© 2008 vWISP PTY LTD
Each connection point between neurons is a synapse. The synapse is much more than
just a connection point though. At each synapse a receptor pad turns a specific neuro-
The neural array is a hierarchy of vWISP neurons, interconnected through dynamic synapses. Each neuron determines the probability that a spatial and temporal nput pulse event sequence is similar to a sequence it has learned previously. Neurons interact through simple rules. Each type of neuron (e.g. Pyramidal or Stellar) has its own set of rules concerning which other neurons they connect to.
The Soma of a neuron contains the cell's nuceus and all the mechanisms to allow the
cell to survive. These are the same as found in other cells. The nucleus contains
genetic material. Action potentials are generated in the axon hillock. Dendrites
integrate post-
The Transfer function for each device in the hierarchy is the sum of (synapses (|n| * delta (t)n) * efficacy(|n|)), whereby delta (t)n is measured from the leading edge of input pulse (n). The output of each neuron is a pulse train expressing the probability that the input matches a previously learned temporal/ spatial event pattern,
The Axon is the output of the neuron. The action potental that originates in the axon hillock is carried by the axon to synapses on other neurons. The axon may repeatedly fire following a trigger event The resulting pulse train is related to the input and efficacy of contributing synapses. Synapses are contantly updated, depending on the timing between action potential and input events.
A hierarchical array of vWISP neurons forms a temporal associative memory that is written to by repeating the same event pattern several times, and read by applying an approximation of a previously stored event pattern.
The vWISP temporal associative memory array works like the brain
