Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.20482

Quantum Spectral Estimation with Multiple Initial States: Improved Error and Cost Scaling

Lingwei Zhang·Zhiyan Ding
View original source ↗Read source document ↗Export metadata ↓

Abstract

We study quantum spectrum and eigenstate property estimation using multiple initial states and generalized Hadamard tests. Building on the QFAMES (quantum filtering and analysis of multiplicities in eigenvalue spectra) framework, we introduce a new method that samples evolution times independently for each initial-state pair. Under the stated assumptions, when statistical error determines the sampling cost, the total Hamiltonian simulation cost required for fixed-accuracy eigenvalue and observable estimation is independent of the number of initial states, up to logarithmic factors. This improves scalability when many initial states are needed to explore the target eigenspaces. The method also reduces the cost of multiplicity recovery and retains Heisenberg-limited scaling for eigenvalue estimation. Numerical experiments support the predicted statistical error scalings.

Explore related subjects

quant-ph

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Lingwei Zhang, Zhiyan Ding. 2026-10-14. Quantum Spectral Estimation with Multiple Initial States: Improved Error and Cost Scaling. https://arxiv.org/abs/2610.20482

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

All you need is Trotter

The work here enables a cost scaling of $\tilde{O}\left( \frac{T ^{1+o(1)}\,{\rm log}^2 (1/ε)}{\varepsilon}\right)$ and maximum depth $\tilde{O}\left( \frac{T^{1+o(1)}\,{\rm log} (1/ε)}{\varepsilon}\right)$ for estimating observables after some Hamiltonian evolution $e^{i H T}$ using no extra block-encoding qubits, where $ε$ is the algorithmic error of the target Hamiltonian evolution and $\varepsilon$ is the statistical error from the observable estimation. Throughout, $\tilde{O}$ suppresses factors that are doubly logarithmic in $1/ε$ (i.e. ${\rm polyloglog}(1/ε)$), which are subdominant to the displayed ${\rm polylog}(1/ε)$ factors. This is achieved through product formulas, stable interpolation (Chebyshev), and to calculate the needed fractional queries, cardinal sine interpolation is used. This is an improvement over similar work using interpolation methods which have a superlinear scaling of the form $5^p T^{1+1/p}$, where $p$ is the order of the product formula used and an improvement on required qubits over Quantum Signal Processing methods while maintaining scaling with ${\rm polylog}{1/ε}$. This is accomplished through an optimized and variable (with $T$) order $p$. The slightly better $\log^2(1/ε)$ compared to $\log^3(1/ε)$ comes from the use of cardinal sine interpolation compared to other methods using Chebyshev interpolation and product formulas. The work presented here also provides an alternative to Quantum Signal Processing techniques to calculate observables or amplitudes after a fractional query evolution with a cost $O(T\log{1/ε})$ through this cardinal sine interpolation.

quant-ph↗
arXiv2026

Relativistic generalization of Feynman's path integral on the basis of extended Lagrangians

In the extended Lagrange formalism of classical point dynamics, the system's dynamics is parametrized along a system evolution parameter $s$, and the physical time $t$ is treated as a \emph{dependent} variable $t(s)$ on equal footing with all other configuration space variables $q^{i}(s)$. In the action principle, the conventional classical action $L dt$ is then replaced by the generalized action $L_{e}ds$. A direct reparametrization of the conventional action yields an extended Lagrangian that is homogeneous of degree one in the velocities. A regular extended Lagrangian that is not homogeneous of degree one can also reproduce the conventional dynamics on a suitably selected invariant hypersurface. With this representation, the physical motion takes place on a hypersurface within a symplectic extended phase space of even dimension. The formal similarity of conventional and extended Lagrange formalisms provides a direct connection with the proper-time path integral approach to relativistic quantum physics. To provide an example, the non-homogeneous extended Lagrangian $L_{e}$ of a classical relativistic spinless point particle in an external electromagnetic field will be presented. This extended Lagrangian is a quadratic function in the velocities, and its selected zero-energy hypersurface is the classical mass shell. With this $L_{e}$, the generalized path integral first yields an evolution equation in $s$. Imposing the corresponding quantum mass-shell condition then yields the Klein-Gordon equation as the physical wave equation. This establishes the relation between the classical extended Lagrangian description and proper-time quantization for a charged spin-zero particle.

quant-ph↗
arXiv2026

Reliable Entropy Estimation from Observed Statistics for Device-Independent Quantum Cryptography

This paper introduces a numerical framework for establishing lower bounds on the conditional von-Neumann entropy in device-independent quantum cryptography and randomness extraction scenarios. Leveraging a hierarchy of semidefinite programs derived from the Navascués-Pironio-Acin (NPA) hierarchy, our tool enables efficient computation of entropy bounds based solely on observed statistics, assuming the validity of quantum mechanics. The method's computational efficiency is ensured by its reliance on projective operators within the non-commutative polynomial optimization problem. The method facilitates provable bounds for extractable randomness in noisy scenarios and aligns with modern entropy accumulation theorems. Consequently, the framework offers an adaptable tool for practical quantum cryptographic protocols, expanding secure communication possibilities in untrusted environments.

quant-ph↗