This book explores the combination of Reinforcement Learning and Quantum Computing in the light of complex attacker-defender scenarios. Reinforcement Learning has proven its capabilities in different challenging optimization problems and is now an established method in Operations Research. However, complex attacker-defender scenarios have several characteristics that challenge Reinforcement Learning algorithms, requiring enormous computational power to obtain the optimal solution.
The upcoming field of Quantum Computing is a promising path for solving computationally complex problems. Therefore, this work explores a hybrid quantum approach to policy gradient methods in Reinforcement Learning. It proposes a novel quantum REINFORCE algorithm that enhances its classical counterpart by Quantum Variational Circuits. The new algorithm is compared to classical algorithms regarding the convergence speed and memory usage on several attacker-defender scenarios with increasing complexity. In addition, to study its applicability on today’s NISQ hardware, the algorithm is evaluated on IBM’s quantum computers, which is accompanied by an in-depth analysis of the advantages of Quantum Reinforcement Learning.
Tabella dei contenuti
Motivation: Complex Attacker-Defender Scenarios – The eternal conflict., The Information Game – A special Attacker-Defender Scenario., Reinforcement Learning and Bellman’s Principle of Optimality., Quantum Reinforcement Learning – Connecting Reinforcement Learning and Quantum Computing.- Approximation in Quantum Computing.- Advanced Quantum Policy Approximation in Policy Gradient Rein-forcement Learning.- Applying Quantum REINFORCE to the Information Game.- Evaluating quantum REINFORCE on IBM’s Quantum Hardware.- Future Steps in Quantum Reinforcement Learning for Complex Scenarios.- Conclusion.
Circa l’autore
About the author
Leonhard Kunczik obtained his Dr. rer. nat. in 2021 in Quantum Reinforcement Learning from the Universität der Bundeswehr München as a member of the COMTESSA research group. Now, he continues his research as a project leader at the forefront of Quantum Machine Learning and Optimization in the context of Operations Research and Cyber Security.
Altri ebook dello stesso autore / Editore
16.795 Ebook in questa categoria
