Politehnika i dizajn, Vol. 6 No. 3, 2018.
Izvorni znanstveni članak
https://doi.org/10.19279/TVZ.PD.2018-6-3-07
INTELIGENCE AND CREATIVITY OF A SINGLE NEURON
Predrag Valožić
; Tehničko veleučilište u Zagrebu, Zagreb, Hrvatska
Sažetak
The design results of the adaptive recursive
system - the determinate signal generator with
embedded linear neurons - were analyzed.
The reference design is classic, derived from
trigonometric identities and z-transformations.
Quite different process is adaptive design of a
generator with linear neuron. The linear neuron
based on the given template of the sinusoidal,
amplitude modulated and triangular signal
independently determines the weight coefficients
of linear filter and generator.
For different initial data after learning, neuron
provides different solutions: weight coefficients
and corresponding difference equation. Some
solutions are stable, some are not, so check it out
before the final selection.
In conclusion, the complementarity between
classical (engineering) and "neural" (artistic)
approaches to designing the system has been
demonstrated. Solutions that give a linear neuron
can be evaluated as creative and intelligent. The
linear neuron generator is a simple example of an
elementary creative neural network.
Ključne riječi
Adaptive recursive system; linear neuron; sine-wave generator, computational creativity
Hrčak ID:
209300
URI
Datum izdavanja:
18.9.2018.
Posjeta: 1.404 *