Discovery of optimal quantum codes via reinforcement learning

The design and optimization of quantum codes is generally hard because of the prohibitive computational cost. The recently introduced quantum lego framework provides a powerful method for generating complex quantum error-correcting codes (QECCs) out of simple ones. We gamify this process and unlock a new avenue for code design and discovery using reinforcement learning (RL). Already for simple code measures and modest qubit numbers, we produce codes that are both optimal and novel. Moreover, we argue that this framework is both scalable, when combined with efficient tensor contraction methods, and flexible, since we can specify arbitrary properties of the code to be optimized. We train on two such properties, maximizing the code distance, and minimizing the probability of logical error under biased Pauli noise. For the first, we show that the trained agent identifies ways to increase code distance beyond naive concatenation, saturating the linear programming bound for CSS (Calderbank, Shor, Steane) codes on 13 qubits. With a learning objective to minimize the logical error probability under biased Pauli noise, we find the best-known CSS code at this task for $\ensuremath{\lesssim}20$ qubits. Compared to other (locally deformed) CSS codes, including Surface, XZZX, and two-dimensional Color codes, our $[[17,1,3]]$ code construction actually has lower adversarial distance, yet better protects the logical information, highlighting the importance of QECC desiderata. Lastly, we comment on how this RL framework can be used in conjunction with physical quantum devices to tailor a code without explicit characterization of the noise model.