I’m working on my own post on NNs that focuses more on deriving backprop from computational graphs. I think that method of doing so also builds up a lot of the Chain Rule intuition, as you can easily see how the derivatives for earlier weights can be derived from those in later weights.
Thanks. I agree with using computational graphs. I think understanding backpropagation using graphs is much easier to understand if you are new to the subject. The reason I didn’t do it here is mainly because there’s already a lot of guides that do that online, but fewer that introduce tensors and how they interact with deep learning. Also I’m writing these posts primarily so that I can learn, although of course I hope other people find these posts useful.
I also want to add that this guide is far from complete, and so I would want to read yours to see what types of things I might have done better. :)
For sure! To be honest, I got a little lost reading your 3-part series here, so I think I’ll revisit it later on.
I’m newer to deep learning, so I think my goals are similar to yours (e.g. writing it up so I have a better understanding of what’s going on), but I’m still hashing out the more introductory stuff.
Thanks for writing this series!
I’m working on my own post on NNs that focuses more on deriving backprop from computational graphs. I think that method of doing so also builds up a lot of the Chain Rule intuition, as you can easily see how the derivatives for earlier weights can be derived from those in later weights.
Thanks. I agree with using computational graphs. I think understanding backpropagation using graphs is much easier to understand if you are new to the subject. The reason I didn’t do it here is mainly because there’s already a lot of guides that do that online, but fewer that introduce tensors and how they interact with deep learning. Also I’m writing these posts primarily so that I can learn, although of course I hope other people find these posts useful.
I also want to add that this guide is far from complete, and so I would want to read yours to see what types of things I might have done better. :)
For sure! To be honest, I got a little lost reading your 3-part series here, so I think I’ll revisit it later on.
I’m newer to deep learning, so I think my goals are similar to yours (e.g. writing it up so I have a better understanding of what’s going on), but I’m still hashing out the more introductory stuff.
I’ll definitely link it here after I finish!