The Heuristic Work of the Brain and Artificial Neural Networks
| Autor: | V. V. Eskov, V. M. Eskov, V. Pyatin, L. K. Ilyashenko |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
0301 basic medicine
Structure (mathematical logic) Quantitative Biology::Neurons and Cognition 030102 biochemistry & molecular biology Artificial neural network Heuristic (computer science) Computer science business.industry Property (programming) Computer Science::Neural and Evolutionary Computation Work (physics) Rank (computer programming) Biophysics 03 medical and health sciences Task (computing) 030104 developmental biology Backward propagation Artificial intelligence business |
| Zdroj: | Biophysics. 64:293-299 |
| ISSN: | 1555-6654 0006-3509 |
| DOI: | 10.1134/s0006350919020064 |
| Popis: | This paper presents two new fundamental principles of the functioning of real neural networks of the brain. These principles have inspired the design of artificial neural networks (a neuro-computer). In addition to well-known properties of artificial neurons (threshold properties, neural network formation, and backward propagation of errors), we describe two new major properties of real neural networks of the brain by which a neuroemulator may work. We discuss the practical usefulness of these properties for the neuro-computer. The first property is permanent statistical chaos in the structure and functions of the brain neural networks. The second property is reverberations, or multiple iterations, in the work of neural networks. The introduction of these two properties to the work of the common artificial neural network provides a new quality, that is, solving the problem of systemic synthesis and finding rank parameters. Modern science has not investigated and solved this problem, which is a general fundamental problem in medicine. It has been shown that this task is identical to the heuristic activity of the brain. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink