Computational Modelling

Question 1

Modelling provides a framework for interpreting data

Answer 1

Select a model based on quantitative and intellectual judgement.

Question 2

Model= abstract framework that captures the structure in the data

Answer 2

Simpler version of what they’re trying to explain.

Question 3

Levels in analysis (Marr, 1982)

Answer 3

• computational level: what the system does/why.
• representational level: how does it do that, what processes build the representations.
• physical system level: how the system is physically realised (eg. biological vision which neural structures build the visual system.

Question 4

Model classification (Lewandowsky + Farrell, 2011)

Answer 4

• data description: describes relationship between variables.
• process characterisation models: peek inside “black box”. Neutral to implementations of the processes they characterise.
• process explanation models: more up-close view of ‘black box’ - try to implement how the processes occur.

Question 5

Why model?

Answer 5

The instantiation of a quantitative model ensures all assumptions of a theory have been identified and tested- take the place of theory: they implement the mechanisms and generate predictions.

Question 6

Process characterisation models:

Answer 6

Explanatory power lies in hypothetic constructs of the mind.
Process explanation models: close up of what’s in the ‘box’.
- implement how processes occur at the models level of specification.

Question 7

Expectations

Answer 7

Bring out relationships between sets of data that we wouldn’t have otherwise notices: emergence of understanding- explore implications of human behaviour.

Question 8

Potential problems?

Answer 8

Need to be falsifiable but not false.

Need to have verisimilitude - ‘partial truth value’.