Artificial vision. Artificial intelligence. Neural models of living sensory systems. Levashov O.V.

The universal interest in the topic of artificial intelligence is constantly growing and causes real excitement in the media and in the scientific literature. We owe this attention, in particular, the successes of mathematicians and programmers in teaching artificial neural networks (“deep training”). However, the sensory systems and the brain of simple living systems (for example, flying insects) are much superior to artificial neural networks in terms of effectiveness. To increase their effectiveness, there are many more tasks to find optimal neural architectures and algorithms for their work. The real help in such work could be provided by real mechanisms of recognition and management, borrowed from wildlife. It was this direction that successfully developed in the laboratory of the Institute of Management of Management, where the author of this book worked. The book compares various versions of computing neural networks, including “deep learning”, they consider their pros and cons. The book will be useful to students and graduate students of technical universities related to information technologies, students of biological faculties of universities involved in the problems of modeling living sensory systems and brain, as well as a wide range of readers interested in artificial intelligence studies.