WSEAS Transactions on Circuits and Systems
Print ISSN: 1109-2734, E-ISSN: 2224-266X
Volume 13, 2014
Fractal Neural Vector - Machines, Tomography and Inheritance of Behaviour
Author:
Abstract: Results of intensive studies of neural systems could not fully explain the astonishing performance of biological nervous systems in complex situations. Therefore alternative models of neural nets and their ways of pattern-processing might be of interest. Fractal neural nets offer interesting rich and flexible connectivity and biomorph aspects as hemispheres, lobes, gyri, sulci, decussatio of fibres, ventricular systems, thalamic structures and a high dynamism of processed patterns. Combining these fractal features with intracellular memory-strings to encode sequences of activities as engrams or vectors, to store, compare and reconstruct patterns of activities, a new tomographic form of information-processing seems to be achievable for such fractal neural vector- machines. Those memory-strings could, though completely hypothetically concerning their biological relevance, at least in principle allow the inheritance of behaviour. Though very preliminary, the results of a first small simulation may shed a light on the interplay of innate talents and learning experiences as well as on hypothetical mechanisms of genetic adjustments of organisms to the environment during evolution.
Search Articles
Keywords: Three-dimensional fractal neural nets, vector-machines, memory-strings, engrams, tomography, inheritance of behaviour
Pages: 464-475
WSEAS Transactions on Circuits and Systems, ISSN / E-ISSN: 1109-2734 / 2224-266X, Volume 13, 2014, Art. #50