arxiv.org/abs/1910.01888
-
-
-
best phrase i ever accidentally invented: self-modern post-parody.
End of conversation
New conversation -
-
-
Math isn't an intrinsic capability of human cognition.
-
arguable, see eg literature reviewed in the number instance bt
@StanDehaene . but arithmetic is definitely innate to most microprocessors and foolish IMHO not to build into AI. - 1 more reply
New conversation -
-
-
I would presume that it comes from the fact that
#AI cannot observe itself while solving a problem and generalize from that observation (which is what human do : they perform an introspection on how they solve problems).Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
To resolve this, note that a computed model of a Higgs boson is not discovery of an artificial Higgs boson. Where is the proof, that everyone assumes true, that an artificial brain is a computed model of a brain? If you can't locate that proof: there's your problem.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Reminds me of how some people say "the brain can't be probabilistic let alone bayesian - we are terrible at those maths!"
-
In many circumstances what matters is causation and not necessarily the quantitative aspect
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.