arXiv:2609.28631
Universal Counterdiabatic Quantum Sensing
Stewart Morawetz, Anatoli Polkovnikov
Abstract
One of the primary applications of quantum mechanics is leveraging coherence to make incredibly sensitive measurements. One way to push the boundary of this sensitivity even further is to initially prepare a system in an entangled state, though the fragility of these states limits their usefulness. In this work, we propose a new approach by which sensitivity comes during the preparation of an entangled state, rather than first preparing the state and then using it for sensing. An adiabatic state preparation process is then accelerated by universal counterdiabatic driving, which is more robust to uncertainty in microscopic parameters than standard shortcuts to adiabaticity. Using a toy Lipkin-Meshkov-Glick model, we show that this approach can also lead to sensitivity beyond the Standard Quantum Limit.
AI Summary
Can a weak field be sensed during entangling state preparation, rather than in a separate interrogation stage? This paper studies that approach using approximate universal counterdiabatic (CD) driving. Unlike system-specific CD optimization, the universal construction fits an odd polynomial to $1/\omega$ over a chosen excitation-frequency window, aiming to tolerate uncertainty in microscopic parameters. The demonstration uses the extensive, all-to-all Lipkin–Meshkov–Glick model with $N=2S$ spins. A smooth ramp interpolates from a transverse-field Hamiltonian to a ferromagnetic interaction plus the unknown longitudinal field $h$. Numerical finite-time evolution determines the final magnetization response, with sensitivity estimated through $\Delta h_\alpha=\Delta S_\alpha/|\partial_h\langle S_\alpha\rangle|$. Longitudinal readout improves with CD driving but retains standard-quantum-limit scaling, $\Delta h_z\propto S^{-1/2}$. Transverse readout achieves fitted scaling approximately $\Delta h_y\sim S^{-2/3}$ for sufficiently high CD expansion order, with favorable durations $\tau\sim1/\chi$. Finite-difference quantum Fisher information calculations independently show beyond-SQL scaling; one example gives $F_Q\sim S^{1.69}$ at $\tau\approx4$ in units with $\chi=1$. The transverse signal arises from field-dependent departures from equilibrium and vanishes under exact ground-state following: imperfect CD driving is essential. Noise simulations incorporating collective decay and individual dephasing show that weak noise can preserve the short-duration signal while destroying slow-protocol sensitivity. The demonstrated scaling is numerical and model-specific. Increasing CD order can introduce high-frequency excitation problems; Floquet implementations are suggested as a remedy but are not demonstrated here.
Central claim may not hold: The direct nested-commutator CD operator explicitly depends on the unknown field even when its polynomial coefficients do not, and the paper does not specify a fixed nominal-field implementation or demonstrate an equivalent unknown-field-compatible implementation for its reported sensing results. The abstract's claim of greater robustness to microscopic uncertainty than standard shortcuts lacks a quantitative comparison or demonstrated guarantee in the presented analysis. The printed individual-dephasing generator is not trace preserving because its subtraction term is outside the sum over spins.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 8 pages, 5 figures