Artificial intelligence is everywhere now, but for most people it still feels like magic. My new book, Neural Networks Using MATLAB: Deep Learning from Scratch for Beginners, with Every Line of Code Explained, takes the magic away. You build a neural network yourself, one small piece at a time, until you can see exactly how a computer learns.

Why I wrote this book
I recorded a lecture on neural networks for my YouTube channel, and the response surprised me. Every day I receive emails from viewers. They ask for a full course, for one-to-one help, or simply: “How do I learn more?”
This book is my answer to all of those emails. It is the complete course I could not fit into one video, written down step by step, so anyone in the world can learn at their own pace.
The problem with most deep learning books
Many books start with pages of heavy mathematics. Others hide everything inside a ready-made toolbox. You copy the code, it runs, and you still cannot explain why it works. That feels like learning, but it is not.
This book does the opposite. You use plain MATLAB, with no toolbox and no shortcuts. You write every part of the network yourself, so nothing is a black box.
How every chapter teaches
Each idea follows the same gentle path, the same one I use in my classroom:
- An everyday picture first. Before any formula, you see the idea in a way that makes sense.
- Then the equation. Short, simple and explained in words.
- Then the code, line by line. Every single line is explained. Nothing is skipped.
By the end of each chapter you will know why a network works, not just how to run it.
What you will be able to do
- Train your first artificial neuron, and see why it needs hidden layers
- Build multi-layer networks that learn using back-propagation and momentum
- Sort data into categories with Softmax and cross-entropy
- Train deeper networks with ReLU and dropout
- Build convolutional neural networks (CNNs) that recognise images
- Find hidden patterns in data with autoencoders
- Teach a computer agent to make decisions with reinforcement learning and Q-learning
- Test a model honestly, using validation sets, confusion matrices and early stopping
Seven real projects
Reading about AI is not enough; you learn by building. The book includes seven hands-on projects drawn from real life, from detecting tumours and recognising cars to recommending films and guiding a warehouse robot.
Made to be friendly
- Nearly 200 figures and tables turn abstract ideas into pictures you can follow.
- Review questions at the end of every chapter let you check your understanding before moving on.
- Free companion code on GitHub lets you run every example yourself.
- No toolbox and no advanced maths, just plain MATLAB.
Who this book is for
It is for anyone who wants to understand AI from the inside: students, self-taught programmers, engineers and researchers. You do not need any earlier experience with deep learning. Teachers can also use it for university courses on machine learning, since it moves in a clear order and every chapter ends with questions.
New · November 2026
Ready to see how deep learning really works? Train your first neural network this week.
Get the book on Amazon →Common questions
Do I need to be good at maths to read this book?
No. Every idea starts with an everyday picture, then a simple equation, then the code. You only need school-level maths and a willingness to try.
Do I need the MATLAB Deep Learning Toolbox?
No. Everything is built in plain MATLAB, with no toolbox. That is the point: you write each part yourself, so you understand it.
I have never studied deep learning. Is this book for me?
Yes. It was written for complete beginners: students, self-taught programmers, engineers and researchers. No earlier experience with deep learning is needed.
Can I run the code myself?
Yes. All the example code is free in the book’s GitHub repository, so you can download it and run every example on your own computer.
Can teachers use it for a university course?
Yes. Each chapter ends with review questions, and the book follows a clear order from a single neuron to reinforcement learning, so it fits a semester course on machine learning.
Where can I buy it?
The paperback is on Amazon worldwide, including amazon.com and amazon.com.au.
About the author
I am an Assistant Professor at Daffodil International University and a doctoral researcher in Data Science and AI at Monash University. I have taught AI and machine learning since 2018, received the Excellence in Teaching Award in 2023 and 2024, and this is my sixth book. You can see my other books on the Books page and my courses on the Teaching page.
Published 9 November 2026 · Paperback · 328 pages · ISBN 979-8141115646