Hi,
There is a question in the Week 2 quiz of DLS course 1 for which I need an explanation on the right answer. The question is -
What does a neuron compute?
Can someone provide an explanation for the right answer to this question ?
We’re not supposed to share the quiz questions and the correct answers on the public forums. I took a look at that question and the other way to phrase it is “what are the computations that happen during forward propagation”. Prof Ng covered this in the lectures. If you carefully read all the multiple choice answers they give you, it should be clear which one matches what Prof Ng showed us.
Well, actually, maybe the problem here is that this question is a bit too “forward looking”. I did a quick scan of the lectures in Week 2 and I’m not sure he uses the terminology “neuron” here. It’s in Week 3 that he really switches to talking about it that way. Maybe that’s really the issue here: the answer seems straightforward to someone who’s already seen the whole course, but not to someone who has only watched the lectures for Week 1 and 2.
Think about it this way: in Logistic Regression, there is only one output neuron, which is the single value that gives the “yes/no” prediction. So what is the formula for how that answer is computed?
Thank you for the explanation. This helped me understand.
1 Like