Book Review: TinyML Cookbook
When I was offered to review a book named “TinyML Cookbook” I was excited, although I did not have a clear idea of what the book was about. Was this a tiny book filled with great recipes for Machine Learning? Did it have tiny snippets for quick ML? Was it directed a teaching robot arms how to cook? — By the way, the last one is actually happening in case you were wondering, here is an interesting video about a 5-star robot chef.
The world is filled with surprises, hence when I started reading the book, I ended up discovering a whole new world where Machine Learning is being applied that was not even in the farthest corners of my periphery. Turns out Tiny ML refers to using embedded Machine Learning directly on low powered devices, yes, like those tiny computers that now make up tons of products from smartwatches to smart umbrellas (in case you were wondering, yes, smart umbrellas do exist).
And here comes the Tiny ML Cookbook by Gian Marco Iodice, bringing us a great book filled with recipes for adding Artificial Intelligence to the growing world of IoT.
What is the book about?
Since this is a book review, and not an article on Tiny ML, I will only give you a small example of it´s application and why it is an interesting and promising field. Many of us are familiar with the Google Assistant or Alexa, and there are many devices, from your smartphone to smart speakers, that connect to these services. Recognizing the magic activation words “Ok Google” or “Alexa” is an actual Machine Learning application. Some of the connected devices, like the Google Nest Product line or the ton of Alexa powered devices, are tiny, and yet, they need to run the magic word recognition locally, otherwise, they would have to take up a ton of bandwidth all the time and it would be very messy on the server end.
Here is where Tiny ML comes to the rescue, it is all about how to implement ML applications in small IoT devices and discovering the opportunities and challenges this brings with it.
And this is what the book is all about, how to implement ML into tiny low powered embedded devices. So, if you have been wondering how to get your smart devices to do more by applying ML, this is the book for you.
A small recommendation, if you want to get the most out of this book, make sure you have the hardware, like Arduino or Raspberry Pi systems (and some components like sensors and the like).
Still, if you do not have said hardware, do not fret, you will still get a lot from this book (but if you can, do get some sweet Arduino or Raspberry gear).
Is the book well written?
Yes, very well. It explains the concepts clearly and comes with clean, understandable schematics for all the hardware, clean explanation drawings for the circuitry and easy to follow coding examples.
In short, it is easy to follow every step of the way, even if you have little experience with IoT devices. Very importantly, this is a book directed at people familiar with Machine Learning concepts, and although it does a good job of explaining them, it might be challenging for people who do not have at least some basic knowledge of it.
Does it present the material well?
One word yes. The drawings, graphics and coding examples are clear and well presented. It is easy to go from writing to drawing, to doing, back to drawing and back to reading without losing yourself most of the time. And, although it has a ton of content, it never feels overwhelming like many technical books tend to.
It is quite simple to read, especially if you understand the concepts, and it is probably even easier if you are familiar with Arduino and/or Raspberry Pi devices (I am not).
Is it easy to read?
Once again, this one is hard to answer. I am leaning towards yes, not just because it was easy for me to read, but because even while I am not very familiar with IoT devices, or hardware in general for that matter, I found that I never struggled to understand and the needed information to grasp ant of the concepts was right there.
I do believe the book can be challenging for people without at least some background on Machine Learning, but I do recognize this might just be my own prejudice from my perspective.
Do you need to be an expert to read this book?
Here I lean towards no, you do not. Although I believe the audience for this book are ML experts looking to apply their craft on IoT devices, truth is, I was able to grasp the concepts for the hardware of these devices without much of a problem, hence, probably anyone might start getting into either IoT or ML through this book.
So, even if I mentioned some people might struggle with the ML side of this book (again, it might be my own bias), do not mind me that much, after all, we all must start somewhere, and this book can be an interesting starting point.
How much will you learn from this book?
A LOT… PERIOD. The two subjects this book brings together are very broad. IoT is huge, and the devices covered are incredibly versatile. On the other side, ML is a massive subject, and yet, Gian Marco is able to bring us some creative solutions to applying complex algorithms (including Deep Learning) into small and very limited devices when it comes to processing.
Even if you have no intention of working with IoT devices, you will learn how to tackle ML problems dealing with processing limitations.
And of course, if you know nothing of IoT, you will learn a lot about it.
Is it easy to reference?
Yes. I will repeat myself: Packt books have a very comprehensive index, and the way it is structured makes it so you can find exactly what you are looking for with a quick look at it and some page flipping. Meaning, going back to any part of the book is fast and easy.
In short, if you are an IoT expert that wants to improve your game by adding ML, get this book. If you are an ML expert that wants to get into TinyML, this is a great book for you. If you are neither and want to learn about both, go for it.
You will enjoy and get a lot out TinyML Cookbook.
The book is already out and you can get it by clicking right here.
NOTE: Packt did send me a free copy of the book in advance to be able to write this review. I do not get anything else, either for this review or book sales. I do get a sense of happiness for having been part of this experience, and to be able to help great authors like Gian Marco Iodice and his book to get noticed so they can keep up their amazing work.
