Deep Learning Neural Network Fundamentals 2

Question 1

What is backpropagation in the context of neural network training?

Accepted Answer

Algorithm to compute gradients via the chain rule

Answer

Forward pass through the network

Answer

Method to initialize network weights

Answer

Technique to prune unused neurons

Question 2

Which loss function is most commonly used for multi-class classification problems?

Accepted Answer

Categorical Cross-Entropy

Answer

Mean Squared Error

Answer

Binary Cross-Entropy

Answer

Hinge Loss

Question 3

What is the universal approximation theorem?

Accepted Answer

A single hidden layer with enough neurons can approximate any continuous function

Answer

Any network can approximate any function with infinite layers

Answer

Deep networks always outperform shallow ones

Answer

Neural networks converge to global optima given enough data

Question 4

What does the learning rate hyperparameter control during gradient descent?

Accepted Answer

The size of the step taken in the direction of the negative gradient

Answer

The number of training epochs

Answer

The number of neurons per layer

Answer

The ratio of training to validation data

Question 5

What is a perceptron in neural network terminology?

Accepted Answer

A single artificial neuron that computes a weighted sum and applies a threshold

Answer

A multi-layer network

Answer

A type of convolutional filter

Answer

A normalization technique

Question 6

Which of the following best describes a hyperparameter in deep learning?

Accepted Answer

A configuration set before training that controls the learning process

Answer

A weight learned during training

Answer

An output of the final layer

Answer

A gradient computed during backpropagation