Machine Learning

Question 1

Machine Learning

Answer 1

Programs that can improve their performance with training data via a learning algorithm.

Question 2

Learning benefit instead of deterministic outputs

Answer 2

• Designer cannot anticipate all possible situations.
• Designer cannot anticipate changes.
• Designer does not know the answer/ how to program the answer.

Question 3

Applications

Answer 3

• Facial detection
• Speech recognition
• Stock prediction
• Digit recognition

Question 4

Representing Instances (Feature vectors)

Answer 4

For example, mushrooms:
x^(1) = <bell, fibrous, …> y^(1) = edible
x^(2) = <convex, scaly, …> y^(2) = non-edible

Question 5

Feature Types:

Answer 5

Nominal/Boolean -> no ordering
Ordinal -> possible values are totally ordered
Numeric -> weight, height etc
Hierarchical -> ordered via hierarchy

Question 6

Feature Space

Answer 6

Way of representing features and the distributions on where the vectors fit. Can be via 3D models or databases etc

Question 7

Data Preprocessing

Answer 7

Techniques that make it easier for data to be used for analysis by machine learning models.

Question 8

Mean Normalisation

Answer 8

Remove mean from every data sample.

Question 9

Standardisation / Normalisation

Answer 9

Requires all features to be on the same scale.

Question 10

I.I.D

Answer 10

Independent and identically distributed.
We assume data collected is sampled independently from the same unknown distribution.

Question 11

Supervised Learning

Answer 11

• Set of instances X
• Unknown target function f : X -> Y
• Set of models H = {h |h : X -> Y}
• Set of training instances (x(1), y(1)),(x(2),y(2)),…,(x(z),y(z))
• Chosen model should most accurately represent the target function X -> Y

Question 12

Regression

Answer 12

A supervised learning technique which estimates parameters (used when Y is continuous)

Question 13

Classification

Answer 13

A supervised learning technique which estimates classes (used when Y is discrete)

Question 14

Unsupervised Learning

Answer 14

• Set of instances without y’s -> x(1), x(2),…,x(z).
• Goal is to discover patterns and regularities.

Question 15

Clustering

Answer 15

An unsupervised learning technique which maps data into different clusters.

Question 16

Dimensionality Reduction

Answer 16

An unsupervised learning technique which lowers dimension features whilst preserving key features.

Question 17

Semi-supervised Learning

Answer 17

Using labelled as well as unlabeled data to perform certain learning tasks.

Question 18

Learned Decision Tree

Answer 18

A tree-like model which shows how models come to conclusions.

Question 19

Batch Learning

Answer 19

When a earner is given training data as a batch.

Question 20

Online Learning

Answer 20

Instances are sent sequentially, updating the model after every instance.

Question 21

Active Learning

Answer 21

Learner can select instances for training.

Question 22

Concept Drift

Answer 22

The target function (f : X -> Y) changes over time.