Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

International Conference on Communication and Electronics Systems (ICCES)

Post-Quantum Secure Network Design using Reinforcement Learning Approaches

M. Tamilarasi, R.K. Soundarayaa, P. Velayutha Perumal, J. Janorious Hermia, C. Balasubramanian, R. Saravanakumar

Published Oct 28, 2025• 0 citations
Publisher
DOI
10.1109/icces67310.2025.11336857

Related papers

Accelerating Security and Efficiency of the BIKE (Bit Flipping Key Encapsulation) through Reinforcement Learning for Industrial IoT

Samir Gherbi, Soumya Banerjee, Yulliwas Ameur

In the context of the recent advancements in post-quantum cryptography, the BIKE (Bit Flipping Key Encapsulation) algorithm has emerged as a prominent candidate for quantum-resistant encryption in resource-constrained environments, particularly within Industrial IoT (IIoT). This chapter explores the integration of reinforcement learning (RL) to dynamically enhance the security and efficiency of the BIKE algorithm. The RL framework allows for real-time adaptation of BIKE's parameters based on network conditions and threat levels, optimizing key generation, reducing computational intensity, and ensuring robust security measures. By leveraging RL, we address the challenges posed by limited resources in IIoT applications and propose a scalable, adaptable solution for maintaining high levels of security against evolving quantum threats. Our experiments demonstrate improved performance and resilience of the BIKE algorithm, offering a valuable contribution to the ongoing development of secure and efficient cryptographic systems in IIoT environments.

Wiley eBooks•Oct 30, 2026•0 citations

Quantum Neural Network Design via Quantum Deep Reinforcement Learning

Anca Ioana Muscalagiu

International Joint Conference on Computational Intelligence•Jan 1, 2024•0 citations

A reinforcement learning approach for quantum state engineering

Jelena Mackeprang, Durga Bhaktavatsala Rao Dasari, Jörg Wrachtrup

arXiv:1908.05981•Quantum Machine Intelligence•May 11, 2020•56 citations

Quantum Machine Learning for Secure Key Generation in Wireless Communication: Addressing Limitations of Classical Cryptographic Protocols

The many forms of wireless communication used these days have created a greater demand for strong and flexible cryptographic methods to protect from new security risks that can happen as quantum computing advances.Diffie-Hellman and RSA approaches have been basic to cryptography, but work on these can now be weakened by attack methods such as Shor's and Grover's algorithms following quantum developments which undermines many post-quantum schemes.In our research, we develop a plan using Quantum Machine Learning (QML) to help ensure the security of future wireless networks.As a result, this design pairs quantum theory with reinforcement learning to create strong keys that cannot be accessed by attackers.Value is drawn from quantum entanglement and random measurements of qubits to create key materials needed for physically unclonable keys and at the same time, reinforcement agents support optimizing agreement policies based on variations detected in the channel.Simulations were run using Rayleigh fading and multi-SNR using Qiskit and TensorFlow.The system achieved entropy greater than 95%, agreed on important issues more than 98% successfully and brought latency down to just 0.13 seconds, outshining both classical and post-quantum cryptography in performance, flexibility and safety from attacks.Moreover, the presented encryption solution managed to resist hackers, even if the environment was noisy.As a quantum-AI combination, it is well-positioned to serve as a platform to safely communicate over wireless networks.It builds the basis for examining embedded QML hardware versions of IoT, vehicle networks and smart grid communications, since fast security and reliable performance without high delay are required.

National Journal Of Antennas And Propagation•Jan 1, 2025•0 citations

Quantum-Aware Reinforcement Learning Framework for Adaptive Quantum Key Distribution

Palisetty Rudrabhishek, Medikonda Mahathi Anushma, Pallapothu Moulika +1 more

Quantum technology is revolutionizing communication networks, which are guaranteed to be secure through Quantum Key Distribution (QKD). Classical cryptography relies on the assumption that certain computational problems are hard, and QKD offers information-theoretic security based on the laws of quantum physics. Nevertheless, a practical realization is challenging. Photon loss, environmental noise, and time-varying channel conditions can all significantly reduce the secure key generation rate . Standard QKD protocols are based on fixed parameters and passive loss compensation, and thus, are not capable of modulating network conditions in real time. In this paper, we propose a Quantum-Aware Reinforcement Learning Framework (QARLF). This is a combined AI-QKD architecture for self-learning, adaptive and predictive quantum networks. Our scheme consists of a Quantum Channel Sensor Layer that continuously monitors the channel conditions, a Reinforcement Learning (RL) Agent that learns to predict the optimal system parameters, and an Adaptive Control Unit (ACU) that conducts real-time adaptation. The RL agent is trained to employ a quantum-adaptive reward function considering the growth of the secure key rate, the reduction of the quantum bit error rate (QBER), and the operational cost. The agent uses Q-learning with a Boltzmann exploration strategy to progressively enhance its policy. The results demonstrate that QARLF can track channel variations effectively and adapt parameters in real time. It significantly boosts the secure key rates over the conventional static or reactive approach. This is a step towards self-adapting, quantum-savvy networks.

IEEE International Conference on Recent Advances in Electronics, Communication, Computing, Automation and Power (RECCAP)•May 1, 2026•0 citations

Quantum Reinforcement Learning

Daoyi Dong, Chunlin Chen, Zonghai Chen

Lecture notes in computer science•Jan 1, 2005•44 citations
In the context of the recent advancements in post-quantum cryptography, the BIKE (Bit Flipping Key Encapsulation) algorithm has emerged as a prominent candidate for quantum-resistant encryption in resource-constrained environments, particularly within Industrial IoT (IIoT). This chapter explores the integration of reinforcement learning (RL) to dynamically enhance the security and efficiency of the BIKE algorithm. The RL framework allows for real-time adaptation of BIKE's parameters based on network conditions and threat levels, optimizing key generation, reducing computational intensity, and ensuring robust security measures. By leveraging RL, we address the challenges posed by limited resources in IIoT applications and propose a scalable, adaptable solution for maintaining high levels of security against evolving quantum threats. Our experiments demonstrate improved performance and resilience of the BIKE algorithm, offering a valuable contribution to the ongoing development of secure and efficient cryptographic systems in IIoT environments.
The many forms of wireless communication used these days have created a greater demand for strong and flexible cryptographic methods to protect from new security risks that can happen as quantum computing advances.Diffie-Hellman and RSA approaches have been basic to cryptography, but work on these can now be weakened by attack methods such as Shor's and Grover's algorithms following quantum developments which undermines many post-quantum schemes.In our research, we develop a plan using Quantum Machine Learning (QML) to help ensure the security of future wireless networks.As a result, this design pairs quantum theory with reinforcement learning to create strong keys that cannot be accessed by attackers.Value is drawn from quantum entanglement and random measurements of qubits to create key materials needed for physically unclonable keys and at the same time, reinforcement agents support optimizing agreement policies based on variations detected in the channel.Simulations were run using Rayleigh fading and multi-SNR using Qiskit and TensorFlow.The system achieved entropy greater than 95%, agreed on important issues more than 98% successfully and brought latency down to just 0.13 seconds, outshining both classical and post-quantum cryptography in performance, flexibility and safety from attacks.Moreover, the presented encryption solution managed to resist hackers, even if the environment was noisy.As a quantum-AI combination, it is well-positioned to serve as a platform to safely communicate over wireless networks.It builds the basis for examining embedded QML hardware versions of IoT, vehicle networks and smart grid communications, since fast security and reliable performance without high delay are required.
Quantum technology is revolutionizing communication networks, which are guaranteed to be secure through Quantum Key Distribution (QKD). Classical cryptography relies on the assumption that certain computational problems are hard, and QKD offers information-theoretic security based on the laws of quantum physics. Nevertheless, a practical realization is challenging. Photon loss, environmental noise, and time-varying channel conditions can all significantly reduce the secure key generation rate . Standard QKD protocols are based on fixed parameters and passive loss compensation, and thus, are not capable of modulating network conditions in real time. In this paper, we propose a Quantum-Aware Reinforcement Learning Framework (QARLF). This is a combined AI-QKD architecture for self-learning, adaptive and predictive quantum networks. Our scheme consists of a Quantum Channel Sensor Layer that continuously monitors the channel conditions, a Reinforcement Learning (RL) Agent that learns to predict the optimal system parameters, and an Adaptive Control Unit (ACU) that conducts real-time adaptation. The RL agent is trained to employ a quantum-adaptive reward function considering the growth of the secure key rate, the reduction of the quantum bit error rate (QBER), and the operational cost. The agent uses Q-learning with a Boltzmann exploration strategy to progressively enhance its policy. The results demonstrate that QARLF can track channel variations effectively and adapt parameters in real time. It significantly boosts the secure key rates over the conventional static or reactive approach. This is a step towards self-adapting, quantum-savvy networks.