machine learning - how to interpret the "soft" and "max" in the SoftMax regression? -

i know form of softmax regression, curious why has such name? or historical reasons?

the maximum of 2 numbers max(x,y) have sharp corners / steep edges unwanted property (e.g. if want compute gradients).

to soften edges of max(x,y), 1 can use variant softer edges: softmax function. it's still max function @ core (well, precise it's approximation of it) smoothed out.

if it's still unclear, here's read.