arXiv:2610.03492
Efficient Sampling for Many-Body Fermionic Non-Gaussianity
Ryota Matsuda, Masahiro Hoshino, Yuto Ashida
Abstract
Fermionic non-Gaussianity is a key resource for universal quantum computation, and its behavior in quantum many-body systems has attracted growing interest. Recently, a convolution-based measure, which we call the magic Rényi entropy (MRE), has been proposed as a resource measure of non-Gaussianity, but its evaluation is limited to small systems due to the computational cost growing exponentially with system size. In this work, we develop a perfect-sampling method to calculate the second-order MRE from matrix product states (MPSs). Our method uses a bounded estimator to control sampling fluctuations and a recursive sweep to draw the samples directly from the input MPS, which allows us to generate each sample with $\mathcal O(nD^3)$ time, where $n$ and $D$ denote the number of modes and MPS bond dimension, respectively. We benchmark our method on the XXZ chain with up to $128$ sites, demonstrating that the MRE captures intricate higher-order correlations at large scales, which are inaccessible to covariance-based measures. These results provide a computational tool that quantitatively evaluates many-body fermionic non-Gaussianity by accounting for higher-order correlations.
AI Summary
The paper addresses how to evaluate convolution-based fermionic non-Gaussianity for many-body states beyond exact wave-function calculations. It focuses on the second-order magic Rényi entropy (MRE), $M_2=-\ln\nu_2$, where $\nu_2$ is the purity of one output after a balanced fermionic beam splitter acts on two identical copies. The method combines exact Born sampling from a definite-parity matrix product state (MPS) with a bounded, unbiased purity estimator. Rao–Blackwellizing the usual swap-amplitude ratio $r$ gives $h=2\operatorname{Re}r/(1+|r|^2)\in[-1,1]$. A parity-resolved recursive sweep handles fermionic signs without explicitly constructing the beam-splitter output MPS. Each estimator sample costs $\mathcal O(nD^3)$ time and $\mathcal O(D^2)$ additional memory for $n$ modes and bond dimension $D$; particle-number symmetry reduces practical costs. Benchmarks use a singly occupied spinful-fermion XXZ chain whose covariance matrix is independent of anisotropy. At $\Delta=-1$, sampling agrees with an exact Dicke-state calculation through $128$ sites, with $M_2\sim\tfrac32\ln L$. For $\Delta<-1$, DMRG calculations resolve MRE saturation associated with domain-wall formation, consistent with an analytical domain-wall reference state. Supplementary spinless-chain calculations reach $120$ sites with extensive MRE. Polynomial cost per sample does not guarantee polynomial total cost: fixed entropy error $\epsilon_M$ requires at most approximately $e^{2M_2}/\epsilon_M^2$ samples, while tested states empirically require order $e^{M_2}/\epsilon_M^2$. Thus extensive MRE retains exponential sampling cost. The $128$-site Dicke benchmark requires approximately $1.4\times10^7$ samples for error $0.01$, or $93$ CPU-hours in the reported implementation.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 21 pages, 11 figures