Suggestion on Changing Induction Step for Backward Propagation

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

I felt an induction step on Z looks simpler than an induction step on A…
Here is a summary.

specs:
a(l)=g(z(l))
z(l)=w(l)'a(l-1)+b(l)
j=-(ylog(a)+(1-y)log(1-a))
J=sum(j)/m

back prop induction:
dZ(l-1)=w(l)dZ(l) * g’(Z(l-1)) [stores j-derivatives columnwise, each example is 1 column]
dw(l-1)=A(l-2)dZ(l-1)’/m [J-derivative]
db(l-1)=sum(dZ(l-1) over columns)/m [J-derivative]

Though, I dont know much about the literature and any specific advantages the original representation may provide. Look forward to your views.

2

If you wait until Prof Ng gets to the fully general treatment of all this in Course 1 Week 4, I think that’s the way he does it. It looks like you’re just addressing the Logistic Regression version here.