Neural net lexicon

From Han Wiki
Jump to navigation Jump to search
layer
a subgroup of processing elements
input layer
usually the first layer
output layer
usually the last layer
hidden layers
layers in between the input and the output layer
cells, neuromimes, artificial neurons
processing elements
threshold function
used to determine the output of a neuron in the output layer
connections
synapses between cells
autoassociation
pattern matching (A => A)
heteroassociation
pattern matching (A => B)
learning
process of changing the weights
training
the act in which a network participates when learning is employed
self-organizing
unsupervised learning
supervised learning
external criteria are used to be matched by the network output
LSTM
long/short term memory -- preservation of intended training on the original set
structure
how many layers, what functions they have (input or output), interconnections & their functions
encoding
paradigm used for the determination of and changing of weights on the connections between neurons
recall
getting an expected output for a given input
correlation matrix
aka weight matrix
hidden layer
aka Kohonen or Grossberg layer
cooperation
the attempt between neurons in one neuron aiding the prospect of another neuron's firing -- vehicle for these phenomena is the connection weight
negative connection value could mean either inhibition or competition
competition
the attempt between neurons to individually excel with higher output
orthogonal
dot product results in zero
Hamming distance
easier method than Euclidean distance to measure proximity between patterns