Parallel Distributed Processing: also referred
to as Neural Networking or Connectionism
McClelland, J. L. and Rumelhart, D. E.
(1988). Explorations in Parallel Distributed Processing: MIT Press
Components of this “computational”
learning/memory theory: (a) processing units, (b) connections --
each unit can be connected to any other unit via a link which has a “weighting” or “strength” (c) the weighting can be either excitatory or inhibitory (d) activation rules, (e)
internal inputs (f) external inputs (g) unit processing output.
Information isn’t input into memory in
a step-by-step manner: consolidated first in sensory memory, then short-term, and then long-term memory. Rather, information is distributed to all parts of the networked memory
system at once.
Connections
Input Layer
Output Layer
Hidden Layers
Input Links to neuron
Output links from neuron
Neuron
Neural (Memory) Network
Input links to neural (memory) network
Output links from neural (memory) network
Arranged
by Gordon Vessels 2005