pith. sign in

arxiv: 2512.10643 · v2 · submitted 2025-12-11 · 🪐 quant-ph · cond-mat.str-el

Efficient simulation of low-entanglement bosonic Gaussian states in polynomial time

Tong Liu , Hui-Ke Jin , Tao Xiang , Hong-Hao Tu This is my paper

Pith reviewed 2026-05-16 23:14 UTC · model grok-4.3

classification 🪐 quant-ph cond-mat.str-el
keywords bosonic Gaussian statesmatrix product statesGaussian boson samplinghafniantensor networksquantum opticsclassical simulation
0
0 comments X

The pith

Pure bosonic Gaussian states with low entanglement convert to matrix product states in polynomial time without computing hafnians.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops an algorithm that represents pure bosonic Gaussian states as matrix product states by combining Gaussian singular value decomposition with a projected-creation-operator mapping. This constructs the local MPS tensors directly and avoids the hafnian calculations that normally bottleneck sampling in the boson occupation basis. The method targets the low-entanglement regime typical of lossy devices, where the resulting bond dimension stays tractable and enables scalable classical simulation. Benchmarks on covariance matrices from the Jiuzhang experiments show clear speedups compared with earlier tensor-network techniques. If correct, the approach extends efficient MPS tools to a wider class of bosonic Gaussian systems where direct Gaussian-formalism methods become impractical.

Core claim

By combining a Gaussian singular value decomposition with a projected-creation-operator mapping, the algorithm constructs local MPS tensors for pure bosonic Gaussian states without computing hafnians. This produces a scalable classical simulation framework in which a target accuracy is reached with a bond dimension that remains computationally tractable for states possessing limited entanglement.

What carries the argument

Gaussian singular value decomposition combined with projected-creation-operator mapping that builds local MPS tensors without hafnians

If this is right

  • Sampling from lossy Gaussian boson sampling experiments becomes faster than previous tensor-network methods.
  • A desired accuracy level is reached with a bond dimension that stays small enough for practical computation.
  • MPS-based techniques now apply directly to a broader range of bosonic systems where Gaussian formalism calculations are inefficient.
  • The method supplies a concrete classical benchmark for verifying quantum advantage claims in the low-entanglement regime.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same decomposition strategy could be tested on non-Gaussian bosonic states to see how far the polynomial scaling extends.
  • Combining this MPS representation with existing continuous-variable tensor-network tools might handle slightly higher entanglement without exponential cost.
  • The approach supplies a natural starting point for studying entanglement growth under loss or decoherence in optical networks.

Load-bearing premise

The bosonic Gaussian states are pure and possess sufficiently low entanglement that the resulting MPS bond dimension remains computationally tractable.

What would settle it

Running the algorithm on a family of pure Gaussian states whose entanglement is known to be low and observing that the required bond dimension or runtime grows exponentially with system size would falsify the claim.

Figures

Figures reproduced from arXiv: 2512.10643 by Hong-Hao Tu, Hui-Ke Jin, Tao Xiang, Tong Liu.

Figure 1
Figure 1. Figure 1: FIG. 1. (a) Schematic of GSVD on the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Performance of the PCO mapping algorithm for the [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Comparison between the hafnian-based and the PCO mapping methods. The performance is benchmarked by con [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Performance of the PCO mapping algorithm for the [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Evolution of the loss [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
read the original abstract

Bosonic Gaussian states are ubiquitous in quantum optics and condensed matter physics. While they are efficiently handled within the Gaussian formalism, sampling requires calculating amplitudes in the boson occupation basis. This step, however, is hindered by a significant bottleneck due to the hafnian. We present an efficient algorithm that converts pure bosonic Gaussian states into matrix product states (MPSs), thereby establishing a versatile tool for probing bosonic Gaussian systems in settings where direct Gaussian-formalism-based calculations become inefficient. Our method combines a Gaussian singular value decomposition with a projected-creation-operator mapping that constructs local MPS tensors without computing hafnians. Benchmarking on covariance matrices from the Jiuzhang 2.0 and Jiuzhang 4.0 Gaussian boson sampling experiments demonstrates substantial speedups over previous tensor-network approaches in the low-entanglement regime relevant to lossy devices. The method provides a scalable classical simulation framework for bosonic Gaussian states with limited entanglement. In this regime, a target accuracy can be achieved with a bond dimension that remains computationally tractable, thereby extending the applicability of MPS-based methods to a broad range of bosonic systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper claims to present an efficient algorithm converting pure bosonic Gaussian states to matrix product states (MPS) via Gaussian singular value decomposition combined with a projected-creation-operator mapping. This construction avoids hafnian evaluation and enables polynomial-time simulation and sampling in the low-entanglement regime, with reported speedups on experimental covariance matrices from the Jiuzhang 2.0 and 4.0 Gaussian boson sampling devices.

Significance. If the central claim holds, the work would supply a practical classical tool for simulating bosonic Gaussian states in quantum optics and condensed-matter settings where entanglement is limited, extending MPS techniques beyond current hafnian-limited approaches and offering scalable simulation for lossy experimental devices.

major comments (2)
  1. [Abstract] Abstract and complexity discussion: the polynomial-time claim rests on the resulting MPS bond dimension remaining computationally tractable (O(poly(N)) or small) under the low-entanglement assumption, yet no explicit bound, scaling argument, or proof is supplied showing that the projected-creation-operator mapping preserves area-law structure without exponential inflation of D for fixed entanglement entropy.
  2. [Results / Benchmarking] Benchmarking and results: speedups are asserted over prior tensor-network methods on Jiuzhang covariance matrices, but the manuscript provides neither the achieved bond dimensions, truncation errors, nor quantitative accuracy metrics, leaving the practical advantage and the regime of validity unverified.
minor comments (1)
  1. [Abstract] The abstract refers to 'sufficiently low entanglement' without a quantitative criterion (e.g., entanglement entropy threshold or scaling with mode number) that would allow readers to assess applicability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address each major comment below and indicate the revisions we will make to strengthen the presentation.

read point-by-point responses

  1. Referee: [Abstract] Abstract and complexity discussion: the polynomial-time claim rests on the resulting MPS bond dimension remaining computationally tractable (O(poly(N)) or small) under the low-entanglement assumption, yet no explicit bound, scaling argument, or proof is supplied showing that the projected-creation-operator mapping preserves area-law structure without exponential inflation of D for fixed entanglement entropy.

    Authors: We appreciate this observation. The low-entanglement regime is characterized by area-law scaling of the entanglement entropy, which in standard MPS constructions implies that a target accuracy can be reached with polynomially bounded bond dimension. Our Gaussian SVD followed by the projected-creation-operator mapping is designed to respect the local correlations of the original bosonic Gaussian state and therefore does not introduce additional long-range entanglement. Nevertheless, we agree that an explicit scaling bound or formal proof that D remains O(poly(N)) for fixed entropy is not supplied. In the revised manuscript we will add a dedicated paragraph in the complexity discussion that provides a heuristic scaling argument based on the preservation of the Gaussian covariance structure under the local mapping, together with numerical evidence from the Jiuzhang data sets confirming that D grows at most polynomially in the experimentally relevant regime. revision: yes

  2. Referee: [Results / Benchmarking] Benchmarking and results: speedups are asserted over prior tensor-network methods on Jiuzhang covariance matrices, but the manuscript provides neither the achieved bond dimensions, truncation errors, nor quantitative accuracy metrics, leaving the practical advantage and the regime of validity unverified.

    Authors: We agree that quantitative metrics are necessary to substantiate the claimed speedups and to delineate the regime of validity. The current version reports only wall-clock runtime improvements. In the revised manuscript we will augment the benchmarking section with (i) tables listing the bond dimensions employed for each Jiuzhang 2.0 and 4.0 covariance matrix, (ii) the corresponding truncation errors after SVD, and (iii) accuracy metrics such as the fidelity of sampled distributions against reference data (where exact comparison is feasible for subsystems) or the convergence of expectation values. These additions will make the practical advantage and the low-entanglement operating window explicit. revision: yes

Circularity Check

0 steps flagged

No circularity: direct algorithmic mapping from Gaussian formalism to MPS tensors

full rationale

The derivation chain consists of a Gaussian SVD step followed by a projected-creation-operator construction that explicitly builds local MPS tensors from the covariance matrix without invoking hafnians. This is a self-contained algorithmic procedure grounded in standard linear-algebra operations and tensor-network representations; no key quantity (bond dimension, runtime, or amplitude) is obtained by fitting a parameter to the target result or by renaming a prior self-citation as a uniqueness theorem. The low-entanglement regime is invoked only to bound computational cost via the standard area-law property of MPS, which is an external fact about tensor networks rather than an assumption smuggled in by the authors' own prior work. Consequently the central claim does not reduce to its inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The algorithm rests on standard linear-algebraic properties of Gaussian states and the existence of an efficient SVD for covariance matrices; no new free parameters, ad-hoc axioms, or invented entities are introduced in the abstract.

axioms (2)
  • standard math Bosonic Gaussian states admit an efficient singular-value decomposition of their covariance matrix
    Invoked to decompose the state before the projected-creation mapping
  • domain assumption Matrix product states with modest bond dimension can represent low-entanglement states accurately
    Central to the claim that the resulting MPS remains tractable

pith-pipeline@v0.9.0 · 5499 in / 1278 out tokens · 25392 ms · 2026-05-16T23:14:38.607101+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

77 extracted references · 77 canonical work pages · 2 internal anchors

  1. [1]

    We refer to this construction as the GSVD

    Gaussian Singular Value Decomposition To formulate the GSVD in a way that closely parallels the MPS construction, it is useful to discuss a Gaussian analogue of the SVD. We refer to this construction as the GSVD. Conceptually, the GSVD factorizes a pure BGS across a bipartition into three pure BGSs by introducing a set of virtual modes. We begin by rewrit...

    work page

  2. [2]

    At each step we factor out one physical mode together with its associated virtual modes, producing one local Gaussian state

    Iterative Decomposition Algorithm Having established the GSVD for a single bipartition, we now build the full MPS representation by applying the decomposition iteratively for the whole system. At each step we factor out one physical mode together with its associated virtual modes, producing one local Gaussian state. The algorithm proceeds as follows: 1.Fi...

    work page

  3. [3]

    Our first task is there- fore to identify, for a given bond dimensionD, the sub- space that retains the maximal amount of physically rel- evant information

    The Kept Subspace and Projection To convert each|A m⟩into a finite-dimensional MPS tensor,A m αpβ, we must project|A m⟩, which lives in an infinite-dimensional bosonic Hilbert space, onto atrun- catedsubspace of fixed dimension. Our first task is there- fore to identify, for a given bond dimensionD, the sub- space that retains the maximal amount of physic...

    work page

  4. [4]

    , nm β,nme ⟩, wheren m β,q denotes the oc- cupation number of the virtual moder † m,q

    Enumerate virtual Fock states|v m β ⟩= |nm β,1, . . . , nm β,nme ⟩, wheren m β,q denotes the oc- cupation number of the virtual moder † m,q

    work page

  5. [5]

    Compute their entanglement energiesE m β = −P q 2nm β,qlog(Λq)

    work page

  6. [6]

    6 The basis for the incoming virtual modesr † m−1,q can be obtained in a similar manner

    Sort states by increasingE m β and retain the firstD states. 6 The basis for the incoming virtual modesr † m−1,q can be obtained in a similar manner. The above procedure allows us to construct the opti- mal virtual bases{|v m−1 α ⟩}and{|v m β ⟩}for the incoming and outgoing virtual modes{r † m−1,q}and{l † m,q}. To- gether with the truncated physical basis...

    work page

  7. [7]

    search-and-update

    The Core Algorithm: Sequential Application of Projected Creation Operators A direct calculation of the MPS tensor entries in Eq. (26) would require calculating hafnians, a compu- tationally expensive task we circumvent with an alterna- tive algorithmic approach. Recall that the local state |Am⟩is a BGS defined on three sets of modes: the incoming virtual ...

    work page

  8. [8]

    Iterate through each of thedD 2 basis states|v k⟩ ∈ V m and its coefficientc k in the current state vector

    work page

  9. [9]

    For each non-zeroc k, apply the PCOP m ζ,ζ ′ to the basis state|v k⟩. This generates a linear combina- tion of distinct target basis states: P m ζ,ζ ′|vk⟩= X κ gζ,ζ ′ κ |vζ,ζ ′ k,κ ⟩, where|v ζ,ζ ′ k,κ ⟩is the unit basis vector inV m such that |vζ,ζ ′ k,κ ⟩ ∝(d † ζd† ζ′)κ|vk⟩,(32) andg ζ,ζ ′ κ is the corresponding expansion coefficient, whose explicit val...

    work page

  10. [10]

    7 We now analyze the computational complexity

    For each resulting target state|v ζ,ζ ′ k,κ ⟩, we identify its index in the basis ofV m and add the corresponding contribution (ck ×g ζ,ζ ′ κ ) to the resulting vector. 7 We now analyze the computational complexity. The cost of applying one PCO is dominated by the loop over thedD 2 coefficients, with each step involvingO(˜n ζ,ζ ′) operations. The cost of ...

    work page

  11. [11]

    Physical excitation through the physical modes in- volved PCOs: PVm exp[gr† m−1,qa† m] (1≤q≤n m−1 e ) PVm exp[ g 2 a† ma† m] PVm exp[gl† m,q′a† m] (1≤q ′ ≤n m e ) (B5)

    work page

  12. [12]

    Therefore, the computation of local tensor elements can be distributed across at leastdCPU/GPU cores

    Independent processing of each occupation sector using virtual-mode PCOs, enabling parallel com- putation across different physical occupation num- bers. Therefore, the computation of local tensor elements can be distributed across at leastdCPU/GPU cores. The parallelization of our algorithm can be further enhanced by incorporating some virtual modes as p...

    work page

  13. [13]

    Aaronson and A

    S. Aaronson and A. Arkhipov, inProceedings of the forty- third annual ACM symposium on Theory of computing (2011) pp. 333–342

    work page 2011

  14. [14]

    M. A. Broome, A. Fedrizzi, S. Rahimi-Keshari, J. Dove, S. Aaronson, T. C. Ralph, and A. G. White, Science 339, 794 (2013)

    work page 2013

  15. [15]

    Tillmann, B

    M. Tillmann, B. Daki´ c, R. Heilmann, S. Nolte, A. Sza- meit, and P. Walther, Nat. Photon.7, 540 (2013)

    work page 2013

  16. [16]

    J. B. Spring, B. J. Metcalf, P. C. Humphreys, W. S. Kolthammer, X.-M. Jin, M. Barbieri, A. Datta, N. Thomas-Peter, N. K. Langford, D. Kundys,et al., Science339, 798 (2013)

    work page 2013

  17. [17]

    Crespi, R

    A. Crespi, R. Osellame, R. Ramponi, D. J. Brod, E. F. Galvao, N. Spagnolo, C. Vitelli, E. Maiorino, P. Mat- aloni, and F. Sciarrino, Nat. Photon.7, 545 (2013)

    work page 2013

  18. [18]

    H. Wang, Y. He, Y.-H. Li, Z.-E. Su, B. Li, H.-L. Huang, X. Ding, M.-C. Chen, C. Liu, J. Qin, J.-P. Li, Y.-M. He, C. Schneider, M. Kamp, C.-Z. Peng, S. H¨ ofling, C.-Y. Lu, and J.-W. Pan, Nat. Photon.11, 361 (2017)

    work page 2017

  19. [19]

    Y. He, X. Ding, Z.-E. Su, H.-L. Huang, J. Qin, C. Wang, S. Unsleber, C. Chen, H. Wang, Y.-M. He, X.-L. Wang, W.-J. Zhang, S.-J. Chen, C. Schneider, M. Kamp, L.-X. You, Z. Wang, S. H¨ ofling, C.-Y. Lu, and J.-W. Pan, Phys. Rev. Lett.118, 190501 (2017)

    work page 2017

  20. [20]

    J. C. Loredo, M. A. Broome, P. Hilaire, O. Gazzano, I. Sagnes, A. Lemaitre, M. P. Almeida, P. Senellart, and A. G. White, Phys. Rev. Lett.118, 130503 (2017)

    work page 2017

  21. [21]

    A. P. Lund, A. Laing, S. Rahimi-Keshari, T. Rudolph, J. L. O’Brien, and T. C. Ralph, Phys. Rev. Lett.113, 100502 (2014)

    work page 2014

  22. [22]

    Zhong, L.-C

    H.-S. Zhong, L.-C. Peng, Y. Li, Y. Hu, W. Li, J. Qin, D. Wu, W. Zhang, H. Li, L. Zhang, Z. Wang, L. You, X. Jiang, L. Li, N.-L. Liu, J. P. Dowling, C.-Y. Lu, and J.-W. Pan, Sci. Bull.64, 511 (2019)

    work page 2019

  23. [23]

    Zhong, H

    H.-S. Zhong, H. Wang, Y.-H. Deng, M.-C. Chen, L.-C. Peng, Y.-H. Luo, J. Qin, D. Wu, X. Ding, Y. Hu, P. Hu, X.-Y. Yang, W.-J. Zhang, H. Li, Y. Li, X. Jiang, L. Gan, G. Yang, L. You, Z. Wang, L. Li, N.-L. Liu, C.-Y. Lu, and J.-W. Pan, Science370, 1460 (2020)

    work page 2020

  24. [24]

    Zhong, Y.-H

    H.-S. Zhong, Y.-H. Deng, J. Qin, H. Wang, M.-C. Chen, L.-C. Peng, Y.-H. Luo, D. Wu, S.-Q. Gong, H. Su, Y. Hu, P. Hu, X.-Y. Yang, W.-J. Zhang, H. Li, Y. Li, X. Jiang, L. Gan, G. Yang, L. You, Z. Wang, L. Li, N.-L. Liu, J. J. Renema, C.-Y. Lu, and J.-W. Pan, Phys. Rev. Lett.127, 180502 (2021)

    work page 2021

  25. [25]

    L. S. Madsen, F. Laudenbach, M. F. Askarani, F. Rortais, T. Vincent, J. F. Bulmer, F. M. Miatto, L. Neuhaus, L. G. Helt, M. J. Collins,et al., Nature606, 75 (2022)

    work page 2022

  26. [26]

    Deng, Y.-C

    Y.-H. Deng, Y.-C. Gu, H.-L. Liu, S.-Q. Gong, H. Su, Z.- J. Zhang, H.-Y. Tang, M.-H. Jia, J.-M. Xu, M.-C. Chen, J. Qin, L.-C. Peng, J. Yan, Y. Hu, J. Huang, H. Li, Y. Li, Y. Chen, X. Jiang, L. Gan, G. Yang, L. You, L. Li, H.-S. Zhong, H. Wang, N.-L. Liu, J. J. Renema, C.-Y. Lu, and J.-W. Pan, Phys. Rev. Lett.131, 150601 (2023)

    work page 2023

  27. [27]

    C. S. Hamilton, R. Kruse, L. Sansoni, S. Barkhofen, C. Silberhorn, and I. Jex, Phys. Rev. Lett.119, 170501 (2017)

    work page 2017

  28. [28]

    Valiant, Theor

    L. Valiant, Theor. Comput. Sci.8, 189 (1979)

    work page 1979

  29. [29]

    Gaussian Noise Sensitivity and BosonSampling

    G. Kalai and G. Kindler, arXiv:1409.3093 (2014)

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  30. [30]

    Rahimi-Keshari, A

    S. Rahimi-Keshari, A. P. Lund, and T. C. Ralph, Phys. Rev. Lett.114, 060501 (2015)

    work page 2015

  31. [31]

    Oszmaniec and D

    M. Oszmaniec and D. J. Brod, New J. Phys.20, 092002 (2018)

    work page 2018

  32. [32]

    H. Qi, D. J. Brod, N. Quesada, and R. Garc´ ıa-Patr´ on, Phys. Rev. Lett.124, 100502 (2020)

    work page 2020

  33. [33]

    Aharonov, X

    D. Aharonov, X. Gao, Z. Landau, Y. Liu, and U. Vazi- rani, inProceedings of the 55th Annual ACM Symposium on Theory of Computing(2023) pp. 945–957

    work page 2023

  34. [34]

    C. Oh, L. Jiang, and B. Fefferman, arXiv:2301.11532 (2023)

    work page arXiv 2023

  35. [35]

    C. Oh, M. Liu, Y. Alexeev, B. Fefferman, and L. Jiang, Nat. Phys.20, 1461 (2024)

    work page 2024

  36. [36]

    Auerbach,Interacting electrons and quantum mag- netism(Springer New York, 2012)

    A. Auerbach,Interacting electrons and quantum mag- netism(Springer New York, 2012)

    work page 2012

  37. [37]

    D. P. Arovas and A. Auerbach, Phys. Rev. B38, 316 (1988)

    work page 1988

  38. [38]

    Sachdev, Phys

    S. Sachdev, Phys. Rev. B45, 12377 (1992)

    work page 1992

  39. [39]

    Wang and A

    F. Wang and A. Vishwanath, Phys. Rev. B74, 174423 (2006)

    work page 2006

  40. [40]

    Keeling, E

    J. Keeling, E. M. Stoudenmire, M.-C. Ba˜ nuls, and D. R. Reichman, arXiv:2509.07661 (2025)

    work page arXiv 2025

  41. [41]

    Link, H.-H

    V. Link, H.-H. Tu, and W. T. Strunz, Phys. Rev. Lett. 132, 200403 (2024)

    work page 2024

  42. [42]

    Chen and R.-B

    C. Chen and R.-B. Liu, arXiv:2509.00424 (2025)

    work page arXiv 2025

  43. [43]

    Chanda, J

    T. Chanda, J. Zakrzewski, M. Lewenstein, and L. Tagli- acozzo, Phys. Rev. Lett.124, 180602 (2020)

    work page 2020

  44. [44]

    Jerrum, A

    M. Jerrum, A. Sinclair, and E. Vigoda, J. ACM51, 671 (2004)

    work page 2004

  45. [45]

    Generalizing and Derandomizing Gurvits's Approximation Algorithm for the Permanent

    S. Aaronson and T. Hance, arXiv:1212.0025 (2012)

    work page internal anchor Pith review Pith/arXiv arXiv 2012

  46. [46]

    Lim and C

    Y. Lim and C. Oh, arXiv:2502.12882 (2025)

    work page arXiv 2025

  47. [47]

    Quesada, R

    N. Quesada, R. S. Chadwick, B. A. Bell, J. M. Arrazola, T. Vincent, H. Qi, and R. Garc´ ıa-Patr´ on, PRX Quantum 3, 010306 (2022)

    work page 2022

  48. [48]

    J. F. F. Bulmer, B. A. Bell, R. S. Chadwick, A. E. Jones, D. Moise, A. Rigazzi, J. Thorbecke, U.-U. Haus, T. V. Vaerenbergh, R. B. Patel, I. A. Walmsley, and A. Laing, Sci. Adv.8, eabl9236 (2022)

    work page 2022

  49. [49]

    T. Dodd, J. Mart´ ınez-Cifuentes, O. T. Brown, N. Que- sada, and R. Garc´ ıa-Patr´ on, arXiv:2511.14923 (2025)

    work page arXiv 2025

  50. [50]

    ¨Ostlund and S

    S. ¨Ostlund and S. Rommer, Phys. Rev. Lett.75, 3537 (1995)

    work page 1995

  51. [51]

    Vidal, Phys

    G. Vidal, Phys. Rev. Lett.91, 147902 (2003)

    work page 2003

  52. [52]

    Perez-Garcia, F

    D. Perez-Garcia, F. Verstraete, M. M. Wolf, and J. I. Cirac, Quantum Inf. and Comp.7, 401 (2007)

    work page 2007

  53. [53]

    Verstraete, V

    F. Verstraete, V. Murg, and J. I. Cirac, Adv. Phys.57, 143 (2008)

    work page 2008

  54. [54]

    Schollw¨ ock, Ann

    U. Schollw¨ ock, Ann. Phys.326, 96 (2011)

    work page 2011

  55. [55]

    J. I. Cirac, D. P´ erez-Garc´ ıa, N. Schuch, and F. Ver- straete, Rev. Mod. Phys.93, 045003 (2021)

    work page 2021

  56. [56]

    Xiang,Density Matrix and Tensor Network Renor- malization(Cambridge University Press, 2023)

    T. Xiang,Density Matrix and Tensor Network Renor- malization(Cambridge University Press, 2023)

    work page 2023

  57. [57]

    M. T. Fishman and S. R. White, Phys. Rev. B92, 075132 (2015)

    work page 2015

  58. [58]

    Schuch and B

    N. Schuch and B. Bauer, Phys. Rev. B100, 245121 16 (2019)

    work page 2019

  59. [59]

    Y.-H. Wu, L. Wang, and H.-H. Tu, Phys. Rev. Lett.124, 246401 (2020)

    work page 2020

  60. [60]

    Jin, H.-H

    H.-K. Jin, H.-H. Tu, and Y. Zhou, Phys. Rev. B101, 165135 (2020)

    work page 2020

  61. [61]

    A. M. Aghaei, B. Bauer, K. Shtengel, and R. V. Mish- mash, arXiv:2009.12435 (2020)

    work page arXiv 2009

  62. [62]

    Petrica, B.-X

    G. Petrica, B.-X. Zheng, G. K.-L. Chan, and B. K. Clark, Phys. Rev. B103, 125161 (2021)

    work page 2021

  63. [63]

    N¨ ußeler, I

    A. N¨ ußeler, I. Dhand, S. F. Huelga, and M. B. Plenio, Phys. Rev. A104, 012415 (2021)

    work page 2021

  64. [64]

    Jin, R.-Y

    H.-K. Jin, R.-Y. Sun, Y. Zhou, and H.-H. Tu, Phys. Rev. B105, L081101 (2022)

    work page 2022

  65. [65]

    Liu, Y.-H

    T. Liu, Y.-H. Wu, H.-H. Tu, and T. Xiang, Quantum Sci. Technol.10, 035033 (2025)

    work page 2025

  66. [66]

    Li, Y.-B

    K. Li, Y.-B. Zhang, and H. C. Po, Phys. Rev. Res.7, 013182 (2025)

    work page 2025

  67. [67]

    Jin, R.-Y

    H.-K. Jin, R.-Y. Sun, H.-H. Tu, and Y. Zhou, AAPPS Bull.35, 16 (2025)

    work page 2025

  68. [68]

    S. H. Hille and A. Szab´ o, arXiv:2510.05227 (2025)

    work page arXiv 2025

  69. [69]

    Local displacements are omitted because they can always be removed locally and do not affect entanglement prop- erties

    work page

  70. [70]

    isometries

    Strictly speaking, the GSVD of a pure BGS takes the form|ϕ⟩=⟨I m|(|ϕ L⟩ ⊗ |ϕR⟩), where |ϕL⟩= Qne q=1 exp h b† L,ql† q i |0⟩bL,l and|ϕ R⟩= Qne q=1 exp h b† R,qr† q i |0⟩bR,r are Gaussian “isometries”, and the bond state|I m⟩= Qne q=1 exp Λql† qr† q |0⟩l,r car- ries the Gaussian “singular values” Λ q. For convenience of the ensuing algorithm, we absorb the ...

    work page

  71. [71]

    Schuch, J

    N. Schuch, J. I. Cirac, and M. M. Wolf, inQuantum Information and Many Body Quantum Systems, Vol. 5, edited by M. Ericsson and S. Montangero (Edizioni della Normale, Pisa, 2008) pp. 129–142, arXiv:quant- ph/0509166

    work page arXiv 2008

  72. [72]

    Apple MacBook Air (M4, 2025)

    work page 2025

  73. [73]

    H.-L. Liu, H. Su, S.-Q. Gong, Y.-C. Gu, H.-Y. Tang, M.- H. Jia, Q. Wei, Y. Song, D. Wang, M. Zheng, F. Chen, L. Li, S. Ren, X. Zhu, M. Wang, Y. Chen, Y. Liu, L. Song, P. Yang, J. Chen, H. An, L. Zhang, L. Gan, G. Yang, J.-M. Xu, Y.-M. He, H. Wang, H.-S. Zhong, M.- C. Chen, X. Jiang, L. Li, N.-L. Liu, Y.-H. Deng, X.-L. Su, Q. Zhang, C.-Y. Lu, and J.-W. Pan...

    work page arXiv 2025

  74. [74]

    sss441803/bosonsupremacy: Version 0.0,

    M. Liu, “sss441803/bosonsupremacy: Version 0.0,” (2023)

    work page 2023

  75. [75]

    K. Liu, F. Xiong, and F. Wang, arXiv:2509.26607 (2025)

    work page arXiv 2025

  76. [76]

    Liao, J.-G

    H.-J. Liao, J.-G. Liu, L. Wang, and T. Xiang, Phys. Rev. X9, 031041 (2019)

    work page 2019

  77. [77]

    Tay and O

    T. Tay and O. I. Motrunich, Phys. Rev. B84, 020404 (2011)

    work page 2011