Noise-resilient quantum algorithm design for adiabatic quantum computers

We have developed a novel architecture for automated design of quantum adiabatic algorithm by combining deep reinforcement learning and simulation annealing techniques. Through numerical test, we find that our architecture is adaptable for noise-resilient quantum algorithm design, which is of great current demand in the era of noisy intermediate-scale quantum computing. We show that in solving Grover search, our method automatically reaches the optimal performance in terms of time-complexity scaling. Our approach offers a recipe to make the noise-resilient adiabatic quantum computing, and is also generalizable to optimizing quantum simulations.