Publication Details

Quantum error correction (QEC) is needed for building reliable and scalable quantum computers because of the susceptibility of quantum states to decoherence and operational noise. Reinforcement learning (RL) has recently emerged as an adaptive approach for quantum control and error mitiga- tion;however, existing RL-based QEC methods are largely re- stricted to small qubit systems and are often dependent on specific hardware architectures. We present a RL–enabled QEC ap- proach for a five qubit quantum system. The proposed approach relies exclusively on syndrome measurements and reward feed- back, ensuring architecture independence and portability across quantum platforms. The framework is evaluated using Qiskit and Cirq simulators under realistic noise models. Experimental results demonstrate improved logical qubit lifetimes and reduced error rates compared to static and smaller qubit correction strategies.