OVERVIEW OF SOFTWARE FOR WORKING WITH NEURAL NETWORKS

Authors

DOI:

https://doi.org/10.35546/kntu2078-4481.2024.2.22

Keywords:

machine learning, neural networks, libraries.

Abstract

One of the most important areas of research and development in modern cybernetics are the fields of machine learning, pattern recognition, and computer vision. The acceleration of the pace of development of information society technologies, the development of robotics, the development of the "smart home" and "smart city" concepts, the development of the Internet of Things and artificial intelligence systems determine this field’s special place in modern scientific knowledge. Many applied tasks in the practice of modern programming use methods of data collection, clustering and classification, methods of statistical inference. In everyday life, as well as in the corporate and industrial environment, technologies are beginning to be introduced that gradually erase the border between real and virtual space, which requires a new qualitative level of universally implemented recognition technologies, whose scope of application has grown enormously in recent years: recognition tasks that were considered the most difficult before. today are solved around the clock by the mobile devices of ordinary citizens. Computerized spaces with a pronounced topology, such as the "smart home" of an ordinary user, designed for many users augmented and designed for many users of virtual reality of various degrees of immersion, artificial intelligence in computer games for various purposes require new ideas and approaches, new level of accuracy and speed of recognition This article is devoted to a comparative analysis of some deep learning software tools, many of which have appeared recently [1]. Such tools include software libraries, extensions of programming languages, as well as independent languages that allow the use of ready-made algorithms for creating and training neural network models. Existing deep learning tools have different functionality and require different levels of knowledge and skills from the user. Choosing the right tool is an important task that allows you to achieve the desired result in the shortest time and with less effort.

References

Новотарський М.А., Нестеренко Б.Б. Штучні нейронні мережі: обчислення. Праці Інституту математики НАН України. Т. 50. Київ : Ін-т математики НАН України, 2004. 408 с.

Кононюк А.Ю. Нейроні мережі і генетичні алгоритми. Київ : «Корнійчук», 2008. 446 с.

M. Gupta Madan, Jin Liang & Homma Noriyasu. Gupta Statistic and Dynamic Neural Networks : From Fundamentals to Advanced Theory. USA : IEEE Press, 2003. 722 p.

Ciresan D.C., Meier U., Gambardella L.M. Schmidhuber J. Deep, Big, Simple Neural Nets for Handwritten Digit Recognition. Neural Computation. 2010. vol. 22, no. 12. pp. 3207–3220.

He K., Zhang X., Ren S., et al. Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (Las Vegas, NV, USA, 27–30 June 2016), 2016. pp. 770–778.

Hinton, G.E.: Learning Multiple Layers of Representation. In: Trends in Cognitive Sciences. Vol. 11. pp. 428–434. (2007)

Published

2024-07-01