Leveraging Machine Learning for Performance Optimization in Post-Quantum Cryptographic Protocols

John Olusegun Fajinmi

Published Apr 14, 2025• 0 citations

Abstract

The advent of quantum computing presents a potential threat to conventional cryptographic systems, prompting the urgent need for post-quantum cryptography (PQC) solutions. However, the complexity of these cryptographic protocols often leads to inefficiencies in performance, hindering their widespread adoption. This paper explores the integration of machine learning (ML) techniques to optimize the performance of post-quantum cryptographic protocols. We propose an innovative framework where ML algorithms are leveraged to enhance the efficiency of key generation, encryption, decryption, and signature processes in PQC schemes. By utilizing data-driven approaches such as reinforcement learning and neural networks, we aim to reduce computational overhead and improve latency, while maintaining the robustness and security of post-quantum algorithms. Experimental results demonstrate significant performance improvements over traditional optimization methods, showcasing the potential of ML in advancing the practicality of PQC protocols. This work lays the foundation for a new class of cryptographic systems capable of withstanding quantum threats while maintaining high operational efficiency.

DOI: 10.22541/au.174466014.40471065/v1

Related papers

Cryptography Innovations for Securing Data in the Quantum Computing Era: Integrating Machine Learning for Enhanced Security

Sheetal Temara, L. Bhagyalakshmi, Sanjay Kumar Suman +3 more

The rise of quantum computing introduces substantial risks to traditional cryptographic protocols, which are vulnerable to quantum decryption methods such as Shor’s algorithm. To address these emerging threats, this paper proposes an innovative cryptographic framework that integrates machine learning (ML) techniques to enhance data security in the quantum computing era. Our approach leverages ML-driven anomaly detection, adaptive key management, and predictive analytics to create a flexible and resilient cryptographic defense. The system’s anomaly detection module utilizes neural networks to identify potential quantum-based decryption attempts, while reinforcement learning optimizes key generation and distribution in response to detected threats. Experimental results demonstrate that the proposed ML-augmented framework significantly improves anomaly detection accuracy and reduces vulnerability to quantum decryption attempts by dynamically adjusting cryptographic parameters. These findings underscore the potential of machine learning to strengthen cryptographic systems, making them adaptable to the advanced threats posed by quantum computing.

International Conference on Computer, Electrical & Communication Engineering (ICCECE)Feb 7, 202511 citations

Machine Learning-Enhanced Post-Quantum Cryptographic Framework For Secure Data Protection In The Quantum Computing Era

Assistant Professor, Mrs.S.Surya Sri, Pasagodugula Ramya +4 more

The rapid advancement of quantum computing poses a significant threat to traditional cryptographic systems that rely on computational hardness assumptions such as integer factorization and discrete logarithms. Quantum algorithms, particularly Shor's algorithm, have the potential to break widely used cryptographic standards including RSA and Elliptic Curve Cryptography. To address these emerging challenges, this research proposes a machine learning-enhanced post-quantum cryptographic framework designed to provide adaptive and resilient data security in the quantum computing era. The proposed framework integrates post-quantum cryptographic algorithms with intelligent machine learning mechanisms to improve the robustness of encryption systems against quantum-based attacks. The system incorporates three major components: a post-quantum encryption module, a machine learning–based anomaly detection module, and an adaptive key management system powered by reinforcement learning. The anomaly detection component utilizes neural network models to monitor encrypted data streams and identify abnormal patterns that may indicate potential decryption attempts. In addition, reinforcement learning dynamically adjusts cryptographic key parameters and rotation schedules to reduce the risk of key compromise. Experimental evaluation demonstrates that the proposed framework significantly improves detection accuracy, reduces key management latency, and enhances resilience against simulated quantum decryption attacks when compared with conventional cryptographic systems. The integration of machine learning enables the framework to dynamically adapt to evolving threats while maintaining strong encryption standards. The results highlight the potential of combining artificial intelligence with post-quantum cryptography to build adaptive, intelligent, and future-proof security systems capable of protecting sensitive data in the presence of powerful quantum computing technologies.\\n\\n

Jan 1, 20260 citations

Using Machine Learning to Enhance PostQuantum Cryptographic Algorithms

MmaduekweEbuka Paul, Femi Osholake Femi Osholake, Je ersonEderhion Je ersonEderhion +1 more

The incoming need for defense against quantum computer attacks has motivated researchers to prioritize post-quantum cryptography (PQC) because this approach develops encryption which quantum computers cannot break. A great number of PQC algorithms create complex challenges regarding both computational requirements and key length as well as potential security weaknesses that require optimization for actual implementation. ML technology provides an approach to enhance the performance of post-quantum cryptography systems while securing their operations and making them more adaptable. The key generation process becomes more efficient while parameter selection reaches optimized results thanks to ML techniques which strengthen attack resilience through implementation protection against anomalies and side-channel attacks. The paper explores the ML and PQC connection while presenting potential ML applications that enhance the effectiveness of PQC methodologies including lattice-based schemes hash-based schemes and multivariate-quadratic and code-based approaches. The research document outlines integration problems and prospective directions between PQC and ML which incorporate considerations for security attacks and operational and computational constraints. Through their union PQC and ML enable the creation of additional secure cryptographic technologies that provide efficient scaling potential to overcome quantum computing threats.

International Journal of Advances in Engineering and ManagementFeb 1, 202515 citations

Post-Quantum Cryptography and Quantum Machine Learning for Resilient Encryption in AI-Driven Cybersecurity

Krishna Kumar, Sathea Sree S, Raj Laxmi

The rapid evolution of quantum computing poses a significant challenge to traditional encryption systems, with the potential to compromise the security of sensitive digital infrastructures. Post-Quantum Cryptography (PQC) has emerged as a vital field, aiming to develop encryption algorithms resilient to quantum attacks. Simultaneously, Quantum Machine Learning (QML) is revolutionizing the way machine learning models process data, offering new avenues for enhancing cybersecurity measures. This chapter explores the integration of PQC and QML to create robust, future-proof encryption systems capable of adapting to the evolving threat landscape. By examining hybrid models that combine the quantum resistance of PQC with the adaptability and efficiency of QML, this work highlights the potential for creating scalable and efficient cryptographic frameworks. The challenges and opportunities presented by the intersection of PQC and QML are discussed, with a focus on resource-constrained environments where computational power and memory are limited. Through this analysis, the chapter offers a comprehensive roadmap for advancing AI-driven, quantum-resistant cybersecurity solutions, addressing both theoretical advancements and practical implementation challenges.

Mar 4, 20252 citations

Post quantum encryption algorithms using AI/ML

Abeer Fatima, Harsh Vardhan Shukla, Shubham Kumar +1 more

Classical cryptographic have a threat because of the emergence of quantum computing, and hence there is an urgent need to research post-quantum cryptography. To evaluate the potential of post-quantum encryption algorithms in countering quantum threats, here the study sees how they can be integrated with artificial intelligence (AI) and machine learning (ML). For their theoretical quantum resistance, the study looks at established post-quantum cryptography methods such as lattice-based, hash-based, and multivariate cryptographic schemes. A review of current research also simulation-based research is also taken into account to review the potential of AI/ML in bolstering security protocols and simplifying encryption processes. The key takeaway is that machine learning and AI do have actual potential to enhance post-quantum encryption through increasing its efficiency, flexibility, and resistance to future quantum attacks. AI also assists in real-time responsiveness in avoiding possible vulnerabilities. Our work contributes to the growing literature by combining insights from cryptographic and AI/ML research, providing a comprehensive evaluation of post-quantum encryption methods. This study emphasizes the highest importance of adopting hybrid cryptography techniques in order to advance information security in the quantum era, leading to further research in practical applications.

CRC Press eBooksNov 6, 20250 citations

AI-BASED POST-QUANTUM CRYPTOGRAPHIC KEY EXCHANGE PROTOCOLS

Narayana Gaddam

One of the promising solutions to overcome this quantum computing attack on conventional encryption protocol is the integration of Artificial Intelligence (AI) withPost Quantum Cryptography (PQC).In this research, AI-driven approaches are investigated in improving post quantum cryptographic key exchange protocol by being efficient, secure and scalable.Based on lattice based cryptography resistant against quantum [2], this work makes use of machine learning models in order to assist in optimizing, or predicting optimal parameter choices, for better key exchange in practice.The system proposed consists of the Learning with Errors (LWE) problem for increased security [4] and a hybrid AI model that minimizes the computation overhead in key exchange protocols [3].The system dynamically adjusts cryptographic parameters based on the side channel attacks that are feasible and up to date, making the system resilient to side channel attacks, and also reduces the latency of secure communication through the incorporation of deep learning algorithms.The key findings show that AI driven model significantly improves the key exchange speed by 27 percent over traditional PQC techniques, and still provides robust security guarantees [5].To sum up, this research brings a significant advancement in PQC protocols by combining the AI techniques to predict and defend against cryptographic VTEs [7].In the future, this approach will be generalized to other quantum-resistant protocols (including SIKE Narayana Gaddam https://iaeme.com/Home/journal/IJCET

INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING & TECHNOLOGYJun 2, 20222 citations