Reinforcement learning for error-robust control on cloud-based superconducting hardware [Part II]

The noisy nature of today’s quantum hardware limits the ability to realize functioning and useful
quantum computers. Yet, a careful design of the systems’ controls allows to narrow the gap between current and desired hardware capabilities. In this work, we study a black-box optimization technique based on reinforcement learning. We show that by employing reinforcement learning, where intermediate information is used in order to optimize a long term goal, we are able to generate two-qubits gates which outperform, when utilized on a real device, the existing model-based optimized pulses. We demonstrate the performance of our learner on IBM quantum hardware accessed using Qiskit Pulse analog programming. The entire learning process occurs on the quantum device itself, aiming to suit the low-level gate implementation to the underlying details of the specific hardware. This allows gate optimization without any prior knowledge or assumptions on the noise model, hardware limitations, parasitic couplings or any other undesired effect that may exist in a real device.