Neural Networks & Deep Learning

About This Series

Deep learning sits behind most of the artificial intelligence you use every day, from the assistant on your phone to the recommendations you scroll past. This series takes you the whole way there, starting from a single artificial neuron and building up to the Transformer architecture behind modern language models, with working code at every step. It is written to be read in order, and each chapter leads into the next.

The aim is to give you both the intuition and the practice. You will see the math where it matters, but every idea is also shown as something you can run, first from scratch in NumPy and then with Keras and TensorFlow.

What the Series Covers

The ten chapters are grouped into four parts that move from fundamentals to practical application:

• Foundations — what an artificial neural network is, and the activation functions that give it its power.
• How Networks Learn — backpropagation and gradient descent, then the optimizers that make training fast and stable.
• Core Architectures — convolutional networks for images, recurrent networks and LSTMs for sequences, and the Transformer and attention mechanism behind today's language models.
• Working with Models in Practice — transfer learning and fine-tuning, and how to tune the hyperparameters that decide whether a model works.

The best place to begin is the introduction to artificial neural networks.

What You Will Be Able to Do

By the end of the series you will understand how a neural network actually learns, be able to build and train one yourself, and know which architecture suits which kind of problem. You will also be ready to move from theory into applied work, such as building real systems on top of large language models in the RAG field manual.

Who It Is For

The series assumes you can read basic Python and are comfortable with simple algebra. Everything else, including the deep learning ideas and the framework code, is explained as it comes up. If you would like to brush up on the language first, the Python Programming series is a good starting point, and the Machine Learning series gives helpful background on the ideas neural networks build upon.