Week 8

Question 1

What is a symbolic processing model?

Answer 1

Different amodal symbols or categories are measures (yes/no, 5 in, etc.) and symbols combine to create a whole

Question 2

What is a connectionist processing model?

Answer 2

A system (similar to action potential) where:
- an input is given and broken down into multiple representational parts, or nodes
- signals are sent through nodes, and strength of signal determines activation level and if node fires
- happens through multiple hidden layers
-Outcome occurs when highest activation level is selected, determining a node from multiple outcome options

Question 3

What’s an example of connectionist processing?

Answer 3

If given a number between 0-9, “nodes” represent each number between 0-9. Signals are sent between nodes, and strength of signal determines activation level

Question 4

What is an input layer?

Answer 4

Detects basic features (edges, colors, phonemes, etc)

Question 5

What is a hidden layer?

Answer 5

Layers in between the input and output layer that signals are sent through and activation levels are measured through

Question 6

What is an output layer?

Answer 6

The node with the highest activation level which determines outcome between possible options (output layer nodes)

Question 7

What is sub-symbolic processing?

Answer 7

No node represents one stand-alone concept (such as face), they all represent smaller pieces of a concept (outline, colors, textures, etc)

Question 8

What are distributed representations?

Answer 8

Many nodes, because they are sub-symbolic, work together to represent a whole concept

Question 9

What is parallel distributed processing?

Answer 9

Nodes are activated and processed at the same time

Question 10

What are connectionist models used for?

Answer 10

Object identification and machine learning (learns to identify things automatically)