0
$\begingroup$

I was wondering how can we show that cross-entropy is equivalent to maximising log likelihood of a training assuming the data can be modeled by this distribution: enter image description here

$\endgroup$
5