Lecture 3 - Marr's Levels of Analysis Flashcards

1
Q

theoretical constructs

A

perception, recognition

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

operational definition

A

to help measure when theoretical constructs are present: “indicates stuff going on in there”

something the scientist comes up with: “when you witness this thing, you can feel confident that this theoretical construct was present”

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Marr’s Levels of Analysis

A

three different perspectives of the same issue

these levels interact

study of one can inform our approach to another

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Computational Level

A

most abstract

what is it doing when faced with an input, how does it behave?

inputs coming in and outputs going out of the “black box”

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Algorithmic level

pt. 1

A

what steps does the system take to try and accomplish that goal or solve that problem

parse it up: “what’s it trying to do,, it does A, B, C, D… to try to solve this problem characterized at the computational level

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Implementation Level

A

the most concrete: nuts and bolts

“ok, how is the system actually accomplishing it? what are the neurons doing? firing pattern? computer wired together?”

have to know what system you’re dealing with

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

How to decide which level it is

A

Focus on: What was the explanation? What was the answer given by the experimenter? What were they explaining or describing to you?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Algorithm Example

A

make some steps or some model that an experiment is working on

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

How to determine the nature of inputs and outputs considered at computational level?

A

the type of problem being solved by the system (“black box”)

what was the system doing? how was it responding when the input came in? what was the behavior?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

which element in model is best matched to the ‘input’ stage of Marr’s computational level?

A

distal stimulus: the input in isolation of anything else, environmental stimulus that is giving you an input into the system and there’s an output coming out

input common to everyone: you would want to present the same distal stimulus to everyone in an experiment

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

proximal stimulus

A

the input as its hitting the sensory organ

has some property of the receiver, so the input coming in is changing at the level of the proximal stimulus (ppl have different vision capabilities)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Algorithmic Level

pt. 2

A

what steps or procedure going on in the “black box”

“to get from that input to that output the system went through the following steps…”

we still don’t know what kind of system (human brain, computer, alien, etc) is working on the problem

We only specify the steps the system takes and what kinds of info (representations) it manipulates

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Real-World Algorithmic example would look like…

A

some kind of model or equation, when it specifies a numerical, regular relationship

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Implementation Level

and the black box

A

get to look into the “black box” and see what kind of system you’re working with and how those components are interacting to embody that algorithm or produce that output

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Implementation Level

Input –> …….. –>

A

Neural Activity

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Example of Implementation Level processes

A

receptors fire, signals go to complex cells, etc.

we can specify how the elements of the system work together

17
Q

the level at which you begin your analysis can determine the types of….and… you will use

A

tools and methods

18
Q

Computational level introduces what kinds of explanations?

A

behavioral explanations: a behavioral output given some input

“when the thing sees this the thing does that”

19
Q

Algorithmic level introduces what kinds of explanations?

A

characterize a model, a psychological theory, a psychophysical law ——-

something that tells you regularities such that if you plug inputs into those regularities you will know what the output will be - tells you the steps being taken

20
Q

Implementation level explanations

A

some explanation about the physiology: “here’s what the cells in this area are doing”