arXiv:2609.31221
Distinct Feedback-Strength Requirements for Quantum-State Ensemble Preparation under Channel-Equivalent Monitoring
Haitao Huang, Qinglin Zhao, Xiaoyu Li
Abstract
Different measurements of the same environment can preserve the average quantum channel yet change the feedback strength needed to prepare a state ensemble. We establish this distinction for a multiqubit register under complete Pauli monitoring, targeting the uniform ensemble of real-amplitude pure states. With unit detection efficiency, public records, and bounded Hamiltonian feedback, fixed transverse noise requires a minimum peak strength proportional to $ε^{-1}$ at distributional error $ε$. Tangent noise instead admits a construction requiring only $O[\sqrt{\log(1/ε)}]$. At fixed preparation time, we determine the optimum over all allowed feedback policies and derive a geometric lower bound beyond radial symmetry. The separation identifies the measurement realization as part of the control-resource specification for monitored ensemble preparation.
AI Summary
The paper asks whether different environment measurements implementing the same discarded-record quantum channel can require different optimal feedback strengths for preparing a prescribed quantum-state ensemble. It studies continuous complete-Pauli monitoring of $n\ge2$ qubits, starting from a publicly known complex-Haar state and targeting the uniform ensemble on $\mathbb{RP}^{d-1}$, with $d=2^n$. Feedback is any predictable Hamiltonian satisfying $\|H_t\|_{\rm op}\le U$; error is trace-distance Wasserstein distance between record-conditioned state laws. Readout phases preserve the depolarizing channel while changing transverse measurement noise. At fixed dimension, preparation time $T$, and monitoring rate, every fixed transverse-noise weight $\delta>0$ requires $\Theta(\epsilon^{-1})$ peak strength. Tangent monitoring, $\delta=0$, admits $U\le\pi/(4T)+\sqrt{(2\kappa/T)\log(1/\epsilon)}$, where $\kappa=\Gamma/(2d)$; no matching tangent lower asymptotic is established. Residual transverse noise retains this budget class exactly when $\delta=O[\epsilon\sqrt{\log(1/\epsilon)}]$. The proof reduces target distance to a scalar diffusion. Pathwise comparison bounds every causal controller, while maximal inward feedback and orthogonal covariance attain the exact finite-time optimum $E^*(U,T)=\mathbb E\sin R_T$, including the correct target sampling weights. A stationary certificate scales as $\kappa\delta(d-1)/(2U)$. A terminal-window argument extends the inverse-accuracy obstruction beyond radial symmetry. The separation assumes unit efficiency, public state reconstruction, continuously active monitoring, and excludes same-increment stochastic feedback and free terminal pulses. Complete-Pauli monitoring uses exponentially many channels and unrestricted global control; structured extensions change the prior or control contract. The readouts differ in information content as well as backaction geometry. Fixed inefficiency produces a positive pure-target error floor; numerical checks do not establish uniform discrete-policy convergence.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 46 pages, 7 figures; includes Supplemental Material