1
$\begingroup$

Why training deep learning models is more likely to suffer from plateaus than local minima?

Could anyone illustrate it.

$\endgroup$