You are on page 1of 2

Simulation of Human Thinking - Artificial Neural Networks

7456
3538
9:16
5.18
Figure 7.6: Biological neuron structure

2.18
The synapse also has the property of modulating the electrical pulse from the axon. The
electrical potential generated by a neuron is, in fact, of the kind on/off; if the electrical
3.18
activity of the neuron exceeds a certain threshold, the impulse is generated; otherwise, not.
2:17

The generated pulse does not differ by intensity from one neuron to another. The potential
propagates along the axon and reaches the synapse with the dendrite of another neuron.
5510

The post-synaptic potential of dendrites depends on the biochemical characteristics of the


synapses. In the presence of the same pre-synaptic potential, two different synapses
:888

generate different post-synaptic potentials. In other words, the synapse weighs the input
potential by modulating it. Post-synaptic potentials propagate through neuron dendrites. At
0990

the soma level, they add up, and only if the result of that sum is greater than a certain
threshold will the neuron trigger the potential that propagates through its axon.
9545
9
one:
N
com:
rvox.
chola
nal.s

[ 260 ]
natio
inter
Simulation of Human Thinking - Artificial Neural Networks

A biological neuron, as we have said, receives various inputs through dendrites. An


artificial neuron also receives different input values. All inputs are added and the result is
the value counted by the neuron. If this value exceeds a certain threshold, the neuron
produces an output signal, or potential; otherwise, it remains silent.

7456
3538
9:16
Figure 7.7: Scheme of a single neuron with the threshold of activation

5.18
The algebraic sum of its inputs is the first function implemented by the artificial neuron.
This sum allows you to build the system response. In the case of an error in the simulation

2.18
of the phenomenon, an appropriate correction is required. To do this, each input must be
assigned a weight, that is, a numeric value that modulates the impact that this input has on
3.18
the total sum to determine the potential of the neuron. In other words, each input
2:17

contributes more or less to the determination of the threshold value being exceeded and the
potential triggering. This property is also inspired by a characteristic of biological neurons.
5510

The feature in question concerns the synapses that the axon of a neuron does with the
dendrite of another neuron.
:888

From the point of view of the post-synaptic neuron, inputs, constituted by the potential of
0990

other neurons whose axons synapse on its dendrites, are modulated precisely by such
synapses. Some of them will have a greater effect on the total sum. Others will even have a
9545

character inhibitory; that is, they will have the effect of lowering the value of the total sum
of the inputs, consequently decreasing the probability that it exceeds the threshold value
9

and triggers a potential.


one:

This property of biological systems is modeled in connection systems with the mathematical
N

weight concept. The weight of a connection is a numeric value for which to multiply the
com:

input value. In this way, the input will have a greater or lesser effect on the total amount
depending on the weight entity.
rvox.
chola
nal.s

[ 261 ]
natio
inter

You might also like