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
-
-
building up a synthetic notion of number and quantity the way embodied human-animals acquired it, via practicalities of countable, re-identifiable, categorizable things in the world... ... my $0.02: there will be other payoffs for commonsense reasoning we can optimize later
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
But that approach misses the point! Why include extra baggage when all evidence shows that it is not needed? Approximate arithmetic is innate, not exact arithmetic.
- 1 more reply
New conversation -
-
-
A calculator will be a tool to an AGI, the same way it is a tool to a human. It's possible to design an AGI with internal effectors to directly operate a calculator but what would be the point? A hand with fingers is much more universal since it can operate all sorts of tools.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.