📄️ Deep learning fundamentals
How neural networks work, from backpropagation to the techniques that make training stable. 8 hours, applied projects, a 40-question exam and a verifiable certificate.
📄️ 1. Perceptron to network
Module 1 of the Deep Learning Fundamentals premium course: the exact computation of a neuron, why a single perceptron fails on XOR, and what depth really buys you.
📄️ 2. Activation functions
Module 2 of the Deep Learning Fundamentals premium course: why non-linearity is indispensable, a comparison of the usual activations, and rules for choosing per layer.
📄️ 3. Forward pass
Module 3 of the Deep Learning Fundamentals premium course: the matrix computation of a layer, batch processing, and choosing the loss function by task.
📄️ 4. Backpropagation
Module 4 of the Deep Learning Fundamentals premium course: the chain rule applied to a network, computing gradients layer by layer, and automatic differentiation.
📄️ 5. Optimizers
Module 5 of the Deep Learning Fundamentals premium course: batch size, momentum, adaptive learning rates, Adam and AdamW, and tuning the learning rate.
📄️ 6. Initialization and gradients
Module 6 of the Deep Learning Fundamentals premium course: why initialization decides convergence, Xavier and He initialization, vanishing and exploding gradients, residual connections.
📄️ 7. Regularization
Module 7 of the Deep Learning Fundamentals premium course: how dropout works and how it behaves at inference, L2 penalty, early stopping and data augmentation.
📄️ 8. Normalization
Module 8 of the Deep Learning Fundamentals premium course: how batch normalization works, its two regimes, its limits, and why layer normalization dominates transformers.
📄️ 9. Learning curves
Module 9 of the Deep Learning Fundamentals premium course: diagnosing a training run from loss curves, telling overfitting from underfitting and leakage, and the deliberate overfit test.
📄️ 10. First network
Module 10 of the Deep Learning Fundamentals premium course: a complete PyTorch project, from data preparation to saving the model, with the training loop annotated.
📄️ Recap and exam
Complete recap of the Deep Learning Fundamentals premium course: neuron, activations, backpropagation, optimizers, regularization and diagnosis, then the 40-question exam.