1. Neural Network Layer

• Each neuron is a logistic regression unit.
• The superscript with square brackets represent the n-th hidden layer.
• The input to layer 2 is the output from layer 1.
  
  
• The final output can be thresholded for categorical data.

2. More Complex Neural Networks

• subscript for j-th logistic unit,
• square bracket superscript for l-th hidden layer activation values.
• activation function is another name for these logictic regression functions.

3. Inference: Making Predictions (forward Propagation)

• Making computations from left to right and forwarding the results.
  • contrasted with backward propagation.
• Number of hidden units decreases as we get closer to the output layer - common architecture.