Generative Adversarial Networks (GANs) have surfaced as a transformative approach in the domain of cryptography, introducing a novel paradigm where two neural networks, the generator (akin to Alice) and the discriminator (akin to Bob), are pitted against each other in a cryptographic setting. A third network, representing Eve, attempts to decipher the encrypted information. The efficacy of this encryption–decryption process is deeply intertwined with the choice of activation functions employed within these networks. This study conducted a comparative analysis of four widely used activation functions within a standardized GAN framework. Our recent explorations underscore the superior performance achieved when utilizing the Rectified Linear Unit (ReLU) in the hidden layers combined with the Sigmoid activation function in the output layer. The non-linear nature introduced by the ReLU provides a sophisticated encryption pattern, rendering the deciphering process for Eve intricate. Simultaneously, the Sigmoid function in the output layer guarantees that the encrypted and decrypted messages are confined within a consistent range, facilitating a straightforward comparison with original messages. The amalgamation of these activation functions not only bolsters the encryption strength but also ensures the fidelity of the decrypted messages. These findings not only shed light on the optimal design considerations for GAN-based cryptographic systems but also underscore the potential of investigating hybrid activation functions for enhanced system optimization. In our exploration of cryptographic strength and training efficiency using various activation functions, we discovered that the “ReLU and Sigmoid” combination significantly outperforms the others, demonstrating superior security and a markedly efficient mean training time of 16.51 s per 2000 steps. This highlights the enduring effectiveness of established methodologies in cryptographic applications. This paper elucidates the implications of these choices, advocating for their adoption in GAN-based cryptographic models, given the superior results they yield in ensuring security and accuracy.

This research introduces a novel hybrid cryptographic framework that combines traditional cryptographic protocols with advanced methodologies, specifically Wasserstein Generative Adversarial Networks with Gradient Penalty (WGAN-GP) and Genetic Algorithms (GA). We evaluated several cryptographic protocols, including AES-ECB, AES-GCM, ChaCha20, RSA, and ECC, against critical metrics such as security level, efficiency, side-channel resistance, and cryptanalysis resistance. Our findings demonstrate that this integrated approach significantly enhances both security and efficiency across all evaluated protocols. Notably, the AES-GCM algorithm exhibited superior performance, achieving minimal computation time and robust side-channel resistance. This study underscores the potential of leveraging machine learning and evolutionary algorithms to advance cryptographic protocol security and efficiency, laying a robust foundation for future advancements in cybersecurity.

The integration of Generative Adversarial Networks (GANs) with Genetic Algorithms (GAs) represents a novel approach to enhancing cryptographic methods, particularly in addressing challenges posed by quantum computing and increasingly sophisticated cyber threats. This research focuses on improving encryption strength, adaptability, and robustness against decryption attempts. By leveraging the optimization capabilities of GAs to evolve neural network architectures within a GAN framework, we significantly enhance the generator's ability to produce secure, quantum-resistant encryptions. The genetic algorithm optimized both the generator and discriminator networks over 300 generations, reducing generator loss from an initial 0.78 to a stable 0.65, while increasing discriminator loss, indicating improved encryption complexity. This study demonstrates the feasibility of using evolutionary techniques and adversarial training to create a dynamic, self-evolving cryptographic system, providing a foundation for future cryptographic innovations in quantum-resistant security. The methodology combines GA-driven network optimization and GAN-based adversarial training to address the challenges of quantum decryption and advanced adversarial attacks, setting new benchmarks for cryptographic security.

Generative adversarial networks (GANs) have gained significant attention in recent years due to their ability to generate realistic and diverse images, videos leading other media contents. However, GANs can also be used for other applications such as cryptography. GAN cryptography is a novel approach to encryption that uses GANs to generate the encrypted messages. In GAN cryptography, a generator network creates a message that is then encrypted using a discriminator network. The discriminator network acts as a decryption function, and only authorized users with the correct decryption key can decrypt the message. The use of GANs for cryptography provides several advantages, such as the ability to generate secure and random encryption keys and the potential for more efficient encryption and decryption. This book chapter provides an overview of GAN cryptography, including the basic principles of GANs and their application in cryptography. It also covers the advantages and limitations of GAN cryptography, as well as its potential applications in areas such as secure communication, data protection, and authentication. This chapter also discusses the current research and future research direction for GAN cryptography, including challenges and opportunities for further development and improvement of the techniques. <!----> <!----> <!----> <!---->

Next Generation Ethernet Passive Optical Networks (NG-EPONs) have emerged as a leading choice for global network connectivity due to their cost-effectiveness, enhanced security, and energy efficiency. As data demands continue to surge with technological advancements, the need for efficient bandwidth management and wavelength selection in NG-EPONs becomes paramount. This paper presents two innovative algorithms—Tensile Wavelength and Dynamic Bandwidth Allocation (TW-DBA) and Premier Dynamic Bandwidth Allocation (PDBA)—designed to optimize bandwidth allocation and maintain Quality of Service (QoS) under varying network conditions. The TW-DBA algorithm achieves a remarkable throughput of 2.34 Gbps, driven by its dynamic wavelength selection mechanism that accounts for factors such as ONU-OLT distance, power availability, and bandwidth demand. Comparative analysis reveals that TW-DBA outperforms existing algorithms like Flexible Wavelength (FW), First-Fit, and Water Filled, both in computational efficiency and resource allocation. Furthermore, the PDBA algorithm demonstrates a minimum blockage probability of 0.0025 for limited ONU scenarios and 0.005 in unlimited scenarios, ensuring uniform bandwidth distribution. Simulation results confirm the superior performance and effectiveness of the proposed models, positioning them as robust solutions for the evolving demands of NG-EPONs.