L2 - The Hull-Spence Model

Question 1

What is the benefit of using a mathematical model?

Answer 1

It can make accurate predictions.

Question 2

What is this formula?

ΔV_A(t)​ = k(λ - V_(t-1))

Answer 2

The Hull-Spence Model

Question 3

What does surprise have to do with learning?

Answer 3

If an animal is surprised by something, it generates large changes in associations and has increased learning.

Question 4

What does ‘V’ represent in the Hull-Spence Model

Answer 4

The change in the strength of the CS(A)-US association on trial (t).

Question 5

What does ‘A’ represent on the Hull-Spence Model?

Answer 5

The conditioned stimulus (e.g. bell)

Question 6

What does ‘Δ’ (delta) represent in the Hull-Spence Model?

Answer 6

In maths, delta means a change. ΔV will always mean the change in the strength of the association.

Question 7

What does ‘(t)’ stand for in the Hull-Spence Model?

Answer 7

It stands for the current trial we are having.

Question 8

What does λ (Lambda) represent in the Hull-Spence Model?

Answer 8

How strong the unconditioned stimulus was.

Actual Outcome (relative):

λ = 1 if US is present

λ = 0 if US is absent.

Question 9

What does V_A(t-1) in the Hull-Spence Model represent?

Answer 9

Expected Outcome = Associative strength that A has acquired up until trial t.

Question 10

In the Hull-Spence Model, what does λ - V_A(t-1) represent?

Answer 10

The extent to which the US wasn’t expected.

Surprise!

Question 11

What does the k( ) around the λ - V_A(t-1) represent?

k(λ - V_A(t-1))

Answer 11

It is the learning rate parameter

(Speed of Learning)

Question 12

In the first trial with an animal, what would you expect the V_A value be?

Why?

Answer 12

It should be at 0.

For the animal has never seen the US before and therefore there is maximum surprise.

Question 13

Using the Hull-Spence Model, if the amount of learning from an animal at the end of your first trial is 2 what would you change in your formula for the second trial?

Answer 13

You would change the Δ (delta) to 0.2 (V_A(2))

Question 14

What does asymptote mean?

Answer 14

A straight line that continually approaches a given curve but does not meet it at any finite distance.

A line that a curve approaches, as it heads towards infinity.

Question 15

What does negatively accelerated mean?

Answer 15

It means that the rate (of learning) is slowing down.

Question 16

In learning, what does extinction mean?

Answer 16

It means to unlearn a learned behaviour.

Question 17

If your Hull-Space Model equation gives you a negative result what does that mean?

Answer 17

It means that you are surprised by the omission of something.

The associative strength will go down, the beginning of extinction if paired with the US.

Question 18

What is the problem with the Hull-Spence Model? What assumption does it make that is wrong?

Answer 18

It assumes that learning about one CS is independent from learning about other co-occurring CSs when it is not.

i.e., it doesn’t matter if other CSs are also present.

Question 19

What is the blocking effect?

Answer 19

If you train an animal to pair 2 CS (A+B) together with a US the they will have a strong CR, but if you have a phase beforehand where you only pair one stimulus (A) and in the second phase pair A+B with US B will have a weak CR.

It’s as if A is blocking learning about B, this is called blocking.