Mathematicians say that deep learning lacks mathematical grounding. I will be vaguely amused when deep learning replaces the mathematicians.
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
DeepMath is an impressive system. But there's still a long path ahead for ML-powered mathematics...
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Mathematics lacks empirical grounding. The metaphysics of mathematics is divorced from reality by a fat ass called Thomas Aquinas.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Except with an extreme word sense disambiguation problem. http://bit.ly/2ck4JK2
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
It gets even harder when math symbols (-constants) don't carry semantic value. Like predicting programs by variable names.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
deep learning will turn machines against us.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
my argument isn't that it should be more formalized. I just don't think math will be terribly learnable.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
try training a model to learn to predict kth element in OEIS data and you'll find it doesn't work well.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
yes, but as sample complexity increases, you begin overfitting. Memorizing vs. Learning.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
That's not the argument either.
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.