II: Machine Learning

Question 1

Define Machine Learning.

Answer 1

A computer is said to learn from experience E with respect to a class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E.

Question 2

What’s the difference between top down and bottom up?

Answer 2

Bottom up: give the system lots of data to discover for itself
Top down: model all different functions and wire these together, and write manual instructions for solving a task

Question 3

Is machine learning top down or bottom up?

Answer 3

Bottom - up

Question 4

What’s the process of using machine learning to reach AI?

Answer 4

1. analyse the task manually
2. select appropriate model/s
3. Training and testing
4. Use the same model to solve unseen tasks

Question 5

What are the 3 pillars of machine learning?

Answer 5

1.Models and algorithms
2.Computation (powerful and cheaper)
3. Massive data warehouse

Question 6

Explain the model and algorithm pillar of machine learning.

Answer 6

Doesn’t require knowledge engineers to find all functions and ‘wire’ them in

Question 7

Explain the data pillar of machine learning.

Answer 7

We have massive amounts of data and the more we can feed into machine learning the more successful it will be.

Question 8

What is data mining?

Answer 8

Exploration and analysis of large quantities of data to discover valid, novel and useful patterns that are understandable

Question 9

What is a big difference between data mining and machine learning?

Answer 9

Data mining gives understandable patterns so it can be explained why a conclusion has been come to as opposed to through machine learning

Question 10

What are the two types of learning in machine learning?

Answer 10

Supervised and Unsupervised Learning

Question 11

Top-down and bottom-up are also known as?

Answer 11

Deduction (top-down)
Induction (bottom-up)

Question 12

Describe supervised learning.

Answer 12

Learns from labelled data which consists of an input and a desired output

Question 13

Describe Unsupervised learning.

Answer 13

Learns from unlabelled data to describe a hidden pattern/structure
e.g. to reach a model or regression or grouping

Question 14

Name three types of supervised learning.

Answer 14

Linear regression, decision tree, ANN

Question 15

What is reinforcement learning? What is the goal?

Answer 15

Learning over time based on trial and error / awards from actions. Goal is to learn a process to maximise reward

Question 16

What are some examples of reinforcement learning?

Answer 16

Deep neural networks, Q-learning, deep Q networks

Question 17

What are the two main differences between supervised and unsupervised learning?

Answer 17

• supervised has a relationship between Y and f(x), unsupervised usually doesn’t.
• supervised uses a labelled data set, unsupervised uses an unlabelled set

Question 18

In data cleaning: what is the meaning of imputation?

Answer 18

Replacing missing values with realistic estimated values

Question 19

What is clustering?

Answer 19

The process of grouping unlabelled data based on a similarity or distance metric

Question 20

What’s the point of a test data set?

Answer 20

To check how well a trained data set performs

Question 21

What is regression training?

Answer 21

Finding a function based on given data that minimises its mean squared error to ‘fit’ the samples

Question 22

What is overfitting?

Answer 22

Fitting a model tightly to a small data set which increases the risk of fitting to an anomaly or noise data

Question 23

What are 3 pros of linear regressions models?

Answer 23

• short training time
• easy to interpret
• easy to implement

Question 24

What are two cons of linear regression models?

Answer 24

• sensitive to noise and outliers
• cannot handle complicated relationships

Question 25

What is classification learning?

Answer 25

Learning to predict which set an instance belongs to based on pre-labelled instances

Question 26

In the decision tree model, what is an internal node?

Answer 26

A decision point where a test is made on an attribute of the data and a decision is made

Question 27

In the decision tree model, what is a branch?

Answer 27

The course of a decision or action

Question 28

In the decision tree model, what is a leaf node?

Answer 28

A predicted class label / output reached through a path of decision points (internal nodes) and branches

Question 29

What are 4 pros of decision tree?

Answer 29

- Reasonable training time - Can handle large number of features - Easy to implement - Easy to interpret

Question 30

What are 4 cons of decision tree?

Answer 30

- Only simple decision boundaries - Problems with lots of missing data - Cannot handle complicated relationships - Over- complex tree: potential for overfitting

Question 31

When we only have one dataset, why do we need to split it?

Answer 31

To have data for training and data for testing

Question 32

What stage happens before the training set is fed into the machine learning method?

Answer 32

Data pre-processing

Question 33

What is a confusion matrix?

Answer 33

It represents how much data is predicted to each outcome and how many of those are predicted correctly and incorrectly

Question 34

What is the concept of perceptron in Neural networks?

Answer 34

There is a set of weighted inputs that if they surpass a certain amount, it causes a set of outputs.

Question 35

What is the differing meaning of positive and negative weights for neural network inputs?

Answer 35

- a positive weight encourages the neuron to fire as a result of that input - a negative weight discourages the neuron from firing as a result of that input

Question 36

What are the 4 logic functions?

Answer 36

- Or - And Not

Question 37

What does the logical OR function do in ANNs?

Answer 37

one or other of the inputs needs to be one

Question 38

What is the logical function 'And Not' in ANN?

Answer 38

A AND NOT B means A must have a value of 1 and B a value of 0