0%
Preparing your experience
English 日本語
0 Start Learning

Your Bag

E-learning materials

Deep Learning Fundamentals

Understand neural networks from the inside out and train your own models with a modern framework, from a single neuron to CNNs and transformers.

Material Overview
Deep Learning Fundamentals

Go from first principles to working networks. You will learn how neurons, layers, activation functions, loss functions, and backpropagation fit together, then build and train models with a modern framework. This E-Learning Material covers convolutional networks for images, sequence models and the attention mechanism behind transformers, plus regularisation, learning rate schedules, and transfer learning when your dataset is small.

Curriculum

Learning Path

Each level is enrolled separately. Work through them in order, or start at the level that matches your experience.

  1. Level 01

    Beginner

    Purpose

    Build an intuition for neural networks by working through what a single neuron and a small network actually compute.

    Outcome

    You can describe what happens inside a neural network during a single training step.

    What You'll Learn
    • Follow inputs, weights, bias, and activation through one forward pass
    • Understand loss as a measure of how wrong the network currently is
    • Train a small network and watch the loss fall as the weights adjust

    278 CREDS

  2. Level 02

    Intermediate

    Purpose

    Train real networks and learn the practical craft of making them converge.

    Outcome

    You can train a network on your own data and fix it when training goes wrong.

    What You'll Learn
    • Understand backpropagation conceptually and choose optimisers with intent
    • Set learning rate, batch size, and epochs and read the training curves
    • Apply dropout, normalisation, and early stopping to control overfitting

    394 CREDS

  3. Level 03

    Advanced

    Purpose

    Work with the architectures behind modern systems, from convolutional networks to attention.

    Outcome

    You can choose and adapt an architecture that suits your data instead of starting from scratch.

    What You'll Learn
    • Build convolutional networks for images and understand what the filters learn
    • Follow sequence models and the attention mechanism that replaced them
    • Use transfer learning to reach good results with a small dataset

    787 CREDS

  4. Level 04

    Expert

    Purpose

    Train at scale and deploy efficiently, where compute cost and inference speed decide what is possible.

    Outcome

    You can deliver a deep learning model that meets both accuracy and runtime requirements.

    What You'll Learn
    • Use mixed precision, checkpointing, and distributed training to manage large jobs
    • Compress models with quantisation, pruning, and distillation for deployment
    • Profile inference and cut latency where it actually matters

    1,053 CREDS