Hebbian theory

The Hebbian learning rule is an established by psychologist Donald Olding Hebb rule to the emergence of learning in neural networks or in an association of neurons that have common synapses.

History

Hebb formulated in 1949 in his book The Organization of Behavior: " When an axon of cell A [ ... ] cell B excited and repeatedly and continuously contributes to the generation of action potentials in cell B, this results in growth processes or metabolic changes in one or two cells, which have the effect that the efficiency of the cell A is greater with respect to the generation of action potentials in B. "

This means: The more often a neuron A is active simultaneously with neuron B, the more preferable the two neurons are successively react ( "what fires together, wires together" ). This Hebb has demonstrated by changes in synaptic transmission between neurons.

Hebb is considered so as the discoverer of synaptic plasticity, which is the neurophysiological basis of learning and memory.

Formula

In artificial neural networks, this change in synaptic transmission is represented as a weight change of the neural graph. The Hebbian learning rule is the oldest and simplest neural learning rule.

: Change of the weight from neuron j to neuron i ( ie, the change in the bond strength of these two neurons)

: Learning rate (one suitable to be selected constant factor )

: Activation of neuron i

: Output of neuron j, which is connected to neuron i