Phase 1 builds the mental model you’ll reuse for every architecture in this module — CNNs, RNNs, Transformers, even generative models. They’re all just neurons, layers, and activation functions arranged differently.

You’ll learn

• What a tensor is - scalars, vectors, matrices, and beyond - plus the element-wise ops, broadcasting, and dot products every layer is built from
• What a biological neuron inspired, and how McCulloch, Pitts, and Rosenblatt turned that idea into the perceptron
• Why a single perceptron can only draw a straight line through your data (and why that’s a problem)
• How stacking perceptrons into a Multi-Layer Perceptron (MLP) fixes that
• What activation functions (ReLU, sigmoid, softmax) do, and why a network without them is secretly just a linear model
• How to describe and build these networks with tf.keras - the Sequential, Functional, and Subclassing APIs, plus saving, loading, and callbacks
• How gradient-based optimization actually works - derivatives, gradients, the chain rule, backpropagation, and a tiny GradientTapeGradientTape example
• Three complete first projects: classifying movie reviews (binary), newswires (multiclass), and predicting house prices (regression)

  flowchart LR
  A["Tensors & tensor ops"] --> B["Perceptron (1 layer of TLUs)"]
  B --> C["Multi-Layer Perceptron (hidden layers)"]
  C --> D["Activation functions (ReLU, sigmoid, softmax)"]
  D --> E["Building models with Keras APIs"]
  E --> G["How networks learn (gradient descent)"]
  G --> H["First examples: IMDB, Reuters, house prices"]
  H --> F["Ready for Phase 2: training"]

Phase checklist

1. Tensors and Tensor Operations
2. Introduction to Neural Networks (The Perceptron)
3. Multi-Layer Perceptron (MLP)
4. Activation Functions (ReLU, Sigmoid, Softmax)
5. Building Neural Networks with Keras (Sequential and Functional API)
6. How Neural Networks Learn (Gradient-Based Optimization)
7. First Example: Classifying Movie Reviews (IMDB, Binary)
8. First Example: Classifying Newswires (Reuters, Multiclass)
9. First Example: Predicting House Prices (Regression)

Prerequisites

• You’ve finished the Machine Learning module (or are comfortable with train/test splits, loss functions, and gradient descent).
• Comfortable with basic NumPy and Python functions.