Please enable JavaScript.
Coggle requires JavaScript to display documents.
Cognitive Psychology - Coggle Diagram
Cognitive Psychology
Similarity
Perceptual and conceptual foundations for similarity:
- We learn about similarity from early age
- Seeing relatedness is the foundation of cognition
How do we perceive similarity?
- Shared perceptual features produce a sense of “likeness”
- We can set aside superficial differences to see a “structural” similarity
- Similarity is not purely a perceptual phenomenon > similarity of conceptual connections
- Similarity can do perceptual and conceptual work at the same time
- DNA has a physical similarity to a zipper, which the visual system detects immediately
- The genetic information in DNA is like a blueprint, which is a conceptual relationship.
Similarity helps us form categories and make generalisation
- Why have a notion of similarity at all?
- The snowflake problem: no two people, events, objects, etc are exactly the same as anything else.
- Unless you have a basis on which to make productions > similarity > it’s hard to get by in the world
- Drawing assumptions, making predictions
- Things are similar enough to make inferential leaps or generalisations or predictions to our advantage.
- Our sense of similarity allows us to order things into kinds so that these can function as stimulus meanings.
- Reasonable expectation depends on the similarity of circumstances and on our tendency to expect that similar causes will have similar effects.
How do we measure similarity?
- Confusability: the probability of mistaking A for B.
- A mistaken identity is a confusion and occurs for more similar items
- The more similar, the more confusion
- Reaction time: time taken to distinguish A from B
- Dissimilar > easy > fast
- Similar > hard > slow
- Forced choice: is X more like A or more like B
- Used with younger participants
- Likert scales: how similar is A to B
Simple theories of similarity: geometric models
- Subjective similarity space: Distant things are dissimilar, nearby things are similar
- Empirical evidence: similarity helps us generalise from one stimulus to another
- The universal law of generalisation
- The probability of generalising from one stimulus to another decreases exponentially as a function of dissimilarity (distance)
- Problems: the symmetry constraint
- The distance from A to B is the same as the distance from B to A, so similarity must be symmetric
- The are statements that are not equivalent: “my butcher is like a surgeon” vs “ my surgeon is like a butcher.”
Simple theories of similarity: featural models
- Breaking things down in terms of horses.
- Asymmetrical knowledge about two terms or concepts
- Makes less sense to liken the less-knowledgable concept to the more-known one. Likening the more known thing to the less known one teaches you about the less known one
- Common and distinctive features
Richer theories of similarity: structural alignment
- The structure does the work
- Removing colour features to see if similarity is still intact
- Blurring out high-frequency spatial information to see if similarity is still intact
- Filtering out everything except high-spatial frequency to see if similarity is still intact
- Object descriptions need to say something about structure > where certain items or colours are
- When two objects share a feature, and that feature appears in the same slot > match in place.
- When the shared feature appears in a different location it is a match out of place
- MIPS have a bigger influence than MOPS
- Changing MIPS increases similarity more than changing MOPS
Richer theories of similarity: stimulus transformation
- We live in a dynamic environment, and everything can be manipulated via processes > rotation, etc.
- Ease or feasibility of transformation processes.
- Short transformations (1) are more similar, more confusable, slower reaction time
- Long transformations are less similar, less confusable and have a faster reaction time.
-
The Cognitive
Revolution
- the mind is a machine > computational metaphor
- Information processing machine that produces empirically testable predictions.
- Cognitive psychology as a computationalist views it:
- Information processing > the computational metaphor suggests that we can use the language of computing to build models
- Why?
- Computational language is precise; we can generate empirically testable predictions (methodological behaviourism)
- Computational language is flexible; we can postulate hidden mechanisms and structure to cognitive processes (goes beyond radical behaviourism).
measuring cognition
- common methods
- Accuracy of response
- Type of response
- Response time
- Neurological deficit
- Brain imaging
- Self-report
What is Cognition
- Words people associate with cognition: knowledge, mind, science, comprehension, understanding, psychology, etc.
- Cognition is concerned with knowledge, reasoning, memory, language, decision-making, etc.
is computation special?
- There's criticism on the idea that the brain works like a computer
- A superficial interpretation of the metaphor is that the brain and a laptop are similar to each other > false.
Alan Turing
- Turing machine is a very simple (hypothetical) device that reads and writes symbols off a piece of tape > universal machine in the sense that anything that can be computed, can be done with the machine.
- Digital computers are much fancier versions of Turing Machines
- Neural networks describe an information processing system that is inspired by the structure of the brain.
- While none of these are similar, a program given to one universal Turing machine can be rewritten using the language of another universal Turing machine.
- At a fundamental level, what underpins the computational metaphor is the fact that these are all information-processing machines > they are computers.
Levels of Explanation
- Bottom-up: cognition is focused on the brain, and our theories of cognition should be informed by the biology of the brain.
- Top-down: cognition is a feature of intelligent agents, and our theories of cognition should be informed by understanding what intelligent agents do (what do they want to achieve, how do they make it happen).
- Both terms agree in the sense that at an abstract level, cognition is a form of computation and the brain does information processing. They disagree in asking “what mechanisms does the brain use to perform computation?” (bottom-up) and “what computational problems does the mind solve?” (top-down).
Marr's levels of analysis
- Abstract computation: what problem does cognition solve
- Algorithm: what processing steps does it follow to do so
- Implementation: how is this instantiated as a physical entity (in the physical sense).