Artificial neural network: Difference between revisions
imported>Subpagination Bot m (Add {{subpages}} and remove any categories (details)) |
imported>Felipe Ortega Gutiérrez No edit summary |
||
| Line 1: | Line 1: | ||
{{subpages}} | {{subpages}} | ||
'''Artificial Neural Networks''' (ANNs for short) are a connectionist processing model inspired on the biological neural networks. Artificial neural networks are composed by simple nodes called [[artificial neuron|artificial neurons]], | '''Artificial Neural Networks''' (ANNs for short) are a connectionist processing model inspired on the biological neural networks. Artificial neural networks are composed by simple nodes called [[artificial neuron|artificial neurons]]. They can be implemented via hardware (i.e: electronic devices) of software (i.e: computer simulations). | ||
In some models, the network behavior is stored in the connections between processing units in values called ''weights'', which represent the strength of each link, equivalent to many components of its biological counterpart. | |||
==Adaptation and Learning== | ==Adaptation and Learning== | ||
Revision as of 00:25, 9 April 2008
Artificial Neural Networks (ANNs for short) are a connectionist processing model inspired on the biological neural networks. Artificial neural networks are composed by simple nodes called artificial neurons. They can be implemented via hardware (i.e: electronic devices) of software (i.e: computer simulations).
In some models, the network behavior is stored in the connections between processing units in values called weights, which represent the strength of each link, equivalent to many components of its biological counterpart.
Adaptation and Learning
When a neuron receives and processes an input signal, it changes the weight value assigned to the input received. The weighted signals are summed to form the activation value, which is filtered by a function called transfer function. Changes in a specific neuron's behavior produces changes in the entire neural network, and this is basically how the neural networks learn.