Connectionist Models Flashcards

1
Q

What is computational cognitive models used for?

A

Can be used to gain insights, build hypotheses and make predictions

Models may be viewed as a theory of phenomenon it describes

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

Cognitive models can be built at different levels of abstraction. Name these levels (BFAFN)

A

Behaviour
Functional description of cognitive modules
Algorithmic account of representations, mechanisms and processes
Functional description of neural networks
Neurochemical description of neural networks

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

What is the rationale behind connectionism?

A

Since human brains are composed of neutron the functional properties of these systems should influence our theories of cognition

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

Which model is thought to replace the classical computational theory of mind?

A

Connectionism (ANNs)

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

What does the degree of activation represent

A

It describes how active a particular neuron is at each point in time (usually between 0-1)

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

Neutrons feed their activation into other neuron? True or false

A

True

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

How do you compute activation of neurons?

A

By looking at the amount of activation fed into it

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

What are the two types of connections between neurons?

A

Excitatory +

Inhibitory -

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

What is the activation rule based on?

A

Input activations

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

Name an activation function

A

Sigma

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

What are neurons typically arranged into (ANN)

A

Input, hidden and output layers

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

What is the output layer

A

Output layer has the vector corresponding to each possible probability of words

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

How can weights be learned? (ANN)

A

Specify training data
Choose Error criterion
Apply backpropagation algorithm

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

What must you apply backpropagation algorithm?

A

To compute weights that minimise error

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

What do you weights and patterns of connectivity represent ?

A

Long term Knowledge

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

What is a usefulness for the model (Ann)

A

It helps us understand learning in real creatures for example by simulating learning curves during child development and observing model behaviour over time