Production editor’s foreword

Iconics is a highly important division of optics, closely associated with process technology and image recognition. Neuroiconics is the science of image processing in the human visual system and of the use of this knowledge to create artificial-intelligence systems. These are two contrary directions in the evolution of the same region of information technologies—namely, neurotechnologies. The knowledge of neurophysiological mechanisms for constructing a visual picture of the world determines ways to develop artificial-intelligence systems and opens up new possibilities of controlling groups and technological processes. It is these neurotechnologies that determine the future prospect of the evolution of modern society.
Several special issues of Optichesk˘ıi Zhurnal in the past 20 years—two in 1999 and one each in 2011, 2015, and 2018— have presented articles on neuroiconics, neurotechnologies, and the performance and evolution of natural and artificial neural networks. Now in 2019, this issue of the journal is again devoted to neural networks and reflects the general trend of the evolution of worldwide research.
The dissertation of Boris Petrovich Babkin had immense significance for understanding the main operating principles of the brain’s neural networks and originated in studies of higher neural activity by I. P. Pavlov himself. This dissertation was defended in 1904 at the Military Medical Academy in St. Petersburg. Babkin introduced the concept of the temporal connections between neurons and their reconstruction. After being expatriated in the 1920s, Babkin accepted a professorial chair in Canada and there served as advisor to Donald Hebb, who founded the North American school of neurophysiology. Hebb’s Rules of the 1940s paraphrase the main result of Boris Petrovich Babkin’s 1904 dissertation. Donald Hebb’s 1949 book (D. O. Hebb, The Organization of Behavior) was one of the most influential books in the area of psychology and neuroscience. However, it is hard to resist the temptation to point out that Babkin’s concepts concerning the temporal connections between neurons were the basis of the rules proposed in Hebb’s book for the weight variations of the interneural connections that result from learning. They subsequently served as the basis for constructing artificial neural networks. A very important step was later made by Teuvo Kohonen in his 1977 book Associative Memory—A System-Theoretical Approach, which introduced new concepts concerning “content-addressable memories.” The book’s very name reflects a Pavlovian approach to the training of neural networks. Despite its simplicity, Hebb’s principle that “cells that are activated together connect to each other” (Babkin’s temporal connection) still inspires the developers of machine-learning systems when they create new algorithms and architectures for neural networks. This principle, for example, is used in the architectural solutions of GoogLeNet—the well-known network winner of competitions involved in image recognition in 2014.
There is no doubt that other principles can be very useful for developing neurotechnologies. This issue of Optichesk˘ıi Zhurnal opens with an article that points out the very important direction in which the technology for training artificial neural networks is evolving in accordance with the principle of least action. This fundamental principle of physics makes it possible both to understand the evolutionary development of natural neural networks and to predict how artificial neural networks will evolve.

Yuri˘ı Evgen’evich Shelepin is a doctor of medical science and a professor. He was born in Rostov-on-Don on January 30, 1945. He graduated from the Lvov State Medical Institute in 1969 and completed graduate studies at the I. P. Pavlov Institute of Physiology in 1972, and he worked starting in 1972 as Junior, Senior, and then Chief Scientific Fellow, and since 1988 has headed the Laboratory of the Physiology of Vision and the Sensory Systems Division of the I. P. Pavlov Institute of Physiology, Russian Academy of Sciences. He has worked in Helsinki, at Cambridge University, and at ITMO University. From 2010 until the present, Yu. E. Shelepin has been a professor at St. Petersburg State University. Yuri˘ı Evgen’evich Shelepin is the author of more than 500 scientific works, including 10 monographs and handbooks on the physiology of recognition of visual scenes and physiological optics and ophthalmology, and has about 50 author’s certificates and patents. He is a member of the editorial staff of Optichesk˘ıi Zhurnal and of the editorial staffs of the journals Sensornye Sistemy, Éksperimental’naya Psikhologiya, and Zhurnal Évolyutsionno˘ı Biokhimii i Fiziologii.