Learning takes place in a connectionist network through a process of <u>back propagation</u> in which an error signal is transmitted starting from the property units.
<u>Explanation</u>:
Back propagation is a set of rules used for learning about the artificial neural networks using gradient descent. The gradient of the error function is calculated with respect to the neural network's weights.
This algorithm can be efficiently used to calculate the derivatives. Connectionist networks helps in arranging the neurons into a network. This network defines the arrangement of neurons, transmission function of neuron and a learning rule.
In back propagation, the error signal is transmitted from the property units.
