Key terms

Question 1

What is supervised learning?

Question 2

What is unsupervised learning?

Question 3

What is unsupervised learning?

Question 4

What is an example?

Answer 4

A feature-label pair (sometimes, when the context is clear, we may use the term examples to refer to a collection of inputs, even when the corresponding labels are unknown)

Question 5

types of supervised learning tasks (5)

Answer 5

1. Regression
2. Classification
3. Searching
4. Recommending
5. Sequence learning

Question 6

What is the difference between regression and classification?

Answer 6

Regression predicts a continuous value whilst classification predicts a categorical outcome

Question 7

What is optimisation?

Answer 7

the process of fitting the model to our observed data by altering the model parameters

Question 8

What is the value produced by the loss function?

Question 9

How is a particular model parameter updated after calculating the loss?

Answer 9

The original value subtract the derivative of the loss with respect to the model parameter

Question 10

What is the gradient vector?

Answer 10

A vector that contains all the partial derivates

Question 11

In order to calculate the gradient with respect to some input features, what must the output be?

Answer 11

A scalar

Question 12

Common regression problem examples (3)

Answer 12

1. Predicting prices
2. Predicting length of stay
3. Forecasting demand

Question 13

What is the bias in linear regression?

Answer 13

a value that represents the value of y when all inputs are 0

Question 14

What do the weights represent in linear regression?

Question 15

What is gradient descent?

Answer 15

an optimisation algorithm that iteratively reduces the error by updating the model parameters in the direction that incrementally lowers the loss function

Question 16

Consequence of a learning rate which is too small?

Answer 16

the optimisation process may converge very slowly

Question 17

Consequence of a learning rate which is too large?

Answer 17

the optimisation process may become unstable

Question 18

Output dimensions of Xw? (linear regression)

Answer 18

n x 1 (one output for every observation/example)

Question 19

What is the purpose of the softmax operation?

Answer 19

convert them to probabilities

Question 20

What is the purpose of activation functions?

Answer 20

to introduce non-linearity into the model, enabling it to learn and represent complex relationships in the data

Question 21

What exactly do activation functions do?

Answer 21

Decide whether whether a neuron should be activated or not by calculating the weighted sum and further adding bias to it.

Question 22

Commonly used activation functions (3)

Answer 22

1. ReLU
2. Sigmoid
3. Tanh

Question 23

What happens if a neuron is not activated?

Answer 23

the neuron does not send its output signal to neurons in the next layer of the neural network

Question 24

What is generalisation?

Answer 24

the ability of a trained model to perform well on new, unseen data that it hasn’t encountered during training

Question 25

What are convolutional neural networks?

Question 26

Why are fully connected layers not suitable for images?

Question 27

What is spatial invariance?

Answer 27

a property in convolutional neural networks (CNNs) where the network learns to recognize patterns or features in the input data regardless of their specific spatial location within the input image.

Question 28

What is the channel dimension?

Answer 28

The number of channels for an image?

Question 29

What is the most popular example of sequence data?

Answer 29

Text