Course Modules

Module 1

Introduction to Neural Networks

What neural networks are, how artificial neurons, weights and layers work, and your first network in Python.

Learn

Module 2

The Perceptron and the Artificial Neuron

The first neural network: weights, bias, the learning rule, the XOR problem, and a perceptron built from scratch.

Learn

Module 3

Activation Functions

Why networks need them: sigmoid, tanh, ReLU and softmax, vanishing gradients, and which one to choose.

Learn

Module 4

Loss Functions and Gradient Descent

How networks measure mistakes and learn from them: MSE, cross-entropy, learning rates, mini-batches and Adam.

Learn

Module 5

Backpropagation

The algorithm behind deep learning, explained with a worked example by hand and a network that learns XOR.

Learn

Module 6

Building Networks with Keras

Build, train, evaluate and save neural networks in a few lines of Python with TensorFlow and Keras.

Learn

Module 7

Training Better Networks

Beat overfitting with early stopping, dropout, L2, data augmentation and batch normalisation.

Learn

Module 8

Convolutional Neural Networks (CNNs)

How computers see: filters, convolution, pooling and CNN architecture, with a CNN built in Keras.

Learn

Module 9

Recurrent Neural Networks and LSTMs

Networks with memory for text and time series, LSTM gates, GRUs, and an energy demand forecaster.

Learn

Module 10

Transformers and Attention

Self-attention, queries, keys and values, and how large language models like ChatGPT and Claude work.

Learn

Module 11

Transfer Learning and Pretrained Models

Reuse powerful pretrained models like MobileNetV2 with feature extraction and fine-tuning.

Learn

Module 12

Capstone Project: CNN Defect Detection

An end-to-end deep learning project: build a CNN that spots faulty parts on a factory production line.

Learn

CX Learn Certification Exam

Earn all 12 module stars to unlock the final 30-question certification exam. Pass the exam to receive your official verified CadetX Neural Networks and Deep Learning Certificate.

Requires 12 more stars

What you'll learn in this deep learning course

By the end of the course, you'll understand how modern AI models learn, and be able to build, train and evaluate your own neural networks.

Explain how artificial neurons, weights, layers and activation functions work

Understand how networks learn with loss functions, gradient descent and backpropagation

Build, train, save and use neural networks in Keras and TensorFlow

Prevent overfitting with dropout, early stopping, L2 and data augmentation

Build convolutional neural networks (CNNs) for image classification

Forecast time series with recurrent neural networks and LSTMs

Explain self-attention, Transformers and how large language models work

Use transfer learning to build on powerful pretrained models

Evaluate deep learning models with confusion matrices and cost-based thresholds

Complete an end-to-end deep learning project for your portfolio

How the course works

Every module follows the same simple pattern, so you always know what comes next.

Read

Short lessons in plain English, with everyday analogies and clear diagrams.

Try

Interactive labs: tune a neuron's weights, watch gradient descent and backpropagation step by step, slide a filter across an image, or explore attention.

Code

Real NumPy and Keras code for every topic, including networks built from scratch, with the real output explained.

Check

A 5-question quiz at the end of every module, marked instantly.

Prove it

Finish the capstone project and pass the exam to earn your certificate.

Who this neural networks course is for

Students and graduates

Preparing for AI, data science or machine learning roles, or building a deep learning project for your CV.

Machine learning learners

Comfortable with classic machine learning and ready for the next step: deep learning.

Developers

Wanting to understand and build the AI models behind image recognition, forecasting and chatbots.

Curious beginners

Wanting to know how ChatGPT, face unlock and self-driving research actually work, without heavy maths.

Why learn neural networks and deep learning?

Neural networks power the biggest breakthroughs in artificial intelligence: large language models like ChatGPT and Claude, face unlock on your phone, voice assistants, instant translation and medical scan analysis. Deep learning skills are in high demand across the UK, India and beyond, for roles like machine learning engineer, AI engineer, computer vision engineer, NLP engineer and data scientist.

This course builds real understanding first, so the tools make sense. You'll build a perceptron and a backpropagation network from scratch in NumPy, then move to Keras and TensorFlow, the industry-standard tools for building deep learning models. The ideas transfer directly to PyTorch and to working with pretrained models from Hugging Face.

New to machine learning? Start with our free Machine Learning course, which covers the foundations this course builds on: training and test data, overfitting and model evaluation.