Questions and Answers
Meor Amer is this book available in epub format, or only pdf?
Hi Ruiz. It’s only in PDF right now. Because the book contains mostly visuals, it might be tricky to have an EPUB format that can give a good reading experience. But TBH I haven’t explored it in great detail, will definitely take a closer look.
Hello everyone, excited to be here! I’m happy to take any questions about the book, creating visuals, or anything in between! I’ll try to answer to my best ability.
In the meantime, here are:
1 - Further details about the book - https://gumroad.com/a/63231091
2 - A 40-page sample of the book - https://bit.ly/34Seg9l
Hi Meor, welcome 👋
Hi Alexey, thanks for having me here.
Hello Meor Amer. Thanks for taking the time here 🙂 In times of the rise of Explainable AI, where do you think you books contributes the most in this area?
Hi Tino. It’s my pleasure! Unfortunately, the book doesn’t cover Explainable AI. Where this book can contribute is by helping readers build a clear framework of understanding before diving deeper into further deep learning topics like Explainable AI.
Got it! Thanks 🙂
Hi Meor Amer, I really like your illustrations posted on LinkedIn!
I’m fascinated by visualization and the art of data storytelling, and I was wondering what makes your book different from other illustrated DL intro books?
Hi Gur, thanks for checking out my posts and for your kind words! Indeed, there are already a number of amazing illustrated DL books from top authors. Where I try to add value is making a book that helps the reader navigate the various concepts with the least friction. For example, the same dataset is used in all chapters so you have the same, consistent reference. And the math is kept to an absolute minimum because the main goal is to help the reader build a strong intuition first.
Great! Thanks, Meor Amer.
I’m looking forward to seeing more of your work :)
Hi Meor Amer. I’m interested in what you mention in intro in the sample as your initiative to get into ML: how is ML applied to prosthetics, and what is that enabling for people?
Hi Dustin. Yes, that’s what got me interested in this field. It’s been quite some time and the progress was much slower then. But now it’s really exciting with the emergence of companies like Elon Musk’s Neuralink. The use case possibilities are huge, stroke rehabilitation and amputee assistance being a couple of examples. Though we are still some way from practical and affordable solutions, it has definitely being given a boost in recent years.
Hello Meor Amer, appreciate your time to answer some questions.. I would like to know:
- Who is your target audience for this book?
- Are there any prerequisites needed to get the most out of the book?
- What the Book Doesn’t Cover?
- What are the most challenging lessons that you have learned while working on this book?
- What advice would you have for beginners in machine learning / deep learning?
Who is your target audience for this book?
People just beginning their journey into deep learning before going further into the technicals. Also leaders looking to understand deep learning from first principles.
Are there any prerequisites needed to get the most out of the book?
No pre-requisites. This book is beginner-friendly.
What the Book Doesn’t Cover?
Mathematical derivations, code examples, and further topics such as optimizers, regularization, embeddings, etc.
What are the most challenging lessons that you have learned while working on this book?
I’m big on ensuring a piece of material has an element of continuity and storyline. So the challenge was making sure all the technical concepts are covered while finding a way to make the reading enjoyable.
What advice would you have for beginners in machine learning / deep learning?
Start with hands-on projects, then reading the theory will make much more sense. The initial steps could be reusing existing code e.g. Kaggle notebooks, then modify and build upon what’s there.
Hi Meor Amer, what gave you inspiration to visualize deep learning concepts in the way you did in your book? Was it challenging to make those visualizations? What tools did you use?
Hi Benedict. My inspiration is the amazing designer Jack Butcher (Visualize Value). Ironically, what he’d taught me is you don’t have to be good at designing (that’s me!) to be able to come up with impactful visuals. It’s more about sharing your perspectives than worrying about the aesthetics. It is still challenging, but through practice, it gets easier. As for the tool, I use Figma - highly recommended!
Hi Meor Amer, really impressive to me and looking forward to your more outstanding books! Here are my questions:
- Beginners like me can easily understand it after reading with the visual intuition, however, after some days, some of the fundamentals may get vague in my mind, can we have some practice tips to consolidate the concept for readers?
- Do you plan to simplify the mathematic algorithms and other complicated stuff like that in ML/DL by the way of visualization in the book?
- What’s your next plan on the book, expanding the other knowledge domain related to ML/DL or go deeper of each part within ML/DL knowledge?
- Do you have plan to spread this type or series of books to more areas, like country as China where has big potential and a huge number of readers who have great interests on ML/DL knowledge?
- Can anyone get the fortune to get full copy If he can translate it into Chinese edition and try to boost its local publication? :)
- I can totally relate. That’s why sometime soon I’ll release the accompanying code used to run the examples in the book. And yes, further down the line I do have ideas for creating additional hands-on exercises.
- Yes, that’s my intention with the book - to help readers build an intuition first (by way of visuals) using the minimum possible math and equations.
- The next one I’m planning is A Visual Introduction to Machine Learning.
- Absolutely, I’m always open to potential collaborations to help make deep learning accessible for more!
- That’s an interesting prospect. Would love to explore this!
Hi Meor Amer, thanks for your response, looking forward to collaborate with you on the Chinese edition of the books🤝
How do you come up with visualization ideas?
Hi Alexey. Here are some that have worked for me:
- As for any generic ML visuals, it helps to observe for phenomenons/things outside of ML/Tech. Because a lot of things around us happen via similar universal principles, we can take one concept from one area and apply it to another.
- As for the book, first I built the storyline of how I want to book’s content to flow. From there, I listed down the key concepts of each page. And then I figured out the appropriate visuals to depict the idea.
What tools do you use for drawing?
I’m using Figma and I really love it!
Have you added concepts like receptive field and how it affects?
Hi Hafiz. Unfortunately that topic is not covered in the book.
Hi Meor Amer, I was wondering if your book also covers some intuition that helps to design and tune neural networks?
Hi Tim. Yes, the book covers these concepts. For designing, it covers the tweaks required in the neural network to address four types of tasks: linear regression, non-linear regression, binary classification, and multi-class classification. For tuning, it covers the common hyperparameters that can be used to improve performance.
Hey Meor Amer thank you for answering our questions.
1-How long did it take you to write the book? And which part of it was the most time intensive?
2-Where do you think the future of Deep learning is heading?
3-Is having knowledge in Machine learning necessary to understand and follow your book?
4-Do you have new visualization books ideas in your mind for future work?
5-How did your self learn deep learning?
Hi Ali. You are most welcome.
1-It took me around 3 months of almost full-time work. The most time-intensive is the continuous structuring of what to cover vs. not to cover to ensure that the whole flow makes sense.
2-Less labels and multi-modal i.e. getting closer to how humans learn.
3-The book is written with the beginner in mind. So the good news is you don’t need any background in ML to read it!
4-Yes, I do. The next I have planned is A Visual Introduction to Machine Learning. It will be more challenging to write though because it’s broad whereas deep learning is specific.
5-Initially by working on domains I am familiar/interested in e.g. telecommunications, biomedical engineering. Combined that with theory learning like courses, books, etc.
Hi Meor Amer ! Thanks for answering questions about your book. I would like to ask your opinion on who would have the most gain from this book? An audience with no prior exposure to deep learning, or an audience who is comfortable with it, or somewhere in between? I’d appreciate if you could elaborate. Thank you in advance!
Hi Roy. This book is beginner-friendly and so it’s fine for those without prior exposure to DL. If you are already comfortable with it, then probably you won’t gain as much new information. But perhaps the visuals can still help give different perspectives to what you already know.
Meor Amer I have gone through your sample pdf. there you said your journey started at 2010 after your son born with a limb difference. how did AI helped in prosthesis?
Hi Meghana, thanks for going through the sample. Yes, that’s how I got started. It’s been some time and I’d say the progress hasn’t been as fast as people would hope it to be. Viable consumer applications are still not there yet. Having said that, I’m excited that companies like Neuralink and Kernel are bringing these ideas more into the mainstream, which help to bring more talents into this field and accelerate the progress.
Thanks Alexey for inviting me here. It’s been so much fun interacting with you all and I appreciate the time you took to ask questions and share comments.
This community is amazing with phenomenal growth, and I’m sure this is just the start!
BTW, I will keep the 40% book discount over the weekend if you are interested to get it. Use the discount code: DATATALKS40