Deep Learning

Question 1

What is the biological inspiration for Deep Learning?

Answer 1

It is inspired by the function and structure of the brain.

Question 2

What are two types of Deep Learning?

Answer 2

Layered Representations Learning and Hierarchical Representations Learning

Question 3

What are the three types of layers in a Deep Learning Model?

Answer 3

Input Layer, Hidden Layer, Output Layer

Question 4

What is the Loss Function?

Answer 4

A function used to calculate how accurate the output of a model is.

Question 5

Which Loss function is used for Binary output?

Answer 5

Binary Cross Entropy

Question 6

What is the result of the loss function called?

Answer 6

Loss Score

Question 7

How is the Loss Score used?

Answer 7

The loss score is fed into an optimizer function and used to update the weights.

Question 8

What is the function of the input layer?

Answer 8

To receive the data.

Question 9

What is the function of the Hidden Layers?

Answer 9

To extract features form the images of increasing complexity.

Question 10

What is the function of the output layer?

Answer 10

To generate probabilities for the specified classes.

Question 11

What is backpropogation used for?

Answer 11

Backpropogation is used to find a set of weights that minimises the error of the model.

Question 12

What does SGD stand for?

Answer 12

Stochastic Gradient Descent

Question 12

Why is an activation function used?

Answer 12

An activation function is used to cope with complex real tasks that require non-linearity.

Question 12

How does SGD work?

Answer 12

SGD updates the weights of the whole network by looking at every example in the training set.

Question 13

What is a Hyperparameter?

Answer 13

The specifications of the Architecture of the model

Question 14

What are Epochs?

Answer 14

The number of time the training samples pass through he network

Question 15

What is the Batch Size?

Answer 15

Partitions of the traning data that pass through the network.

Question 16

What is Overfitting?

Answer 16

An issue that arises where a model becomes to specific and memorises the correct classification for the training data, and is unable to correctly classify the actual data being used.

Question 17

Why is Dropout used?

Answer 17

It is used to overcome overfitting

Question 18

How does Dropout work?

Answer 18

It randomly drops connections that have been created in order to force generalisation rather than memorisation

Question 19

What are the three main Loss Functions?

Answer 19

Mean Squared Error, Binary Cross Entropy, Logarithmic Cross Entropy.

Question 20

What are the two sections contained in a Convolutional Neural Network?

Answer 20

Feature Extraction and Classification

Question 21

What are the three layers contained within the feature extraction on a CNN?

Answer 21

Input, Convolution, Pooling

Question 22

How does a Convolutional Layer extract features?

Answer 22

A square filter is passed over the image and the dot product of the features contained in the filter is taken.

Question 23

What is the advantage of a convolutional layer?

Answer 23

Convolutional Layers retain information about the spatial relationship between the pixels in the image, whereas previous models have converted everything into vectors which loses spatial knowledge.

Question 24

What is a pooling layer?

Answer 24

The third layer in the feature extraction section it decreases the size of the feature map to reduce computational cost.

Question 25

What are the two types of pooling?

Answer 25

Max Pooling and Average Pooling.

Question 26

What is Dimensionality Reduction?

Answer 26

Representing Multi-Dimensional data in a lower dimension.

Question 27

What is the Curse of Dimensionality?

Answer 27

Sample Density decreases exponentially with in the increase in dimensionality.

Question 28

What is an Autoencoder?

Answer 28

A types of unsupervised ANN that aims to learn representation of the data in a way that captures the essential features, reducing the dimensionality of the data.

Question 29

What is the Objective Function used for in an Autoencoder?

Answer 29

It measures the difference between the Input data and its reconstruction and aims to minimise this difference.