Algorithm 6.2's notation confuses me a little:
\sum_{i:j \in Pa(u^{(i)})} grad\_table[u^{(i)}]\frac{\partial u^{(i)}}{\partial u^{(j)}}
What does $i$ represent here? How do I read $i:j \in Pa(...)$?
Thanks
Read more… (20 words)
Active In
Deep Learning Book
Deep Learning
Artificial Intelligence
Featured Contributions
NaN.
discussion
Chapter 6: Deep Feedforward Networks
If you're reading Deep Learning Book, feel free to ask questions, discuss the material or share resources.
Below is a video from a book club in San Francisco, ...
Read more…(37 words)
reply in this discussion
NaN.
discussion
Chapter 6: Deep Feedforward Networks
If you're reading Deep Learning Book, feel free to ask questions, discuss the material or share resources.
Below is a video from a book club in San Francisco, ...
Read more…(37 words)
reply in this discussion
John, I also find this section less clear. However, this is what I think the book meant: probability distributions can be described by parameters, e.g. univariate Gaussian PDF is defined by its mean and variance. When we try to learn about a Gaussian distribution from data, we try to learn both of these parameters, variance, like mean, is just a parameter. I agree that some concrete examples would have been helpful.
Read more… (71 words)