Skip to main content

Deep learning fundamentals

How neural networks work, from backpropagation to the techniques that make training stable.

Course Duration: 8h​

What You'll Learn​

  • Describe the computation of a neuron, then of a full layer
  • Follow the gradient through a network with backpropagation
  • Choose an activation function and understand its effects
  • Diagnose a training run that fails to converge
  • Apply dropout, batch normalization and early stopping appropriately

Prerequisites​

  • Python and NumPy
  • Notions of derivatives and gradients

Course Modules​

  1. From the perceptron to the multilayer network
  2. Activation functions: ReLU, sigmoid, tanh, GELU
  3. Forward pass and loss computation
  4. Backpropagation explained step by step
  5. Optimizers: stochastic gradient descent, momentum, Adam
  6. Weight initialization and vanishing gradients
  7. Regularization: dropout, L2 penalty, data augmentation
  8. Batch and layer normalization
  9. Learning curves: reading overfitting
  10. A first network trained end to end

Assessment and certificate​

The course ends with a 40-question exam covering every module. On success, a certificate of completion is issued; its number is verifiable on the platform.

Free courses, by contrast, end with a 5-question quiz and a preview of the certificate, without certification.