Efficient sampling from shallow Gaussian quantum-optical circuits with local interactions

Diego Cifuentes; Haoyu Qi; Kamil Br\'adler; Nicol\'as Quesada; Robert Israel; Timjan Kalajdzievski

REVIEW 3 major objections 6 minor 60 references

Efficient sampling from shallow Gaussian quantum-optical circuits with local interactions

T0 review · 3 major / 6 minor · reviewed 2026-08-27 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A classical computer can efficiently sample from Gaussian states made by shallow local optical circuits.

desk verdict A repairable proof gap in Lemma 4, but the central simulation result for shallow local Gaussian Boson Threshold Sampling is sound and worth a serious referee. read the letter →

arxiv 2009.11824 v1 pith:GF46BEHV submitted 2020-09-24 quant-ph

Haoyu Qi , Diego Cifuentes , Kamil Brádler , Robert Israel , Timjan Kalajdzievski , Nicolás Quesada This is my paper · ORCID

classification quant-ph MSC 81P6815A15

keywords GaussianBosonSamplingshallowcircuitslocalgatesclassicalsimulabilityloophafnianbandedmatricesthresholdcontinuous-variablequantumcomputing

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proves that Gaussian Boson Sampling becomes classically simulable when the optical circuit is shallow, meaning depth logarithmic in the number of modes, and built from gates that only couple neighboring modes. Because any real photonic experiment has detectors with finite resolution, the paper defines Gaussian Boson Threshold Sampling, in which an outcome above a fixed threshold is lumped into a single '#' event. For this coarse-grained task, the runtime is polynomial in the number of modes when the threshold is constant. This removes a class of near-term photonic devices as candidates for quantum computational advantage, since realistic local circuits would be shallow to avoid loss and would therefore be simulable. The key structural fact is that the Gaussian-state adjacency matrices involved have small bandwidth proportional to circuit depth.

What carries the argument

The central object is the bandwidth of the adjacency matrix. A matrix is banded with bandwidth $w$ when its entries vanish beyond $w$ positions from the diagonal. One layer of local gates multiplies bandwidth by a constant, D layers give bandwidth $D$, and after tracing to a reduced state and coarse-graining with threshold $c$, the relevant loop-hafnian matrix is banded with bandwidth at most $8Dc$. The paper's new algorithm computes a loop hafnian of an $n \times n$ banded matrix in $O(nw 4^w)$ operations by dynamic programming over subsets of a sliding window of size $2w$, and a second algorithm handles repeated rows and columns in $O^*(n (2c+2)^{2w+1})$ operations using convolutions over weight vectors. The loop hafnian, the sum over perfect matchings with loops of products of matrix entries, is the quantity that gives Gaussian Boson Sampling detection probabilities.

What would settle it

Compute the adjacency matrix of a reduced $k$-mode state for a $D$-layer local circuit with uniform loss and check whether any off-diagonal entry with mode distance greater than $8Dc$ is nonzero; one nonzero entry violates the banded-lemma at the core of the proof. Alternatively, run Algorithm 1 on a moderate number of modes with $D = \lceil \log M \rceil$ and $c=1$, and compare its output distribution to exact probabilities from Eq. (11); any statistically significant mismatch would falsify the correctness claim.

Watch

Extended reading notes

Core claim

The central claim is Theorem 9: an M-mode, uniformly lossy Gaussian Boson Threshold Sampling problem whose unitary consists of D layers of commuting local gates can be simulated by a classical computer in time $T = O^*(M^2 \operatorname{poly}(c) (2c+2)^{16D})$; hence when $D = O(\log M)$ and $c$ is constant, the simulation is polynomial in the number of modes. The proof shows that each conditional probability in a sequential sampling algorithm is a loop hafnian of a banded matrix, with bandwidth at most $8Dc$, and supplies dynamic-programming algorithms that evaluate loop hafnians of banded matrices in time exponential in the bandwidth rather than the matrix size. The result holds for the coarse-grained threshold distribution, which the paper argues matches the realistic finite resolution of photon-number detectors, not for the exact unbounded photon-number distribution.

Load-bearing premise

The theorem's polynomial runtime collapses if the detector threshold $c$ is not a fixed constant, because the $(2c+2)^{16D}$ factor then ceases to be polynomial-scale; exact unbounded sampling over all photon-number patterns is outside the theorem.

Editorial extensions

If this is right

For depth $D = O(\log M)$ and constant threshold $c$, Gaussian Boson Threshold Sampling from a uniformly lossy local circuit is a polynomial-time classical computation, so such devices do not offer a quantum sampling advantage in the standard sense.
The runtime bound grows exponentially in circuit depth, so the polynomial-speed result is tied to logarithmic depth and cannot be stretched to arbitrary local circuits by this argument.
The threshold coarse-graining is essential: the proof does not give a finite-time algorithm for exact sampling over all $\mathbb{N}^M$ photon patterns, only for the distribution with overloading outcomes merged into '#'.
Because realistic photon loss grows exponentially with depth, deep lossy circuits were already known to be asymptotically simulable; together with this result, both shallow and deep local photonic circuits sit in classically simulable regimes.
A similar banded-structure argument should carry over to standard Boson Sampling via banded permanent algorithms, though the authors note this requires matching a fast conditional-sampling scheme to the banded-permanent subroutine.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

An exact-photon-number version of the sampling problem is left untouched; if finite detector resolution is not accepted as part of the computational task, the theorem does not constrain the hardness of sampling from the full unbounded distribution.
The banded loop-hafnian algorithm is a general matrix gadget: any Gaussian or bosonic sampling instance whose covariance or adjacency matrix has small bandwidth inherits a speedup, independent of the optical implementation.
A natural testable extension is to map the simulable region in the depth-threshold-loss plane: for slowly growing $c$ or mildly superlogarithmic $D$, the bound predicts subexponential rather than polynomial algorithms, and numerical experiments could check whether the crossover is as sharp as the theorem suggests.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Request a human review

A listed scientist reviews the paper for a fee and the review publishes here regardless of verdict. See the reviewers or get listed.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This paper gives the first classical polynomial-time simulation for shallow local Gaussian Boson Sampling, and the main result survives scrutiny. The key structural claim—that reduced adjacency matrices of shallow local circuits are banded—is believable and the proof can be repaired. The paper also contributes two reusable algorithms: a dynamic program for loop hafnians of banded matrices (Theorem 6) and a repetition-aware variant (Theorem 8). These are clearly explained and are of independent value. The GBTS model is a sensible way to make the sampling task well-defined, and the coarse-graining is defended honestly via finite-resolution detectors. Prior work, including incomplete attempts to use banded matrices, is credited fairly.

The soft spots are real but not fatal. Lemma 4's proof contains an incorrect identity: conjugating the projection W_k†W_k by a depth-D banded unitary gives an identity block only on the first k-D modes, not on all first k modes, and the boundary block is 2D wide, not something that starts after position k. The stress-test note is right about that. The claim that both blocks of V†(W†W)V equal the same Y also needs justification. However, after correcting the identity to I_{k-D} ⊕ K ⊕ 0_{M-k-D}, the inverse still has bandwidth 2D, and after conjugation by V the bandwidth stays bounded by 4D. So the bandedness conclusion survives; the lemma needs a rewritten proof, not a new idea.

Two smaller issues. The abstract says "sampling from the photon-number probability distribution" without the threshold coarse-graining; Theorem 9 is for GBTS with constant c, not the exact unbounded distribution. The authors should qualify that in the abstract. Also, Lemma 11's key bijection is just asserted ("It can be shown"), and Theorem 9's runtime doesn't explicitly budget the cost of forming the reduced matrices at each step. Both are minor; the latter is absorbed by the polynomial factors.

Bottom line: the paper is worth a serious referee. It goes beyond the qubit shallow-circuit results, the algorithms are formal and checkable, and the central claim looks correct. Send it to review with the expectation that Lemma 4 be rewritten and the abstract be made precise.

Referee Report

3 major / 6 minor

Summary. The paper claims that a classical computer can efficiently sample from a coarse-grained photon-number distribution (Gaussian Boson Threshold Sampling, GBTS) of Gaussian states produced by shallow, local optical circuits. The main result, Theorem 9, states that for M modes, D layers of commuting local gates, and detector resolution threshold c, sampling can be simulated in time O*(M^2 poly(c) (2c+2)^{16D}), which is polynomial when D = O(log M) and c is constant. The proof strategy is: (i) show that the reduced adjacency matrices of the Gaussian state are block-banded with bandwidth O(D); (ii) develop fast algorithms for loop hafnians of banded matrices (Theorem 6) and of banded matrices with repeated rows/columns (Theorem 8); (iii) combine these with a sequential conditional-sampling routine (Algorithm 1) so that each output sample requires only polynomially many probability evaluations. The paper also discusses implications for photonic quantum advantage, arguing that shallow local circuits cannot demonstrate supremacy.

Significance. If the technical gaps are repaired, this is a valuable result that extends the classical-simulability of shallow local circuits from qubit systems to continuous-variable photonics. The banded loop-hafnian algorithm of Theorem 6 is a clean and novel algorithmic contribution, likely of independent interest in combinatorial computing. The paper is self-contained, uses standard tools (Woodbury identity, banded permanent algorithms), and makes an honest effort to define a well-posed sampling task (GBTS) that reflects finite detector resolution. The main limitation is that the result concerns the threshold-coarse-grained distribution, not the exact unbounded photon-number distribution, and two key proofs currently contain gaps. Credit is due for clearly identifying the bandwidth of reduced adjacency matrices as the central structural property and for providing a fully worked dynamic program for banded loop hafnians.

major comments (3)

[Sec. IV, Lemma 4, Eqs. (38)-(40)] The proof of Lemma 4 misstates the boundary structure of V†(W_k†W_k)V. For a banded unitary U of bandwidth D, conjugating the projection P_k = I_k ⊕ 0 by U gives (U^T P_k U*)_{ij} = δ_{ij} only for i,j ≤ k-D, and 0 for i,j > k+D; the nontrivial entries occupy a 2D×2D block on indices k-D+1,...,k+D. The claimed identity Y = I_k ⊕ K ⊕ 0_{M-2D-k} in Eq. (39) should therefore be Y = I_{k-D} ⊕ K ⊕ 0_{M-k-D}. The subsequent assertion that G and F in Eq. (40) are diagonal except for a 2D×2D block also needs justification: it follows from the Woodbury identity because T^{-1} is block diagonal in the mode basis and the correction is supported on the boundary subspace, but the text does not provide this argument. Since Lemma 4 is the foundation for the bandwidth bound used in Theorem 9, the proof must be corrected and completed.
[Appendix C, Lemma 11; Theorem 8] The proof of Lemma 11 asserts 'It can be shown that this function is a bijection' without proof, but this bijection is the core of the correctness of Algorithm 3 and hence of Theorem 8. Because Theorem 9's stated runtime uses Theorem 8, the runtime claim is not fully established. The gap is likely repairable, and the qualitative polynomial-time claim for constant c can be obtained from Theorem 6 alone (with a larger exponent), but as written the proof of the stated bound is incomplete.
[Abstract and Sec. VII] The abstract and the conclusion state that a classical computer can efficiently sample from the photon-number probability distribution of a Gaussian state prepared by a shallow local circuit. The theorem actually proved (Theorem 9 with Definition 1) is limited to the coarse-grained Gaussian Boson Threshold Sampling distribution with a fixed resolution threshold c; the exact unbounded photon-number distribution over N^M is not sampled. The finite-resolution justification in Sec. II is reasonable, but the abstract and conclusion should be reworded to state the GBTS qualification, since the current wording overstates the scope of the result.

minor comments (6)

[Sec. IV, Lemma 5] The interleaving permutation in Eq. (41) maps an entry with block distance w to position distance 2w+1, so the resulting bandwidth is 2w+1, not 2w. This does not affect the main result but should be corrected.
[Appendix A, Eq. (A2)] The product in Eq. (A2) is written as ∏_{l=1}^k p(n_k | n_1...n_{k-1}); the subscript should be p(n_l | n_1...n_{l-1}).
[Sec. II, Definition 1, Eq. (10)] The index j in Σ_j = {s : ∃j s_j > c} is used both as the set index and as the mode index; please disambiguate the notation.
[Sec. V, Lemma 7 proof, Eq. (52)] In Eq. (52), 'Aij' should be 'A_it'.
[Sec. IV, after Lemma 5] The text says the extended banded matrix has bandwidth at most 8Dc; this is correct for the matrix As, but the parameter w in Theorem 8 refers to the bandwidth of the unrepeated matrix A. Please clarify this distinction to avoid confusion.
[Sec. IV, Lemma 4] The lemma assumes uniform loss, but the proof only needs the input covariance T to be diagonal in the mode basis. The statement could be generalized, or the proof should state this assumption explicitly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main theorem is derived from a self-contained bandwidth argument plus cited technical algorithms that do not presuppose the target result.

full rationale

The derivation chain is not circular. The central claim (Theorem 9) rests on Lemma 3 (bandwidth of the unitary, proved directly from the layer structure), Lemma 4 (bandedness of the reduced adjacency matrix, proved via the covariance matrix and the Sherman-Morrison-Woodbury identity), Lemma 5, Theorem 6, and Theorem 8 (banded loop hafnian algorithm, proved in Appendix C). The GBTS distribution in Definition 1 is a new sampling task rather than a renamed version of the output distribution, and no parameter is fitted to data and then renamed as a prediction. Citations to Refs. [26], [37], [39], [42], and [43] supply standard GBS probability formulas, the sequential sampling strategy, and the banded permanent precursor; these are external technical tools, and the present paper proves its own correctness for Algorithm 1 in Appendix A and proves Theorem 8 in Appendix C. Even though some cited works share authors with this paper, the proof does not invoke those citations for the load-bearing bandedness claim; that claim is argued from the unitary structure. A possible flaw in the proof of Lemma 4 would be a correctness issue rather than circularity, because the lemma's conclusion is not assumed as an input to the argument.

Assumptions & free parameters 1 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard Gaussian-state probability formulas and a new dynamic-programming algorithm. The only new formal object is the GBTS sampling task, which is a definition rather than an invented physical entity. No free parameters are fitted to data; c is a fixed model parameter.

free parameters (1)

detector resolution threshold c = constant (not fitted)
Defines the coarse-graining in GBTS (Definition 1). Runtime is exponential in c, so polynomiality requires c to be a fixed constant independent of M. It is a model parameter, not fitted to data.

assumptions (6)

standard math Photon-number probability formula for Gaussian states (Eqs. 3-7)
Taken from Ref. [26]; used throughout to compute conditional probabilities.
standard math Sherman-Morrison-Woodbury identity (Eq. 30)
Used in Lemma 4 to derive bandedness of the inverse reduced covariance matrix.
standard math Bandwidth additivity under matrix multiplication and the loop-hafnian perfect-matching interpretation
Used in Lemma 3 and Theorem 6 for the banded matrix algorithms.
domain assumption Input state is a product of single-mode Gaussian states with moments n_i, m_i and uniform loss (Eqs. 17-18)
The bandedness proof in Lemma 4 depends on this parametrization; it excludes correlated input modes and nonuniform loss.
domain assumption Detectors have finite resolution c and overload events are coarse-grained into '#' (Definition 1)
This redefines the sampling task as GBTS; the exact unbounded GBS distribution is not sampled. The polynomial runtime requires c constant.
domain assumption The circuit is a product of D layers of commuting local gates, each layer having bandwidth 1 (Lemma 3)
Defines 'shallow and local'; if gates were nonlocal or layers did not have bandwidth 1, the bandwidth argument would fail.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Efficient sampling from shallow Gaussian quantum-optical circuits with local interactions." pith.science (2026). https://pith.science/paper/GF46BEHV

@misc{pith2026200911824,
  author       = {Pith},
  title        = {Pith review of: Efficient sampling from shallow Gaussian quantum-optical circuits with local interactions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GF46BEHV}},
  note         = {Machine review of arXiv:2009.11824}
}

read the original abstract

We prove that a classical computer can efficiently sample from the photon-number probability distribution of a Gaussian state prepared by using an optical circuit that is shallow and local. Our work generalizes previous known results for qubits to the continuous-variable domain. The key to our proof is the observation that the adjacency matrices characterizing the Gaussian states generated by shallow and local circuits have small bandwidth. To exploit this structure, we devise fast algorithms to calculate loop hafnians of banded matrices. Since sampling from deep optical circuits with exponential-scaling photon loss is classically simulable, our results pose a challenge to the feasibility of demonstrating quantum supremacy on photonic platforms with local interactions.

Figures

Figures reproduced from arXiv: 2009.11824 by the authors.

**Figure 2.** FIG. 2: A two-mode example of the distribution of GBTS. We [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** FIG. 3: Block diagram of our algorithm for 3 modes. We explicitly [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 53 canonical work pages

[1]

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

work page 2017
[2]

⨁M i=1 mi⨁M i=1 m∗ i ⨁M i=1 (ni + 1 2) ) . (21) We can write the adjacency matrix of the M modes as [3, 5 48] A =X(I2M−Q−1) = (B C CT B∗ ) , (22) B =U ( M⨁ i=1 λi ) UT =BT, (23) C =U ( M⨁ i=1 µi ) U† =C†, (24) λi = mi (1 +ni)2−|mi|2, (25) µi = 1− 1 +ni (1 +ni)2−|mi|2. (26) For pure states, one ﬁnds thatµi = 0 for alli. More generally, it holds thatB andC ...

work page
[3]

Kruse, C

R. Kruse, C. S. Hamilton, L. Sansoni, S. Barkhofen, C. Silber- horn, and I. Jex, Phys. Rev. A 100, 032326 (2019)

work page 2019
[4]

Rahimi-Keshari, A

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

work page 2015
[5]

A. P. Lund, A. Laing, S. Rahimi-Keshari, T. Rudolph, J. L. OBrien, and T. C. Ralph, Phys. Rev. Lett.113, 100502 (2014). 11

work page 2014
[6]

Barkhofen, T

S. Barkhofen, T. J. Bartley, L. Sansoni, R. Kruse, C. S. Hamil- ton, I. Jex, and C. Silberhorn, Phys. Rev. Lett. 118, 020502 (2017)

work page 2017
[7]

Aaronson and A

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

work page 2011
[8]

T. R. Bromley, J. M. Arrazola, S. Jahangiri, J. Izaac, N. Que- sada, A. D. Gran, M. Schuld, J. Swinarton, Z. Zabaneh, and N. Killoran, Quantum Sci. Technol. 5, 034010 (2020)

work page 2020

Show all 60 references

[9]

J. Huh, G. G. Guerreschi, B. Peropadre, J. R. McClean, and A. Aspuru-Guzik, Nat. Photonics 9, 615 (2015)

2015
[10]

Jahangiri, J

S. Jahangiri, J. M. Arrazola, N. Quesada, and A. Delgado, arXiv preprint arXiv:2006.13339 (2020)

2020 arXiv
[11]

J. M. Arrazola and T. R. Bromley, Phys. Rev. Lett.121, 030503 (2018)

2018
[12]

J. M. Arrazola, T. R. Bromley, and P. Rebentrost, Phys. Rev. A 98, 012322 (2018)

2018
[13]

Banchi, M

L. Banchi, M. Fingerhuth, T. Babej, C. Ing, and J. M. Arrazola, Sci. Adv. 6, eaax1950 (2020)

2020
[14]

Schuld, K

M. Schuld, K. Br ´adler, R. Israel, D. Su, and B. Gupt, Phys. Rev. A 101, 032314 (2020)

2020
[15]

Br ´adler, S

K. Br ´adler, S. Friedland, J. Izaac, N. Killoran, and D. Su, arXiv preprint arXiv:1810.10644 (2018)

2018 arXiv
[16]

Br ´adler, P.-L

K. Br ´adler, P.-L. Dallaire-Demers, P. Rebentrost, D. Su, and C. Weedbrook, Physical Review A98, 032310 (2018)

2018
[17]

K. K. Sabapathy, H. Qi, J. Izaac, and C. Weedbrook, Phys. Rev. A 100, 012326 (2019)

2019
[18]

D. Su, C. R. Myers, and K. K. Sabapathy, Phys. Rev. A 100, 052301 (2019)

2019
[19]

Banchi, N

L. Banchi, N. Quesada, and J. M. Arrazola, Phys. Rev. A 102, 012417 (2020)

2020
[20]

M. Reck, A. Zeilinger, H. J. Bernstein, and P. Bertani, Phys. Rev. Lett. 73, 58 (1994)

1994
[21]

W. R. Clements, P. C. Humphreys, B. J. Metcalf, W. S. Kolthammer, and I. A. Walmsley, Optica3, 1460 (2016)

2016
[22]

de Guise, O

H. de Guise, O. Di Matteo, and L. L. S ´anchez-Soto, Phys. Rev. A 97, 022328 (2018)

2018
[23]

Garc ´ıa-Patr´on, J

R. Garc ´ıa-Patr´on, J. J. Renema, and V . Shchesnovich, Quantum 3, 169 (2019)

2019
[24]

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

2020
[25]

Neville, C

A. Neville, C. Sparrow, R. Clifford, E. Johnston, P. M. Birchall, A. Montanaro, and A. Laing, Nat. Phys. 13, 1153 (2017)

2017
[26]

Clifford and R

P. Clifford and R. Clifford, in Proceedings of the Twenty-Ninth Annual ACM-SIAM Symposium on Discrete Algorithms(SIAM, 2018), pp. 146–155

2018
[27]

Quesada, L

N. Quesada, L. Helt, J. Izaac, J. Arrazola, R. Shahrokhshahi, C. Myers, and K. Sabapathy, Phys. Rev. A100, 022341 (2019)

2019
[28]

Clifford and R

P. Clifford and R. Clifford, arXiv preprint arXiv:2005.04214 (2020)

2020 arXiv
[29]

J. Wu, Y . Liu, B. Zhang, X. Jin, Y . Wang, H. Wang, and X. Yang, Natl. Sci. Rev.5, 715 (2018)

2018
[30]

B. Gupt, J. M. Arrazola, N. Quesada, and T. R. Bromley, Quan- tum Information Processing 19, 1 (2020)

2020
[31]

Y . Li, M. Chen, Y . Chen, H. Lu, L. Gan, C. Lu, J. Pan, H. Fu, and G. Yang, arXiv preprint arXiv:2009.01177 (2020)

2020 arXiv
[32]

Jozsa, arXiv preprint quant-ph/0603163 (2006)

R. Jozsa, arXiv preprint quant-ph/0603163 (2006)

2006 arXiv
[33]

Vidal, Phys

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

2003
[34]

Montangero, Introduction to Tensor Network Methods (Springer, 2018)

S. Montangero, Introduction to Tensor Network Methods (Springer, 2018)

2018
[35]

A. E. Lita, A. J. Miller, and S. W. Nam, Opt. Express 16, 3032 (2008)

2008
[36]

Z. H. Levine, T. Gerrits, A. L. Migdall, D. V . Samarov, B. Calkins, A. E. Lita, and S. W. Nam, J. Opt. Soc. Am. B: Opt. Phys. 29, 2066 (2012)

2012
[37]

R. H. Hadﬁeld, Nat. Photonics 3, 696 (2009)

2009
[38]

Cifuentes and P

D. Cifuentes and P. A. Parrilo, Linear Algebra Appl. 493, 45 (2016)

2016
[39]

Barvinok, Combinatorics and complexity of partition func- tions, vol

A. Barvinok, Combinatorics and complexity of partition func- tions, vol. 9 (Springer, 2016)

2016
[40]

Quesada and J

N. Quesada and J. M. Arrazola, Phys. Rev. Research 2, 023005 (2020)

2020
[41]

Weedbrook, S

C. Weedbrook, S. Pirandola, R. Garc´ıa-Patr´on, N. J. Cerf, T. C. Ralph, J. H. Shapiro, and S. Lloyd, Rev. Mod. Phys. 84, 621 (2012)

2012
[42]

Seraﬁni, Quantum continuous variables: a primer of theo- retical methods (CRC press, 2017)

A. Seraﬁni, Quantum continuous variables: a primer of theo- retical methods (CRC press, 2017)

2017
[43]

Bj ¨orklund, B

A. Bj ¨orklund, B. Gupt, and N. Quesada, J. Exp. Algor. 24, 1 (2019)

2019
[44]

Br ´adler, R

K. Br ´adler, R. Israel, M. Schuld, and D. Su, arXiv preprint arXiv:1910.04022 (2019)

2019 arXiv
[45]

Y . Shi, M. Dehmer, X. Li, and I. Gutman, Graph polynomials (CRC Press, 2016)

2016
[46]

B. Wu, B. Cheng, F. Jia, J. Zhang, M.-H. Yung, and X. Sun, Sci. Bull. 65, 832 (2020)

2020
[47]

Quesada, J

N. Quesada, J. M. Arrazola, and N. Killoran, Phys. Rev. A 98, 062322 (2018)

2018
[48]

A. M. Dalzell, A. W. Harrow, D. E. Koh, and R. L. La Placa, Quantum 4, 264 (2020)

2020
[49]

Jahangiri, J

S. Jahangiri, J. M. Arrazola, N. Quesada, and N. Killoran, Phys. Rev. E 101, 022134 (2020)

2020
[50]

W. W. Hager, SIAM Rev. 31, 221 (1989)

1989
[51]

L. G. Valiant, Theoretical computer science 8, 189 (1979)

1979
[52]

H. J. Ryser, Combinatorial mathematics , vol. 14 (American Mathematical Soc., 1963)

1963
[53]

Bj ¨orklund, in Proceedings of the twenty-third annual ACM- SIAM symposium on Discrete Algorithms (SIAM, 2012), pp

A. Bj ¨orklund, in Proceedings of the twenty-third annual ACM- SIAM symposium on Discrete Algorithms (SIAM, 2012), pp. 914–921

2012
[54]

Cygan and M

M. Cygan and M. Pilipczuk, Inf. Comput. 243, 75 (2015)

2015
[55]

Schwartz, Linear Algebra Appl

M. Schwartz, Linear Algebra Appl. 430, 1364 (2009)

2009
[56]

Temme and P

K. Temme and P. Wocjan, arXiv preprint arXiv:1208.6589 (2012)

2012 arXiv
[57]

Muraleedharan, A

G. Muraleedharan, A. Miyake, and I. H. Deutsch, New J. Phys. 21, 055003 (2019)

2019
[58]

Lundow and K

P. Lundow and K. Markstr¨om, arXiv preprint arXiv:1904.06229 (2019)

2019 arXiv
[59]

Lubasch, A

M. Lubasch, A. A. Valido, J. J. Renema, W. S. Kolthammer, D. Jaksch, M. S. Kim, I. Walmsley, and R. Garc´ıa-Patr´on, Phys. Rev. A 97, 062304 (2018)

2018
[60]

R. Kan, J. Multivar. Anal. 99, 542 (2008)

2008

Pith tools

Reviewed August 27, 2026 · model on record in the stance chip above.

[1] [1]

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

work page 2017

[2] [2]

⨁M i=1 mi⨁M i=1 m∗ i ⨁M i=1 (ni + 1 2) ) . (21) We can write the adjacency matrix of the M modes as [3, 5 48] A =X(I2M−Q−1) = (B C CT B∗ ) , (22) B =U ( M⨁ i=1 λi ) UT =BT, (23) C =U ( M⨁ i=1 µi ) U† =C†, (24) λi = mi (1 +ni)2−|mi|2, (25) µi = 1− 1 +ni (1 +ni)2−|mi|2. (26) For pure states, one ﬁnds thatµi = 0 for alli. More generally, it holds thatB andC ...

work page

[3] [3]

Kruse, C

R. Kruse, C. S. Hamilton, L. Sansoni, S. Barkhofen, C. Silber- horn, and I. Jex, Phys. Rev. A 100, 032326 (2019)

work page 2019

[4] [4]

Rahimi-Keshari, A

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

work page 2015

[5] [5]

A. P. Lund, A. Laing, S. Rahimi-Keshari, T. Rudolph, J. L. OBrien, and T. C. Ralph, Phys. Rev. Lett.113, 100502 (2014). 11

work page 2014

[6] [6]

Barkhofen, T

S. Barkhofen, T. J. Bartley, L. Sansoni, R. Kruse, C. S. Hamil- ton, I. Jex, and C. Silberhorn, Phys. Rev. Lett. 118, 020502 (2017)

work page 2017

[7] [7]

Aaronson and A

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

work page 2011

[8] [8]

T. R. Bromley, J. M. Arrazola, S. Jahangiri, J. Izaac, N. Que- sada, A. D. Gran, M. Schuld, J. Swinarton, Z. Zabaneh, and N. Killoran, Quantum Sci. Technol. 5, 034010 (2020)

work page 2020