The Author Online Book Forums are Moving

The Author Online Book Forums will soon redirect to Manning's liveBook and liveVideo. All book forum content will migrate to liveBook's discussion forum and all video forum content will migrate to liveVideo. Log in to liveBook or liveVideo with your Manning credentials to join the discussion!

Thank you for your engagement in the AoF over the years! We look forward to offering you a more enhanced forum experience.

ctoomey (4) [Avatar] Offline
#1
In chapter 2 you give a nice overview of parametric and non-parametric learning, followed by sections on parametric supervised and parametric unsupervised learning. After completing the latter I fully expected to come to sections on non-parametric supervised and non-parametric unsupervised learning. Instead there was an abrupt end of chapter with the conclusion section. I literally thought I'd somehow skipped a couple pages so scrolled back to find the skipped sections, to see that they were in fact omitted. For completeness and balance, you should really include at least small sections of those 2 of 4 learning methods that you've introduced.

Another thing to fix is the conclusion, which is not that at all -- it's some brand new information (that DL uses parametric learning) without a recap of the chapter's material.
Golo Roden (5) [Avatar] Offline
#2
+1
ebert (13) [Avatar] Offline
#3
+1
422782 (2) [Avatar] Offline
#4
+1
Andrew Trask (Grokking Deep Learning Author) (26) [Avatar] Offline
#5
Groovy. I'll add more meaty dialogue between parametric learning and Deep Learning.

How strongly do you feel the need for a section on non-parametric learning (given that this is indeed completely separate from Deep Learning).
ctoomey (4) [Avatar] Offline
#6
375112 wrote:Groovy. I'll add more meaty dialogue between parametric learning and Deep Learning.

How strongly do you feel the need for a section on non-parametric learning (given that this is indeed completely separate from Deep Learning).


Sorry, I'm just seeing your reply as for some reason I didn't get the notification of a reply to my thread.

I think it'd be really helpful to readers (certainly to me) to be given at least some more information about how non-parametric learning works and well as some of the reasons one might choose parametric vs. non-parametric. It certainly doesn't need to be very long, and you could/should start off by saying that you'll just give a quick overview since it's beyond the scope of the book.

The other option, which I would like less but would prevent setting the wrong expectation, is to clearly state in the intro. section that the following sections would only cover parametric learning since the other is beyond the scope.
Andrew Trask (Grokking Deep Learning Author) (26) [Avatar] Offline
#7
Alright! By popular vote, chapter 2 will be revised to include a broader section on Non-Parametric learning. smilie Thanks for the feedback everyone.
10135 (7) [Avatar] Offline
#8
Another few comments on chapter 2 besides this which is totally spot on:

1. You introduce terms w/o defining them - if this is a book for non-experts in the field not everyone is going to know what General AI is
2. You mention deep learning is a subset of ML because it's "neural networks" but you don't really give any information about other ML techniques or say very much about what neural networks are. Even though the book will go into that in great detail you have to give some general info from the start. Moreover, while this is a book about DL, without understanding the difference bet. DL & other ML techniques your readers will have no clue or motivation why DL is so exciting and why so many people are interested in it as opposed to other techniques.

Here is an example of an article that explains the differences very well and motivates the reader to learn about and use DL. Your fiirst 2 chapters very badly needs something similar:

https://medium.com/intuitionmachine/why-deep-learning-is-radically-different-from-machine-learning-945a4a65da4d

3. A minor quibble but you betray your U.K.origin: it's the Red Sox. Stick to Man City or Chelsea smilie
monger (2) [Avatar] Offline
#9
10135 wrote:
2. You mention deep learning is a subset of ML because it's "neural networks" but you don't really give any information about other ML techniques or say very much about what neural networks are.

+ 1

I got quite confused while following this chapter. First it explains shortly what deep learning is, and then drops it completely. Everything after that is titled "machine learning". So, are these topics in the scope of this book, or just mentioned for completeness?

I would have expected a structure like that:
What is machine learning?
What kind of machine learning areas are there?
What kind of areas does deep learning cover?

Every chapter that refers to deep learning should be somehow recognizable via name. Names are important.