This is the second edition. So just to make a little note here to be your first edition. I believe it is in black and white it covers TensorFlow one, but the new updated version which is not much more to this version 2 or addition to that we’re going to talk about today. It has more information than ever before and of course It’s in color and it covers tensorflow 2.0 so let’s just dive in here to the book review and I’ll tell you the ins and outs of what I think. All right, so just to start off with you I gave this book 4.5 stars out of 5, so we’ll talk a little bit about this. one of the most important aspects in book reviews. In my opinion it is being able to really understand. What is the book good for and who is a good for and then who is it not good for me? because even though it might be a stellar book on one topic. It might not be a good book for someone else on a different topic. So let’s just talk a little bit about the book itself why I like it and then I’m going to talk about who it’s for and I’m going to talk a little bit about the cons and I’m going to contrast this to two other books and will kind of show you some examples so you can kind of get an idea what I’m talking about on who it’s for who it’s not for. Okay, so this book is for just to start off with you’re the first person who is going to be the beginner.
You want to get into data science and want to get into machine learning. You want to learn some deep neural networks, right? You don’t really know where to start but you want to start getting hands-on experience. This is the book for you. This is probably my favorite book. I have used thus far because one the book explains everything from a more detailed level but not necessarily a mathematical level. So again, that’s a con for me write the math level, but if you’re just diving into the first time I’m this book is very good at introducing Concepts and ideas, but he does a really good job of explaining like the details of why things are kind of working and how they’re structured and how they get put together and then he has amazing graphics on kind of making these ideas clear. So he talks about you know, how this is put together. So for example, like gradient descent or svm’s any talks about, you know shows a little chart and you see like, okay like on an svm. This is where the whines come on the screen if you have different lines you can move them further or closer apart.
results and there’s different colored dots to indicate different groupings. So I like it. It’s very simple. It’s very straightforward. The other big Pro that I love about the book is that it is what I like to call a practitioner’s book. It’s not a textbook so textbooks are typically more rigorous and academically driven, but this is going to be more of an industry practitioner book meaning. It’s more Hands-On. It’s more about learning the ideas and then learning how to code. And then actually utilize them to get results.
Textbooks on the other hand are usually more math and rigor driven and they cover really fine details but a lot of times those textbooks have absolutely zero coding in them. And so while you understand the math and the implementation really well on how to actually do it in different terms. There’s no code on how you actually Implement that into the real world. So this book has the label right, his hands on it is Stellar. It is amazing. I really really enjoyed reading this book. It’s a great book again for beginners to get their hands dirty.
The second person this could be for if there’s any three people here. The second person is going to be someone who focuses or as an expert more on statistical modeling or perhaps you have a lot of focus in one area of machine learning or data science. So let’s just hypothetically say, I don’t know if you’ve trained and studied your work and you will lead you neural networks and deep learning, but you’re wanting to do more machine learning.
like bringing boosting clustering algorithms, data processing and what not. This book is going to be for you because it’s going to give you enough information that you can leverage your statistics background or your other subject matter expert area and you’re gonna go with the leverage that to the point where you get a really solid foundation and understanding what’s going on again if you have a strong math and stats background, you should be able to connect the dots behind the scenes as in like a textbook kind of style fairly quickly because you’ll know the ideas at
your job so when you read this and explains how it all fits together. It makes a lot more sense. So that’s the second person. This book is good for the third person. This book is good for being the expert and this is not going to be a great book for learning if you’re an expert, but this is what I like to call a reference book. So let’s say again working machine learning took the work I don’t know. It’s a gradient boosting a random forest or even just like clustering for example, and you want to do something different. So you need to go back and you need to buff up on it, and I’ve read it in the past. Let’s say I have a masters or PhD. I’ve already done all the work. I’ve already read a lot, you know, but it’s been like five six years ago. I don’t quite recall the exact details. It’s an excellent reference book. We need you to just flip up and quickly get to the section. Reading the chapter you get the basic idea is refresh your memory and then dive in from there. So let’s talk about who this book is not for: this book is not for someone who already has a very deep understanding of data science and machine learning.
Yeah, some advantages here be if you’re not running in pythons. Let’s say you’re coming from the other side. This might be a good idea to help with the python coding is to put some python could help the tensor flow. For example, I’m setting things up and programming but in general it’s not going to be one of those books. That is very rigorous indeed. This is one of these books you start off with your first year and use it for many years and like practice. It’s the base foundation for you. And then as you specialize in focus in a career and specific tools that your area
Require you to go out and find other books to help give you more of that rigor and that deeper understanding of what’s going on behind the scenes. Okay, so the three books we’re going to compare here is a course on machine learning, but that was reviewing and this is more or less a practitioners book as I mentioned. The second book I’m going to mention is going to be deep learning with python. This is one of my other favorite books. again. This is more of a practitioner book, but it is specialized on neural networks and deep learning and then finally here we’re going to review
pattern recognition and machine learning by Christopher and Bishop. I haven’t reviewed this book for you guys, but I have read most of this book different parts of this book, right but to compare all three of these books to really show you the differences we’re going to just do neural networks and deep learning because of course one of these is specialized in it. Okay, so to start off with your hands on machine learning with scikit-learn curiosity tensorflow, we’re going to start off with you’re at the very beginning of artificial intelligence, which is Page 279.
He’s going to do an introduction. And so if you scroll through it’s going to be a lot of disparate verbiage discussions chatting. It’s going to be basic ideas talking about, you know, logic and computations and the perceptron and there’s a little bit you notice here. There’s a little bit of math but not really a lot. There’s some good diagrams here right talking about how it gets processed through the input layer is the output layers and again, if you look here there’s going to be a little bit of mathematical notation, but not a lot. We’re starting to get into the coding. So this is the big advantage of this book. It has python code. It’s going to use, you know, different numpy and different, you know libraries here.
and as we keep going through here again, there’s a lot of explanations and a lot of thoughts spent on just explaining what’s going on in trying to get the intuition behind what you’re doing, which is very good and like this in the market. again as we continue to scroll through you’ll see more code and you’ll see examples. I’ve done this example before and miss a fashion in and IST dataset category is different sets of clothing it gets it to be a great example to start with for image processing. And again, he provides a lot of code in layers and explanations. And so the book in general does a great job at getting you that Hands-On.
kind of nitty-gritty in the weeds feeling a bit. I can do this for another problem. So this book is a really good job at explaining things. It doesn’t a good job at providing code and realistically there are a lot a lot of material in this book just on neural networks in general, but I actually think if you’re running to neural networks the specifically so again you are going to specialize I think deep learning with python is a better book for this which I’ve done a book review and I’ll put a link in the description.
Below if you’re interested, but this book again. It does the coding but I think this book does a far better job at explaining line by line by line each set of code. So we flip open here. Let’s say you can see here. He has notes in all the code like you know, this starts from a great image with some noise and there’s another line of code. It says magnitude. and of each grade and update and then he talks about, you know, these steps are going to be run gradient descent for 40 steps. It’s just more detailed and more specified and again this points off into the book review here. There are specific books that you’re going to want just to really dive deep on one topic: hands on machine learning with scikit-learn Keras tensorflow 2.0 is a good book. It’ll get you started. You’ll get much of the same details, but I actually found this deep learning book far better by Charles area.
I just think it’s a well-written book. But again, he’s only focusing on deep learning and they’re all that works which is why I like this book better for neural networks. But again, you see here that scikit-learn is far far thicker. It’s like double the size and again, it’s going to cover a lot of topics. So it’s a much better book for an introduction and the next year.
We have pattern recognition and machine learning. This book again is more of a textbook and so like I mentioned there’s not going to be a lot of code or anything. Here and here we can scroll through and will start right off the bat here. You see we’re talking about feed-forward Network functions, and there’s just more math notation. It’s just more mathematical because this book is more or less intended to be used at a university and then if we scroll through like you see pictures and diagrams again trying to explain things in a well-put-together manner again, but as you go, there’s just more math.
There’s more explanations back with more mathematics is we continue to scroll through here. Just looking at a few different sections of this right you’ll see if for example when you get into like, you know is near back propagation and we flip over you really start seeing again. It’s just the same stuff over and over right we’re looking at mathematical equations. We’re taking derivatives here. So you’re going to need a math background to really understand this in to really grasp this again if you’re coming from a math background and you’re wanting to learn machine learning data science, write this book will be a really good book for you because it will come
From it at that map perspective. Now that being said, right you’re gonna have to learn how to implement this at some point as I’ve noted here. You’ve been looking through the book. There is zero code in this book. There’s no python. There’s no art. There’s no code. So this book again is very good for someone looking to get a very rigorous mathematical understanding of what’s going on behind the scenes. How would you optimize how to do this in a very professional setting but again, you have to buy a book so that you can really understand how to implement this in Python.
This is a good textbook in general and this hands on machine learning sci kit learning here is some tensorflow. For those that are just starting out that are beginning at it’s going to give you a little bit of theories of knowledge a little bit of math in the book is we saw but it’s really going to dive in on how to implement this in practice. So that’s why I gave this book a 4.5 and a half Stars again if you’re looking for a very academic rigorous mathematical book, it’s not going to be the book for you if you were to have a solid base. It’s not really going to be a big enhancement for you if you’re looking for a great reference book, I love it. I think it’s amazing that four out of five stars grateful for that.