arXiv:2610.22069
Generalized $m$-body mean-field projections for classical simulation of $d$-dimensional Heisenberg-picture dynamics
Federico Tomás B. Pérez, Gabriela Wójtowicz, Matthias Kost, Martin B. Plenio
Abstract
Classical simulation of quantum many-body dynamics is challenged by the rapid growth of operator complexity, particularly in higher dimensions where tensor-network methods become increasingly demanding. We develop state-adapted Pauli-weight propagation based on generalized $m$-body mean-field projections, which compress many-body observables into optimal low-body representations in a covariance geometry induced by the initial state. Combined with adaptive Krylov-like restricted Heisenberg dynamics, this provides an alternative to conventional Hilbert--Schmidt Pauli-weight truncation. We derive necessary and conditional sufficient bounds on low-body approximation accuracy using standard operator complexity measures and characterize possible dynamical failure mechanisms of the state-adapted method. Numerical benchmarks demonstrate substantial state-dependent accuracy gains, while identifying regimes of reduced accuracy or violations of Pauli spectral bounds. For a three-dimensional system of $80$ interacting spins, our method achieves strong body-order convergence, close agreement with independently converging tensor-network simulations, and substantially smaller sparse Pauli representations. These results establish state-adapted Pauli propagation as a promising approach to classical simulation of higher-dimensional coherent many-body dynamics.
AI Summary
The paper asks how to truncate Heisenberg-picture operator dynamics when the goal is an expectation value in a known initial state, rather than reconstruction of the full observable. It retains Pauli strings of weight at most $m$, but replaces Hilbert–Schmidt truncation with a projection minimizing $\mathrm{Tr}[\sigma_0(O-O_m)^2]$. For product reference states, this projection truncates the expansion in centered local operators and is explicitly computable. Projected dynamics are propagated using adaptive, finite-depth Krylov-like bases. The analysis separates instantaneous representation error from discarded-sector feedback and finite-basis errors. Projection preserves the reference-state expectation exactly, so prediction error arises through the latter dynamical contributions. Operator-space Rényi-2 entanglement supplies a necessary condition for low-body Hilbert–Schmidt approximation; sufficient covariance-error bounds require an independently assumed exponentially decaying body-order tail and controlled return coupling. Rank-deficient reference states introduce null directions whose equivalence need not survive projected evolution. Benchmarks use transverse-field Ising models with a longitudinal field. A 10-spin purity sweep finds approximately two orders of magnitude lower time-averaged error than Hilbert–Schmidt propagation near polarization $r=0.9$. An 18-spin orientation sweep also finds unfavorable states with approximate Pauli expectations reaching about $2.5$. For an 80-spin three-dimensional lattice over $tJ\leq0.6$, body orders $m=2,3$ agree at roughly the $10^{-3}$ scale and agree with independently converging tensor-network calculations. Covariance propagation uses roughly $10^4$ versus $10^5$ nonzero coefficients at $m=3$, but both require order $10^5$ seconds. No general convergence guarantee, large-system error certificate, or proportional runtime advantage is established; the adaptive threshold is not an error bound.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- Companion paper to arXiv:2609.12840; 35 pages; 6 figures; supplementary numerical results and analytical proofs included. Comments and suggestions are appreciated