04. Volchikhin V. I., Ivanov A. I., Neural Network Molecule: a Solution of the Inverse Biometry Problem through Software Support of Quantum Superposition on Outputs of the Network of Artificial Neurons

UDC 577.33:004.4

DOI: 10.15507/0236-2910.027.201704.518-529

Neural Network Molecule: a Solution of the Inverse Biometry Problem through Software Support of Quantum Superposition on Outputs of the Network of Artificial Neurons

Vladimir I. Volchikhin
President of Penza State University (40 Krasnaya St., Penza 444000, Russia), Dr.Sci. (Engineering), Professor, ORCID: ORCID: http://orcid.org/0000-0002-9986-531X, This email address is being protected from spambots. You need JavaScript enabled to view it.

Alexander I. Ivanov
Head of Biometric and Neuronal Nets Technology Laboratory, Penza Scientific Research Electrotechnical Institute (9 Sovetskaya St., Penza 440026, Russia) Dr.Sci. (Engineering), Associate Professor, ORCID: http://orcid.org/0000-0002-3854-2660, This email address is being protected from spambots. You need JavaScript enabled to view it.

Introduction: The aim of the study is to accelerate the solution of neural network biometrics inverse problem on an ordinary desktop computer.
Materials and Methods: To speed up the calculations, the artificial neural network is introduced into the dynamic mode of “jittering” of the states of all 256 output bits. At the same time, too many output states of the neural network are logarithmically folded by transitioning to the Hamming distance space between the code of the image “Own” and the codes of the images “Alien”. From the database of images of “Alien” 2.5 % of the most similar images are selected. In the next generation, 97.5 % of the discarded images are restored with GOST R 52633.2-2010 procedures by crossing parent images and obtaining descendant images from them.
Results: Over a period of about 10 minutes, 60 generations of directed search for the solution of the inverse problem can be realized that allows inversing matrices of neural network functionals of dimension 416 inputs to 256 outputs with restoration of up to 97 % information on unknown biometric parameters of the image “Own”.
Discussion and Conclusions: Supporting for 10 minutes of computer time the 256 qubit quantum superposition allows on a conventional computer to bypass the actual infinity of analyzed states in 5050 (50 to 50) times more than the same computer could process realizing the usual calculations. The increase in the length of the supported quantum superposition by 40 qubits is equivalent to increasing the processor clock speed by about a billion times. It is for this reason that it is more profitable to increase the number of quantum superpositions supported by the software emulator in comparison with the creation of a more powerful processor.

Keywords: neural network converter biometry-code, biometric data, large dimensions, software support of quantum superposition, artificial neurons

For citation: Volchikhin V. I., Ivanov A. I. Neural Network Molecule: a Solution of the Inverse Biometry Problem through Software Support of Quantum Superposition on Outputs of the Network of Artificial Neurons. Vestnik Mordovskogo universiteta = Mordovia University Bulletin. 2017: 27(4):518–529. DOI: 10.15507/0236-2910.027.201704.518-529

This work is licensed under a Creative Commons Attribution 4.0 License.