en ru

Docs

NAME

computeAbstractNeurons - perform a discrete simulation of the abstract neurons to compute their output signals

SYNOPSIS

function computeAbstractNeurons(neurons, times);

DESCRIPTION

The function is used to perform a discrete simulation of the abstract neurons with ids neurons. Simulation consists of times steps. At each step, the consecutive computation of all output signals will be performed, starting from the first neuron. More than one step is required only if the neural network has a feedback.

RETURN VALUES

None.

EXAMPLES

A simple case:

connectors = createAbstractConnectors(6);
weights = createAbstractWeights(5);
procUnit = createProcUnit(PROC_UNIT.WEIGHTED_SUM);
actFunc = createActFunc(ACT_FUNC.LINEAR, 1.0, 0.0);
setSignals(connectors, 0, {1.0}); -- Connector 0 represents a bias input;

neuron1 = createAbstractNeuron(3, {0, 2, 1}, connectors, 3, weights, 0, procUnit, actFunc);
setAbstractWeights(neuron1, {7.0, 10.0, 5.0}); -- Set neuron1 weights;
setSignals(connectors, 1, {0.2, 0.1}); -- Set neuron1 input signals;

neuron2 = createAbstractNeuron(2, {0, 4}, connectors, 5, weights, 3, procUnit, actFunc);
setAbstractWeights(neuron2, {4.0, 2.0}); -- Set neuron2 weights;
setSignals(connectors, 4, {0.5}); -- Set neuron2 input signals;

computeAbstractNeurons({neuron1, neuron2}, 1);
out1 = getSignals(connectors, 3, 1); -- {9.0}
out2 = getSignals(connectors, 5, 1); -- {5.0}

SEE ALSO

createAbstractNeuron

Verbatim copying and distribution of this entire article are permitted in any medium, provided this notice is preserved.
Send suggestions, questions and bug reports to neurowombatmail@gmail.com
Copyright © 2009, 2010, 2011, 2012 Andrew Timashov