IEEE International Conference on Quantum Computing and Engineering (QCE)
Reinforcement Learning for Gate Synthesis in Noisy Quantum Systems
Amara Katabarwa, Collin Farquhar, Hyeongrak Choi, Marc Davis, Dirk Robert 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.