Reinforcement Learning for Gate Synthesis in Noisy Quantum Systems

Amara Katabarwa, Collin Farquhar, Hyeongrak Choi, Marc Grau Davis, Dirk R. Englund, Yudong Cao, Mekena Metcalf

Abstract: We present a gate synthesis application of the Deep Deterministic Policy Gradient (DDPG) algorithm for continuous control in a noisy quantum environment. Gate synthesis plays a crucial role in quantum computing, and optimizing control strategies for gate synthesis is of significant interest. The proposed approach leverages a DNN-based actor-critic architecture to optimize gate synthesis performance. The effectiveness of the algorithm is evaluated through simulations in a noiseless and noisy quantum environment, demonstrating its ability to achieve high-fidelity gate synthesis, $\mathcal{O}(10^{-3})$, despite the presence of Markovian damping terms. The results highlight the potential of DDPG and reinforcement learning techniques for addressing continuous control challenges in quantum computing applications.

External IDs:dblp:conf/qce/KatabarwaFCDECM23