Why g' in š‘‘ā¢š‘[š‘™] =š‘‘ā¢š“[š‘™] āˆ—š‘”[š‘™]ā€²ā”(š‘[š‘™]) and not dg?

Course Q&A Deep Learning Specialization Neural Networks and Deep Learning
,
1

in š‘‘ā¢š‘[š‘™] =š‘‘ā¢š“[š‘™] āˆ—š‘”[š‘™]ā€²ā”(š‘[š‘™]) gā€™ derivative of g (activation function)

why gā€™ in š‘‘ā¢š‘[š‘™] =š‘‘ā¢š“[š‘™] āˆ—š‘”[š‘™]ā€²ā”(š‘[š‘™]) and not dg?

š‘‘ā¢š‘ is dL/dZ is it because its not dL/dg?

2

As I mentioned on your previous thread about this, this is beyond the scope of this course. I gave links on the other thread that cover the derivations.

Suppose g^{[l]} is ReLU or sigmoid. What do you mean by dg = \displaystyle \frac {\partial L}{\partial g}?