Quantum sensing is one of the key applications in quantum information processing. In typical protocols such as Ramsey-type measurements, an external field shifts the qubit frequency, and this shift can be detected to estimate the field strength. Improving sensitivity requires increasing the number of qubits, while enhancing spatial resolution necessitates confining qubits within a small volume. However, a high qubit density leads to strong qubit-qubit interactions, which limits the achievable dynamic range.
In this work, we propose a quantum circuit learning approach—a technique from quantum machine learning—to overcome this limitation and enhance the dynamic range of quantum sensing. After exposing the qubits to a target magnetic field, we apply parameterized quantum control pulses and measure a suitable observable. By optimizing the pulse parameters, we numerically demonstrate that the dynamic range of the sensor is significantly improved. Our results suggest that quantum circuit learning offers a promising strategy for high-performance quantum sensing in densely packed qubit systems.