CGSC 2001 Study Guide - Quiz Guide: Connectionism, Artificial Neural Network, Simple Algorithm

• School: Carleton University
• Department: Cognitive Science
• Course: CGSC 2001
• Professor: Sterling Somers

Document Summary

Parallel distributed processing: rumelhart and mccelland other time you might have to ignore a rule and favour a different one instead would be complicated. It"s job is to influence the network to relay the information and come out with the right output. The input is a simple algorithm that allows the program to send out information to the rest of the network. The input neuron influences the network in a certain way; if the output is a 9 but the input was a 2, we can go back and see how we came up with that mistake. When we look backwards to see where the influence is coming from is called backpropogation. Then we can change the weights in the network to fix the error. After the war we want to get rid of these mines so we don"t die. We go down in a submarine and listen to echoes.