Laroo-Pece Division, Gulu City, Northern Uganda
Uganda time
+256-779-712-990 Learner login Chat on WhatsApp
Lumina Africa Foundation Igniting Light, Inspiring Change.
Donate
Artificial intelligence - Advanced

Deep Learning and Neural Networks Explained

Understand how neural networks learn, from neurons and gradient descent to CNNs and transformers, then train, evaluate…

About this course

Deep learning powers the tools many of us use every day: phone cameras that find faces, voice typing, translation and AI chat assistants. Behind the mystery is a small set of ideas: simple units that multiply and add numbers, a score that says how wrong the answer is, and a method that nudges thousands of numbers in the right direction until the answer improves.

This advanced course explains those ideas in plain words with small worked numbers you can follow with a calculator. You will learn why networks overfit, how to split and prepare data, how to judge a model with accuracy, precision and recall, and how convolutional networks, attention and transformers fit together. You will then run real experiments in a free Python notebook, reuse a pretrained model through transfer learning, think about cost and energy, and deploy a small model.

The course ends with a capstone: you build, test and document your own image or text classifier on a problem from your community, school or business.

What you will learn

  • Explain how a neuron, a layer, weights, biases and activation functions combine to make a network
  • Calculate a forward pass, a loss value and one gradient descent step by hand on a tiny example
  • Describe the difference between training and inference and how epochs, batches and learning rate fit in
  • Detect overfitting with a validation set and reduce it with regularisation, dropout and early stopping
  • Prepare, split and label data properly and avoid data leakage
  • Evaluate a classifier with a confusion matrix, accuracy, precision, recall and F1 score
  • Explain convolutional networks, attention, tokens and transformers, and how large language models are built
  • Train and fine-tune a small model with Keras in a Google Colab notebook, using transfer learning
  • Estimate the compute, cost and energy of a project and deploy a small model responsibly

Course content

5 modules · 17 lessons · 6 h 10 min · 6 quizzes

Module 1: How a neural network works4 lessons

Build the core ideas with small numbers: neurons, activations, loss, gradient descent and the training loop.

  1. Neurons, weights and layers 20 min
  2. Activation functions and the forward pass 20 min
  3. Loss and gradient descent 25 min
  4. Backpropagation, batches and the training loop 20 min
  5. Quiz: 8 questions
Module 2: Training well: data, overfitting and evaluation4 lessons

Prepare data properly, spot overfitting, regularise, and judge a model with the right measures.

  1. Overfitting and the validation set 20 min
  2. Regularisation and tuning 20 min
  3. Preparing and labelling data 20 min
  4. Evaluating models: confusion matrix, precision and recall 25 min
  5. Quiz: 8 questions
Module 3: Architectures: images, sequences and transformers3 lessons

See how convolutional networks read images and how attention and transformers handle language.

  1. Convolutional neural networks for images 25 min
  2. Sequences, tokens and attention 20 min
  3. Transformers and how large language models are built 25 min
  4. Quiz: 8 questions
Module 4: Building models in practice3 lessons

Run experiments in a free notebook, reuse pretrained models, and plan for compute, cost and energy.

  1. Your first experiment in a Python notebook 25 min
  2. Transfer learning and fine-tuning 25 min
  3. Compute, cost and energy 15 min
  4. Quiz: 7 questions
Module 5: Deployment and capstone project3 lessons

Save and serve a small model responsibly, then plan, build and report your own classifier.

  1. Deploying a small model 20 min
  2. Capstone planning: define the problem and the data 20 min
  3. Capstone build, test and report 25 min
  4. Quiz: 8 questions
Final exam

28 questions. Pass with 70% or more to earn your certificate.

What you need

  • Completed Introduction to Machine Learning (machine-learning-introduction)
  • Completed Introduction to Programming with Python (python-programming-intro)
  • A computer with internet access and a free Google account for the notebook activities
  • Comfortable with basic school mathematics: multiplying, adding, averages and percentages

Further reading

More courses to explore

Intermediate 1 learner

Generative AI at Work: Prompting, Safety and Productivity

Use generative AI tools well and safely at work or school: write strong prompts, draft, summarise and analyse, check facts, protect private data, follow policy and build repeatable workflows.

What you will learn

  • How generative AI works
  • The art of prompting
  • AI for everyday work
  • 17 lessons
  • 5 h 35 min
  • Online, at your pace
View course

Advanced 1 learner

Introduction to Machine Learning

Learn how machines learn from data: prepare data, train and judge regression, tree, nearest-neighbour and clustering models with scikit-learn, avoid overfitting and bias, and finish a complete project.

What you will learn

  • What machine learning is and how to prepare for it
  • Predicting with simple models
  • Trees, forests and clusters
  • 17 lessons
  • 4 h 46 min
  • Online, at your pace
View course

To install on iPhone or iPad: tap the Share button in Safari, then choose "Add to Home Screen".