Machine learning for efficient generation of universal hybrid quantum computing resources

We present numerical results from simulations using deep reinforcement learning to control a measurement-based quantum processor-a time-multiplexed optical circuit sampled by photon-number-resolving detection-and find it generates squeezed cat states quasideterministically, with an average success rate of 98%, far outperforming all other proposals.Since squeezed cat states are deterministic precursors to the Gottesman-Kitaev-Preskill (GKP) bosonic error code, this is a key result for enabling fault tolerant photonic quantum computing.Informed by these simulations, we also discovered a one-step quantum circuit of constant parameters that can generate GKP states with high probability, though not deterministically.