Concept Learning

Question 1

What is a concept? English Definition.

Answer 1

A description of attributes/characteristics that can identify a particular set of object/events over a larger set of varying objects/events

Question 2

What is a concept? Math Definition.

Answer 2

Boolean values function that returns true if an instance matches a description of attributes/characteristics…

Question 3

What are instances?

Answer 3

Set of items over-which a concept is defined

Question 4

What is a hypothesis?

Answer 4

Set of constraints on instance attributes.

Question 5

What is a conjunctive concept?

Answer 5

Conjunctive concepts mandate such properties like mother mandate that both features (i) female, and (ii) parent, be present with omission of either being impossible for mother to remain true

Question 6

What is a disjunctive concept?

Answer 6

A disjunctive concept such as a car that has many attributes but can still be a car without one of the attributes

Question 7

What is the maximally general hypothesis?

Answer 7

Accepts all instances of X.

Question 8

What is the maximally specific hypothesis?

Answer 8

Accepts no instances of X.

Question 9

What is the target concept?

Answer 9

The concept we are trying to learn.

Question 10

What is the goal of concept learning?

Answer 10

To find all hypothesis s.t h(x) = c(x) for all x

Question 11

On what proof is concept learning based on?

Answer 11

Induction

Question 12

What is the inductive learning hypothesis

Answer 12

Any hypothesis that can approximate the target function over sufficient training examples will do the same for test examples

Question 13

What is the Find S algorithm

Answer 13

Finds the maximally specific hypothesis of a training set.

Basically one half of candidate elimination: only generalise the maximally specific hypothesis

Question 14

What are some of the advantages of find S?

Answer 14

Speed
Simple
Scalable
Always guarantees one viable hypothesis

Question 15

What are some of the disadvantages of find S?

Answer 15

Won’t always converge
Guarantees only one viable hypothesis
Sensitive to noise

Question 16

What is a version space?

Answer 16

Set of all hypotheses consistent with training data

Question 17

What is the List then Eliminate Algorithm?

Answer 17

Brute force

Generate every feasible hypothesis then eliminate ones where h(x) /= c(x)

Question 18

Advantages of LTE

Answer 18

Exhaustive

Returns list of consistent hypotheses

Question 19

Disadvantages of LTE

Answer 19

Exhaustive
Scales bad
Pretty much doo doo unusable dog water algorithm

Question 20

What is the candidate elimination algorithm?

Answer 20

On positive examples, we generalise the maximally specific hypothesis and on negative examples, we constrain the maximally general hypothesis.

(You have to watch a youtube video)

Question 21

Does C.E. converge

Answer 21

Yes if no noise and a valid hypothesis exists

Question 22

What type of examples should learners request? (yes they can request examples)

Answer 22

Aim to remove as many competing hypotheses from the version space as possible.

Question 23

How can we use partially learned concepts?

Answer 23

Have each hypothesis in the version space vote on whether an instance should be accepted or not.

An instance will be positive if it is accepted by every member of S.

An instance will be negative if it is rejected by every member of G

Question 24

What is an unbiased learner?

Answer 24

define an unbiased hypothesis space by allowing arbitrary conjunctions, disjunctions, and negations.

Basically overfitting.