arXiv:2609.35016
Experimental quantum-computing-enhanced sensing using Grover's algorithm
Mathieu Ouellet, Purnendu Sen, Xiangqin Wang, Saswata Roy, Xingrui Song, Vladimir Kremenetski, Sridhar Prabhu, Valla Fatemi, Peter L. McMahon
Abstract
The combination of quantum sensing with quantum computing to provide an enhancement over conventional quantum sensing has recently emerged as a promising potential application of quantum computing that could give advantages without needing large-scale or fault-tolerant hardware. In this work, we report an experimental demonstration of a recent theoretical proposal to repurpose Grover's search algorithm to improve the ability to detect signals with unknown frequency within a large detection bandwidth. Our experiments were based on a system comprising a single superconducting qubit coupled to a single superconducting cavity, highlighting the modest hardware requirements for realizing the protocol. We found that Grover-based sensing was able to outperform the natural non-Grover baseline for our experimental platform for detection bandwidths $>$10~MHz, with an advantage that empirically grew superlinearly with the bandwidth beyond that break-even point. The use of the Grover-based protocol reduced the amount of signal that needed to be sensed to make an accurate detection decision by more than 10$\times$ for choices of larger detection bandwidth and higher desired detection accuracy. Our results provide a proof-of-principle validation that Grover-based quantum computational sensing can be realized in near-term hardware and provide a metrological advantage well beyond break-even in spite of the additional protocol complexity.
AI Summary
The paper experimentally tests whether Grover amplitude amplification can reduce the signal-interaction cost of deciding whether an unknown-frequency microwave signal lies within a continuous detection band. The sensor is a superconducting transmon dispersively coupled to a cavity, with photon-number-dependent qubit transitions encoding frequency channels. A coherent cavity state supplies the initial frequency coverage. Signal-driven qubit rotations mark resonant Fock components, while displacements and a vacuum-selective phase gate implement reflection about the initial state. A final signal-driven rotation maps the amplified response onto qubit excitation. Finite-bandwidth pulses bridge the discrete transitions to provide continuous coverage. The comparison uses a coherent-probe baseline without Grover iterations, adjusting probe amplitudes to match detection bandwidths. Resource cost is $n_{\mathrm{shots}}(2N+1)$ signal interrogations, rather than total runtime. Classification costs are estimated by resampling measured binary outcomes and training and validating a classifier. One- and two-iteration protocols outperform the baseline above approximately $8$ and $13\,\mathrm{MHz}$, respectively; two iterations outperform one above approximately $22\,\mathrm{MHz}$. Reported cost reductions reach about $4\times$ at $75\%$ classification accuracy and $18\times$ at $90\%$. The stated limitations include synchronized, repeatedly available signals; no proof that the baseline is optimal; nonuniform sensitivity from coherent rather than uniform probes; bandwidths limited to tens of MHz; and coherence limiting the advantage experiments to two Grover iterations. Input attenuation limits sensitivity to room-temperature signals, and applicability to practical sensing remains unresolved. The results do not establish an optimal non-Grover comparison or an end-to-end runtime advantage.
Error found: The equal-bandwidth protocol comparisons generally use different band centers, and the paper does not establish equivalent performance for the same prescribed frequency window. The shot-cost appendix incorrectly assigns a 1/p² measurement cost to distinguishing excitation probabilities zero and p, and assumes a constant amplified peak where its fixed-iteration ideal response eventually decreases, weakening the justification for its fitted resource curves and extrapolated bandwidth limits.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 18 pages, 11 figures