Authors V.N. Gridin, V.I. Solodovnikov
Month, Year 07, 2016 @en
Index UDC
DOI DOI 10.18522/2311-3103-2016-7-114122
Abstract A prerequisite for the neural networks use as a mathematical basis for the cryptographic in-formation protection methods could be distinguished their ability to restore the distorted signals and recognize objects with differences from the reference characteristics. An additional advantage includes the hardware realizability of neural network algorithms, which increases speed of data encryption and decryption. One of the main problems that hinder the advancement of neural net-work encryption methods is the insufficiently studied reliability issues that make it urgent to study the characteristic features and vulnerabilities of the neural network cryptographic algorithms. The article investigates the questions of the neural networks usage for cryptographic information pro-tection. Algorithms for encryption, decoding and data preprocessing were proposed. The encryption algorithm is based on the generation of different variants of the distorted code that could be restored and identified by the used network. Moreover, the formation of the neural network uses the information on the frequency of occurrence of the source alphabet symbols, which makes it difficult to apply the methods of frequency cryptanalysis. At the stage of decryption the neural network classifies the input encrypted signals and converts them into source symbols. Thus, the proposed algorithm belongs to a symmetric cipher, since key encryption and decryption is a neural network itself, specifically the selected paradigm, its parameters and structural characteristics. The mathematical model of the neural network algorithm for symmetric encryption was carried out. Also it was noted the similarity of neural network algorithm with the proportional replace cipher, but with the feature characteristic of neural network data processing methods. The classical methods of cryptanalysis and their applicability in relation to the neural network algorithm were analyzed. Possible directions of cryptanalysis, as well as ways to improve the reliability have been offered. It should be noted that more prolonged and expert analysis of the algorithm and its implementation were conducted, then more accurate its cryptographic strength could be assumed.

Download PDF

Keywords Neural network; cryptographic protection; encryption; decryption; cryptographic strength; cryptanalysis.
References 1. Ido Kanter, Wolfgang Kinzel, Eran Kanter. Secure exchange of information by synchronization of neural networks, Europhys., Lett. 57, 141, 2002.
2. Kinzel W., Kanter I. Interacting neural networks and cryptography, Advances in solid state physics, 2002, Vol. 42, зз. 383-391.
3. Michal Rosen-Zvi, Ido Kanter. Wolfgang Kinzel Cryptography based on neural networks–analytical results, Journal of Physics A: Mathematical and General, 2002, Vol. 35, No. 47.
4. Klimov Alexander, Mityaguine Anton, and Shamir Adi. “Analysis of Neural Cryptography”, Computer Science department, The Weizmann Institute, Rehovot 76100 Israel.
5. Chervyakov N.I., Galushkin A.I., Evdokimov A.A., Lavrinenko A.V., Lavrinenko I.N. Primenenie iskusstvennykh neyronnykh setey i sistemy ostatochnykh klassov v kriptografii [Application of artificial neural networks and the system of residual classes in cryptography]. Moscow: Fizmatlit, 2012, 280 p.
6. Kallan R. Osnovnye kontseptsii neyronnykh setey [Basic concepts of neural networks]. Moscow: Izd. dom «Vil'yams», 2001, 287 p.
7. Ezhov A., Shumskiy S. Neyrokomp'yuting i ego primenenie v ekonomike i biznese [Neurocomputing, and its application in Economics and business]. Moscow: MIFI, 1998, 224 p.
8. Rassel S., Norvig P. Iskusstvennyy intellekt: sovremennyy podkhod [Artificial intelligence: a modern approach]. Moscow: Izd. dom «Vil'yams», 2007, 1408 p.
9. Medvedev V.S., Potemkin V.G. Neyronnye seti. MATLAB 6 [Neural network. MATLAB 6]. Moscow: Dialog MIFI, 2002, 496 p.
10. Gridin V.N., Solodovnikov V.I., Evdokimov I.A. Neyrosetevoy algoritm simmetrichnogo shifrovaniya [Neural network symmetric encryption algorithm], Informatsionnye tekhnologii [Information Technology], 2015, Vol. 21, No. 4, pp. 306-311.
11. Gridin V.N., Solodovnikov V.I., Evdokimov I.A. Primenenie neyrosetevogo podkhoda na osnove LVQ-seti dlya shifrovaniya tekstovoy informatsii [Application of neural network approach based on the LVQ network to encrypt text information], Sistemy vysokoy dostupnosti [System High Availability], 2011, Vol. 7, No. 1, pp. 65-68.
12. Evdokimov I.A., Gridin V.N., Solodovnikov V.I., Solodovnikov I.V. Predobrabotka dannykh s uchetom zadannykh znacheniy otdel'nykh priznakov [Preprocessing given the values of individual traits], Informatsionnye tekhnologii i vychislitel'nye sistemy [Information Technologies and Computing Systems], 2009, No. 1, pp. 14-17.
13. Alferov A.P., Zubov A.Yu., Kuz'min A.S., Cheremushkin A.V. Osnovy kriptografii [The basics of cryptography]. Moscow: Gelios ARV, 2005, 480 p.
14. Bauer F. Rasshifrovannye sekrety. Metody i printsipy kriptologii [Decrypted secrets. Methods and principles of cryptology]. Moscow: Mir, 2007, 550 p.
15. Shnayer B. Prikladnaya kriptografiya. Protokoly, algoritmy, iskhodnye teksty na yazyke Si [Applied cryptography. Protocols, algorithms, and source code in C language]. Moscow: Triumf, 2003, 610 p.
16. Shnayer B., Fergyuson N. Prakticheskaya kriptografiya [Practical cryptography]. Moscow: Vil'yams, 2005, 424 p.
17. Panasenko S.P. Sovremennye metody vskrytiya algoritmov shifrovaniya [Modern methods of opening the encryption algorithms]. Part 1. CIO-World. 23.10.2006.
18. Stinson D.R. Cryptography: Theory and Practice. CRC Press, 1995.
19. Agranovskiy A.V., Khadi R.A. Prakticheskaya kriptografiya: algoritmy i ikh programmirovanie [Practical cryptography: algorithms and their programming]. Moscow: Solon-Press, 2009, 258 p.
20. Smart N. Kriptografiya [Cryptography]. Moscow: Tekhnosfera, 2005, 528 p.
21. Ryabko B.Ya., Fionov A.N. Kriptograficheskie metody zashchity informatsii [Cryptographic methods of information protection]. Moscow: Goryachaya Liniya – Telekom, 2005, 229 p.

Comments are closed.