I'm reading Tariq Rashid's book "Make Your Own Neural Network".
It is a brilliant introduction to NNs (rare to see a technical book so engaging to all ranges of abilities)...but I think there is a typo which is not in the online errata for the book.
See the attached image. I don't think the 2 in the last equation is correct, shouldn't it be simply 2?
It could be that the author has anticipated himself, because the error function derivative is actually taken away from the weights in the neural network. But as far as the maths go, should that be 2 or simply 2?
It could also be that I don't understand the chain code!
It is a brilliant introduction to NNs (rare to see a technical book so engaging to all ranges of abilities)...but I think there is a typo which is not in the online errata for the book.
See the attached image. I don't think the 2 in the last equation is correct, shouldn't it be simply 2?
It could be that the author has anticipated himself, because the error function derivative is actually taken away from the weights in the neural network. But as far as the maths go, should that be 2 or simply 2?
It could also be that I don't understand the chain code!
Attachments

15.2 KB Views: 8