arXiv:2610.01125
Multivariate Quantum Signal Processing
Guang Hao Low
Abstract
We develop Multivariate Quantum Signal Processing for designing matrix-valued transformations of multiple noncommuting nonnormal block-encoded matrices. Our framework equips quantum transducers with the analytic structure of multiport Infinite-Impulse-Response (IIR) filters and establishes a transfer-function methodology for designing quantum query algorithms. Efficiently computable analytic bounds that hold uniformly across the oracle promise yield efficient quantum circuits with logarithmic error dependence and near-unity success probability, paralleling the approximation-theoretic guarantees of univariate QSP/QSVT. We provide a complete characterization of achievable transfer functions and a principled library of analytic IIR components. Our modular designs greatly improve or resolve the per-oracle query complexity of open multi-oracle problems: (1) Hamiltonian simulation under $H=\sum_j H_j$ with weighted quasinorm cost $\approx t\langle\vecλ,\vec C\rangle_{1/2}+\mathcal{O}(\|\vec C\|_1\log(1/ε))$ for block-encoding normalizations $\vecλ$ and costs $\vec C$; (2) sparse simulation with optimal time--norm $Θ(\sqrt{d}\|H\|_{\rm 1\to2}t)$ and additive error dependence; (3) solving linear systems and (4) differential equations under minimal matrix assumptions with optimal weighted state-preparation and multi-term cost; (5) ground-state preparation and energy estimation; (6) generalized eigenvalue problems. We apply these to exactly simulate downfolded Hamiltonians with leading cost $t\langle\vecλ_{\rm eff},\vec C_{\rm eff}\rangle_{1/2}$ further reduced by fractional dynamic correlation energy contributions from coarse-grained Hamiltonians, and show similar classical-data-input-reduction results for multiscale graph classification via Kron reduction. We also show the online generalization with anisotropic Kalman filtering of streaming quantum data.
AI Summary
The paper asks how to design quantum query algorithms for several noncommuting, potentially nonnormal block-encoded matrices while controlling individual oracle costs, approximation error, and success probability. It develops multivariate quantum signal processing using analytic, matrix-valued infinite-impulse-response transfer functions: feedback networks describe the desired transformation, and analytic variables track oracle queries. The construction extends quantum transducers with a unitary Toeplitz lift and input/output clock states that compile feedback into finite circuits. Clock normalization is characterized by a nuclear-norm optimization; analytic continuation and growth bounds provide constructive clocks with logarithmic error dependence. Schur–Agler realization conditions characterize achievable transfer functions, including every finite causally ordered query circuit. Lossless composition of primitive modules supplies structured implementations without explicitly expanding ordered query paths. For $H=\sum_jH_j$, with block-encoding normalizations $\lambda_j$ and query costs $C_j$, the simulation bound is $(1+\delta)t(\sum_j\sqrt{C_j\lambda_j})^2+O(\delta^{-2}\sum_jC_j\log(1/\epsilon))$. Sparse simulation attains optimal leading time–norm scaling $\Theta(t\sqrt d\|H\|_{1\to2})$, with additional precision-dependent terms. Linear-system solving simultaneously achieves $O(\kappa\log(1/\epsilon))$ matrix queries and $O(1/\sqrt p)$ state-preparation queries. Further constructions address differential equations, ground-state problems, generalized eigenvalues, downfolding, graph reduction, and online filtering. The general characterization is not itself an efficient synthesis algorithm: uniform operator constraints can be exponentially large. Efficient constructions require structured modules and provable analytic bounds. Differential-equation bounds do not supersede every specialized solver. Online Kalman filtering additionally requires succinct covariance structure, and high-success replication incurs costs quadratic in the preparation-to-output norm ratio.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 305 pages, 34 figues, 14 tables