I am reading THIS blog post, and I do not understand the logic behind this part:
Why is the relu here
max(0, Xavg - X) ? And even so, it does not really explain the values obtained in the two situations where
X1 = X2 = 0,
X1 = 1 and
X2 = 0, and
X1 = X2 = 1.
Can someone please explain this in more detail?