Eğitim gerektirmeyen kodlayıcı-tabanlı Transformer ile DNA-alfabeli sözde rastgele sayı üreteci

Abstract

DNA temelli uygulamalarda çoğunlukla bit düzeyinde çalışan klasik sözde rastgele sayı üreteçleri (SRSÜ) kullanılmaktadır. Ancak bu üreteçler DNA uzayına doğrudan çıkış veremediğinden, ek kodlama/adaptasyon katmanları gerektirir. Bu durum hem hesaplama maliyetini hem de anahtar tasarımının karmaşıklığını artırmakta, DNA düzeyindeki rastgelelik özelliklerinin kontrolünü zorlaştırmaktadır. Rastgelelik özelliklerini doğrudan DNA alfabesi üzerinde üretebilen mekanizmalara ihtiyaç duyulmaktadır. Bu çalışmada, eğitim gerektirmeyen ve rastgele başlatılmış ağırlıklara sahip kodlayıcı-tabanlı bir Transformer mimarisi kullanılarak DNA-alfabeli bir SRSÜ önerilmektedir. Model pozisyon bilgisi içermemekte; çıkış katmanında kullanılan düşük genlikli Dirichlet gürültüsü ile çalışmaktadır. Blok içi dengeli örnekleme ve tek seferlik global karıştırma birlikte kullanılarak A/C/G/T sembollerine ait yerel frekans yanlılıkları ve kısa menzilli örüntüler zayıflatılır. Sekiz bijektif DNA→ikili eşlemeden rastgele seçilen biriyle dönüştürülen 500 000 bazlık dizilerden 1 Mbit ikili on bağımsız çıktı üretilmiştir. CPU üzerinde, blok boyutu 100 iken üretim süresi yaklaşık 49 sn. olarak ölçülmüştür. On bağımsız akışın tamamı NIST SP 800-22 p>0.01 eşiğini sağlamış, ENT analizinde ≈8 bit/byte entropi ve ihmal edilebilir seri korelasyonu gözlenmiştir. Bu bulgular, önerilen DNA-alfabeli SRSÜ’nün istatistiksel rastgelelik açısından tatmin edici bir davranış sergilediğini göstermektedir. Çalışma formel güvenlik ispatı sunmak yerine bu davranışın deneysel olarak değerlendirilmesine odaklanmaktadır.

Keywords

• Sözde rastgele sayı üretimi
• Yalnızca kodlayıcı Transformer
• Eğitim gerektirmeyen model
• İstatistiksel rastgelelik testleri

Training-free encoder-based Transformer for a DNA-alphabet pseudo-random number generator

Abstract

In DNA based applications, classical bit-level pseudo-random number generators (PRNGs) are mostly used. However, since these generators cannot produce outputs directly in the DNA space, additional encoding/adaptation layers are required. This increases both the computational cost and the complexity of key design, and makes it more difficult to control randomness properties at the DNA level. Therefore, mechanisms that can generate randomness properties directly on the DNA alphabet are needed. In this study, a DNA-alphabet pseudo-random number generator is proposed using a training-free, encoder-based Transformer architecture with randomly initialized weights. The model does not incorporate positional information and operates with a low-amplitude Dirichlet noise applied at the output layer. Within-block balanced sampling and a one-time global shuffling are used together to attenuate local frequency biases and short-range patterns in the A/C/G/T symbols. From 500,000-base sequences transformed using one of eight bijective DNA→binary mappings chosen at random, ten independent 1 Mbit binary outputs were generated. On CPU, with a block size of 100, the generation time was measured to be approximately 49 s. All ten independent streams satisfied the p>0.01 threshold in the NIST SP 800-22, and in the ENT analysis an entropy of ≈8 bits/byte and negligible serial correlation were observed. These findings show that the proposed DNA-alphabet PRNG exhibits satisfactory behavior in terms of statistical randomness. The study focuses on the experimental evaluation of this behavior rather than providing a formal security proof.

Keywords

• Pseudo-Random number generation
• Encoder-Only transformer
• Training-Free generation
• Statistical randomness tests

References

1. [1] Jeong YS, Oh KJ, Cho CK, Choi HJ. “Pseudo-random number generation using LSTMs”. The Journal of Supercomputing, 76(10), 8324-8342, 2020.
2. [2] Wu X, Han Y, Zhang M, Zhu S, Cui S, Wang Y, Peng Y. “Pseudorandom number generators based on neural networks: a review”. Journal of King Saud University - Computer and Information Sciences, 37(3), 18, 2025.
3. [3] Wu X, Han Y, Zhu S, Li Y, Cui S, Wang X. Learned Pseudo-Random Number Generator Based on Generative Adversarial Networks. Editors: Yang H, Lu R. Frontiers in Cyber Security, 517–530, Singapore, Springer, 2024.
4. [4] Bhattacharjee K, Das S. “A search for good pseudo-random number generators: Survey and empirical studies”. Computer Science Review, 45, 100471, 2022.
5. [5] Pasqualini L, Parton M. “Pseudo random number generation through reinforcement learning and recurrent neural networks”. Algorithms, 13(11), 307, 2020.
6. [6] Mukherjee P, Garg H, Pradhan C, Ghosh S, Chowdhury S, Srivastava G. “Best fit DNA-based cryptographic keys: the genetic algorithm approach”. Sensors, 22(19), 7332, 2022.
7. [7] Kalsi S, Kaur H, Chang V. “DNA cryptography and deep learning using genetic algorithm with NW algorithm for key generation”. Journal of Medical Systems, 42(1), 17, 2018.
8. [8] Basu S, Karuppiah M, Nasipuri M, Halder AK, Radhakrishnan N. “Bio-inspired cryptosystem with DNA cryptography and neural networks”. Journal of Systems Architecture, 94, 24-31, 2019.

9. [9] Reddy MI, Kumar AS, Reddy KS. “A secured cryptographic system based on DNA and a hybrid key generation approach”. Biosystems, 197, 104207, 2020.
10. [10] Sakr AS, Shams MY, Mahmoud A, Zidan M. “Amino acid encryption method using genetic algorithm for key generation”. Computers, Materials & Continua, 70(1), 123-134, 2022.
11. [11] Wu X, Han Y, Zhang M, Li Y, Cui S. “GAN-based pseudo random number generation optimized through genetic algorithms”. Complex & Intelligent Systems, 11(1), 31, 2025.
12. [12] De Bernardi M, Khouzani MHR, Malacaria P. “Pseudo-random number generation using generative adversarial networks”. ECML PKDD 2018 Workshops, Dublin, Ireland, 10–14 September 2018.
13. [13] Okada K, Endo K, Yasuoka K, Kurabayashi S. “Learned pseudo-random number generator: WGAN-GP for generating statistically robust random numbers”. PLOS ONE, 18(6), e0287025, 2023.
14. [14] Iqbal N, Khan M, Khurshid K, Hussain I. “An efficient hybrid encryption model based on deep convolutional neural networks, deoxyribonucleic acid computing and chaotic system”. Multimedia Tools and Applications, 82(9), 13881-13903, 2023.
15. [15] Park S, Kim K, Kim K, Nam C. “Dynamical pseudo-random number generator using reinforcement learning”. Applied Sciences, 12(7), 3377, 2022.
16. [16] Almardeny Y, Benavoli A, Boujnah N, Naredo E. “A reinforcement learning system for generating instantaneous quality random sequences”. IEEE Transactions on Artificial Intelligence, 4(3), 402-415, 2023.
17. [17] Lin T, Wang Y, Liu X, Qiu X. “A survey of transformers”.AI Open, 3, 111-132, 2022.
18. [18] Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser L, Polosukhin I. “Attention is all you need”. Advances in Neural Information Processing Systems 30, Long Beach, California, USA, 4-9 December 2017.
19. [19] Lian Z, Liu B, Tao J. “CTNet: Conversational transformer network for emotion recognition”. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 29, 985-1000, 2021.
20. [20] Han K, Xiao A, Wu E, Guo J, Xu C, Wang Y. “Transformer in Transformer”. Advances in Neural Information Processing Systems 34, 15908-15919, Virtual Conference, 6-14 December 2021.
21. [21] Jiang J, Ke L, Chen L, Dou B, Zhu Y, Liu J, Zhang B, Zhou T, Wei GW. “Transformer technology in molecular science”. Wiley Interdisciplinary Reviews: Computational Molecular Science, 14(4), e1725, 2024.
22. [22] Tao T, Doshi D, Kalra DS, He T, Barkeshli M. “(How) Can Transformers Predict Pseudo-Random Numbers?”. CoRR, abs/2502.10390, 2025. https://doi.org/10.48550/arXiv.2502.10390
23. [23] Parfentiev KV. “Generation of pseudo-random sequences using generative predictive neural network transformers”. 2025 7th International Youth Conference on Radio Electronics, Electrical and Power Engineering (REEPE), Moscow, Russian Federation, 8-10 April 2025.
24. [24] Joo W, Lee W, Park S, Moon IC. “Dirichlet variational autoencoder”. Pattern Recognition, 107, 107514, 2020.
25. [25] Gordon-Rodriguez E, Loaiza-Ganem G, Cunningham J. “The continuous categorical: a novel simplex-valued exponential family”. Proceedings of the 37th International Conference on Machine Learning, Virtual Event, 13-18 July 2020.
26. [26] Hobbhahn M, Kristiadi A, Hennig P. “Fast predictive uncertainty for classification with Bayesian deep networks”. Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence, 822-832, Eindhoven, The Netherlands, 1-5 August 2022.
27. [27] Kang G, Dong X, Zheng L, Yang Y. “PatchShuffle regularization”. CoRR, abs/1707.07103, 2017.
28. [28] Kasantikul R, Kusakunniran W, Wu Q, Wang Z. “Channel-shuffled transformers for cross-modality person re-identification in video”. Scientific Reports, 15(1), 15009, 2025.
29. [29] Feng S, Hu K, Fan E, Zhao L, Wu C. “Kalman filter for spatial-temporal regularized correlation filters”. IEEE Transactions on Image Processing, 30, 3263-3278, 2021.
30. [30] Shannon CE. “A mathematical theory of communication”. The Bell System Technical Journal, 27(3), 379-423, 1948.
31. [31] Li J, Zhang L, Li H, Ping Y, Xu Q, Wang R, Tan R, Wang Z, Liu B, Wang Y. “Integrated entropy-based approach for analyzing exons and introns in DNA sequences”. BMC Bioinformatics, 20(Suppl 8), 283, 2019.
32. [32] Pimenta-Zanon MH, Kashiwabara AY, Vanzela ALL, Lopes FM. “GRAMEP: an alignment-free method based on the maximum entropy principle for identifying SNPs”. BMC Bioinformatics, 26(1), 66, 2025.
33. [33] Afify YM, Sharkawy NH, Gad W, Badr N. “A new dynamic DNA-coding model for gray-scale image encryption”. Complex & Intelligent Systems, 10(1), 745-761, 2024.
34. [34] Fan W, Li T, Wu J, Wu J. “Chaotic color image encryption based on eight-base DNA-level permutation and diffusion”. Entropy, 25(9), 1268, 2023.
35. [35] Bassham III LE, Rukhin AL, Soto J, Nechvatal JR, Smid ME, Barker EB, Leigh SD, Levenson M, Vangel M, Banks DL, Heckert NA, Dray JF, Vo S. “A statistical test suite for random and pseudorandom number generators for cryptographic applications”. National Institute of Standards and Technology, Gaithersburg, Maryland, USA, NIST Special Publication, 800-22 Revision 1a, 2010.
36. [36] Barker EB, Kelsey JM. “Recommendation for random number generation using deterministic random bit generators”. National Institute of Standards and Technology, Gaithersburg, MD, USA, NIST Special Publication 800-90A Revision 1, 2015.
37. [37] Hernandez-Castro J, Barrero DF. “Evolutionary generation and degeneration of randomness to assess the indepedence of the Ent test battery”. 2017 IEEE Congress on Evolutionary Computation (CEC), Donostia-San Sebastián, Spain, 5-8 June 2017.
38. [38] Susanti BH, Jimmy J, Ardyani MW. “ENT randomness test on DM-PRESENT-80 and DM-PRESENT-128-based pseudorandom number generator”. 2021 4th International Seminar on Research of Information Technology and Intelligent Systems (ISRITI), Virtual Conference from Yogyakarta, Indonesia, 16 December 2021.
39. [39] Petitcolas FAP. Kerckhoffs’ Principle. Editors: Jajodia S, Samarati P, Yung M. Encyclopedia of Cryptography, Security and Privacy, 1342–1343, Cham, Switzerland, Springer, 2025.