Table 1 Error strengths for the configurations of the custom noise model used in Fig. 14. Depolarization (1Q) indicates the single qubit depolarising channel applied after each single-qubit gate, and similarly for 2Q for two-qubit gates. Configuration a) in bold is based on error rates published by IBM at the time of writing, as described in the main text

From: Robustness of quantum reinforcement learning under hardware errors