Chapter 6 Quiz: Perceptrons & Neural Networks

Chapter 13: From Perceptrons to Deep Networks

Learning objectives

Test your understanding of perceptrons, activation functions, and building neural networks

This quiz covers both the lecture material and lab exercises from Chapter 6.

Perceptron: $z = \mathbf{w}^T\mathbf{x} + b$ , output = step( $z$ ). Only linearly separable functions.
Activations: ReLU $= \max(0, z)$ ; sigmoid for binary output; softmax for multi-class.
Backpropagation: Forward pass $\to$ loss $\to$ backward pass (chain rule) $\to$ update weights.
Parameters: Dense( $n_{in}, n_{out}$ ) has $n_{in} \times n_{out} + n_{out}$ parameters.
Vanishing Gradient: Sigmoid derivatives < 1 cause gradients to shrink. ReLU helps.

Goodfellow, I., Bengio, Y., Courville, A. (2016). Deep Learning, ch. 6. MIT Press.
LeCun, Y., Bengio, Y., Hinton, G. (2015). Deep learning. Nature 521, 436–444.
Bishop, C.M. (2006). Pattern Recognition and Machine Learning, ch. 5. Springer.