Pith. sign in

REVIEW 2 major objections 4 minor 32 references

Linear Program Witness for Network Nonlocality in Arbitrary Networks

T0 review · 2 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

Pith's one-line read A linear program can certify network nonlocality in arbitrary networks by testing an auxiliary distribution q(a,λ) with five classes of linear constraints; infeasibility proves the observed correlations are not network-local.

desk verdict Useful LP construction for a restricted token model, but the completeness gap makes the network-nonlocality claim unproven. read the letter →

arxiv 2512.21962 v1 pith:A3FWU4KZ submitted 2025-12-26 quant-ph

classification quant-ph MSC 81P4081P1590C05 PACS 03.65.Ud03.67.-a
keywords networknonlocalitylinearprogramwitnessnetwork-localcorrelationsringsingle-photonWstatedomainasymmetryhidden-variableenumerationsufficientcriteria
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 introduces a linear-programming witness for network nonlocality: five classes of linear constraints on an auxiliary distribution q(a,λ) defined over a restricted set of outcomes and the hidden-variable strategies that produce them. If no q(a,λ) satisfies the constraints while matching the observed probabilities p(a), the authors conclude p(a) is network-nonlocal. The constraint classes are designed to be network-agnostic in kind, with explicit forms tailored to each network. They demonstrate the method on a six-party, four-source ring network distributing single-photon W states, where each party uses a tunable beamsplitter; the LP becomes infeasible for beamsplitter transmissivities t∈(0,0.292)∪(0.708,1), which they present as a certification of network nonlocality. The approach's decision variables scale as the number of observed outcomes times the number of restricted strategies, avoiding the combinatorial growth of existing approaches.

What carries the argument

The machinery is the auxiliary distribution q(a,λ) defined on the outcome subset O_S (where the number of single-click outputs equals the number of photons) and its pre-image S in the enumerated strategy space D, together with the five constraint classes: distribution validity, marginal agreement with p(a), strategy distribution derived from inferred λ-marginals, conditional independence for parties sharing the same hidden-variable pair, and domain asymmetry that equates differences between disjoint pre-image regions with statistics computable from p(a). Theorems 1 and 2 provide the bridge from outcome patterns to hidden-variable values: because there are fewer photons than parties, zero-cli

What would settle it

For a transmissivity in the certified range (say t=0.1), exhibit an explicit network-local model of the general form in Eq. (32) — arbitrary hidden-variable domains and arbitrary response functions — whose outcome distribution equals the quantum p(a) computed from Eq. (31). If such a model exists, the LP's infeasibility would be an artifact of the restricted strategy space rather than a certificate of network nonlocality.

Watch

Extended reading notes

Core claim

The central claim is that network nonlocality can be witnessed by linear programming once the hidden-variable space is carefully enumerated and the outcome set is restricted to events where each single photon is detected as a click at a distinct party. Under this restriction, Theorems 1 and 2 show that the positions of zero-click outcomes reveal the value of individual source variables λ_m, letting the authors express strategy probabilities, conditional independence, and 'domain asymmetry' as linear constraints on q(a,λ). For the six-party, four-source ring network, the resulting LP is infeasible for t∈(0,0.292)∪(0.708,1), certifying that the observed correlations cannot be produced by a net

Load-bearing premise

The load-bearing premise is that the enumeration of hidden-variable strategies — each source variable limited to naming one of its three target parties, and each party's response depending only on how many photons it receives — can represent every network-local model of the observed statistics; the paper only proves that every realizable outcome has at least one such strategy, not that no network-local model is lost by the restriction.

Editorial extensions

If this is right

  • If the LP infeasibility genuinely reflects network nonlocality, then for the six-party, four-source ring, the W-state correlations at beamsplitter transmissivities in (0,0.292) and (0.708,1) are certified nonlocal using only observed probabilities and the tunable parameter.
  • The five constraint classes provide a general template for constructing network-nonlocality witnesses: classes 1 and 2 are generic for any network, while classes 3–5 are adapted to the network's structure.
  • The decision-variable count |O_S|·|S| is upper bounded by d^N · P^M, meaning the witness can handle networks where existing combinatorial approaches become intractable.
  • The witness is sufficient but not necessary: a feasible LP does not imply network-locality, so the feasible regions (including t=0 and t=1) do not establish locality.

Reading between the lines

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

  • The paper leaves open whether the restricted strategy space (target-party-valued λ_m and photon-count-only responses) covers all network-local models; if it does not, the infeasibility at t∈(0,0.292)∪(0.708,1) could be a false positive. A proof of completeness for the enumeration, or a counterexample, would settle whether the witness is sound in general.
  • The success of the witness appears tied to the 'fewer photons than parties' structure, which makes hidden variables partially observable from zero-click patterns; similar LP witnesses may exist for other networks with conserved quantities that allow such inference.
  • The uniform strategy distribution and analytic domain asymmetries found for the six-party ring suggest symmetry reductions could scale the approach to larger rings without full enumeration.
  • The same five constraint classes could be adapted to detect full or genuine network nonlocality once the relevant network-structured notions are defined, since the constraints already enforce source independence among all sources.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper proposes a linear-programming witness for network nonlocality. It defines an auxiliary distribution q(a,λ) on a restricted outcome subset O_S and strategy set S, and imposes five classes of linear constraints: distribution validity, marginal agreement with observed p(a), strategy distribution, conditional independence, and domain asymmetry. The claim is that infeasibility of this LP is a sufficient certificate that p(a) is not network-local. The method is specialized to ring networks with tripartite single-photon W states and click/no-click detectors, and is demonstrated for a 6-party, 4-source ring, where the LP is reported infeasible for approximately t∈(0,0.292)∪(0.708,1).

Significance. If valid, the approach would be a significant practical alternative to inflation: the decision-variable count is |O_S|·|S| rather than the combinatorial clone count, and the paper provides explicit analytic constraint forms and a comparison of ECOS, SCS, and GLPK. The derivations leading to Eq. (66) are internally consistent for the specific token strategy model defined in Def. 1. However, the central gap is that this strategy model is a strict subset of the network-local models of Eq. (32); the paper does not prove that every network-local decomposition can be represented in D. The witness therefore does not certify standard network nonlocality as claimed.

major comments (2)
  1. [Sec. III, V (Def. 1)] The central claim (Sec. III, before Eq. (15)) that LP infeasibility implies p(a) is network-nonlocal requires that the LP constraints are satisfied by every distribution admitting the decomposition in Eq. (32). Definition 1 restricts each λ_m to a three-valued 'target party' and the image construction in Step 1 assumes a party's output depends only on how many photons it receives. Eq. (32) permits arbitrary λ_m domains and arbitrary response functions p_n(a_n|λ_m,λ_m'), including stochastic and photon-number-nonconserving response functions. The verification in Step 1 ('there exists a valid strategy λ_j∈D for every realizable outcome a_i∈O') is only a support condition F(D)=O; it does not show that an arbitrary network-local model can be expressed as a convex mixture over D. Hence Classes 1–5 are not necessary conditions for network locality, and the numerical infeasibility in Sec. VI do
  2. [Sec. V (Thm. 1, Eq. (66))] Theorem 1 infers λ_m=A_n from the pattern of zeros in O_S. This inference is valid only for the token model of Def. 1, where a source's λ_m value is the party receiving a photon and a receiving party must produce a click. Under Eq. (32), a network-local model may have p_n(0|λ_m,λ_m')>0 or may output L/R/2 according to an arbitrary function of the two incoming λ's; the observed zero pattern then carries no information about λ_m. Therefore the marginals μ(λ_m=A_n) computed from p(a) via Eq. (66) are not the marginals of a general network-local model, and the strategy-distribution and domain-asymmetry constraints built on them are not necessary for network locality. This is a load-bearing gap in the derivation of Classes 3 and 5.
minor comments (4)
  1. [Sec. VI (Results)] The sentence 'For any value of T>0, the program is infeasible' is confusing; the tolerance-minimized LP is feasible by construction for sufficiently large T. Please rephrase to state that the exact LP with zero tolerances is infeasible and T is the minimal total violation required.
  2. [Fig. 5] The axes and the meaning of T should be given in the caption; currently only transmissivity t is mentioned.
  3. [Sec. III, footnote 3] The reduction of inputs to fixed settings by mapping inputs to outputs of new parties may change the network structure; a reader would benefit from a precise statement of how the LP constraints adapt when inputs are present.
  4. [Sec. VI, Table I] The strategy labels λ_0...λ_29 and the notation F(λ)↦O_S would be easier to follow if the table also explained in text that each outcome pattern has two supporting strategies.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the LP constraints are derived necessity conditions on p(a), not fitted predictions; the main concern is a soundness gap, not circularity.

full rationale

The paper's central derivation takes observed statistics p(a) and builds a feasibility LP over an auxiliary distribution q(a,λ). The constraints in Classes 1 and 2 are exactly the definition of q(a) as the renormalized p(a) (Eqs. 19-20), so they are not predictions but bookkeeping. Class 3 computes μ(λ_m=A_n) from p(a) via Eq. (66) and then imposes factorization (Eq. 62); this is a derived necessary condition for a network-local model, not a fitted parameter being relabeled as a prediction. Class 5's Γ_Op is computed from p(a) and then equated to Δ_Op; while this is close to an identity after the definitions, it is presented as a constraint derived from the restricted model, and the paper does not claim it alone certifies nonlocality. The only self-citation, Ref. [1], introduces the general methodology but the current paper re-derives the procedure in full, so it is not load-bearing. The reviewer's main concern—that Definition 1 restricts λ_m to target-party values and response functions are effectively restricted, so LP infeasibility only rules out the restricted model rather than all network-local models of Eq. (32)—is a soundness/completeness gap, not a circularity. There is no evidence that constraints were chosen after seeing p(a) to force infeasibility: the beamsplitter transmissivity t is scanned, not optimized, and the witness is explicitly only sufficient. Therefore no circular step can be exhibited, and the circularity score is low. The soundness gap should be addressed as a correctness issue, not as circularity.

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

The main load-bearing assumptions are the strategy-space restriction and photon-number conservation. No fitted numerical parameters are used: t is scanned, and the tolerance objective is an infeasibility measure rather than a fitted constant. No new physical entities are introduced.

assumptions (6)
  • ad hoc to paper Each source variable λ_m can be restricted to one of the three parties it connects to (single-photon token model); outcomes are determined by the number of tokens received.
    Used in Def. 1 and Step 1 to enumerate |D|=3^M strategies. Load-bearing: if this is not without loss of generality, LP infeasibility does not certify standard network nonlocality. The paper provides no proof against arbitrary network-local models of Eq. (32).
  • domain assumption Photon number is conserved in the local model; each source emits exactly one photon, so impossible outcomes are excluded from O_S.
    Used in Def. 3, Prop. 2 and Eq. (31). It matches the quantum W-state experiment but is not a consequence of network locality alone.
  • domain assumption Detectors are non-photon-number-resolving and the POVM is Eq. (30) with equal transmissivity t and phase φ=0.
    Needed to compute p(a) and define the outcome alphabet; assumed from the photonic setup.
  • domain assumption The ring network has N≥6, N≡0 mod 3, M=2N/3 tripartite sources, each party connected to exactly two sources, and no two sources signal to exactly the same set of parties.
    The family of networks in Sec. IV; the concrete 6-party/4-source example satisfies it. This limits the 'arbitrary network' claim.
  • domain assumption Observed statistics are memoryless (no temporal memory across rounds).
    Mentioned in Sec. III: 'we note for network scenarios, this corresponds to the memoryless regime [24]'.
  • standard math Linear programming solvers (ECOS, SCS, GLPK) produce correct feasibility answers at the stated tolerances.
    The numerical certificate relies on solver convergence; not machine-checked.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Linear Program Witness for Network Nonlocality in Arbitrary Networks." pith.science (2026). https://pith.science/paper/A3FWU4KZ

@misc{pith2026251221962,
  author       = {Pith},
  title        = {Pith review of: Linear Program Witness for Network Nonlocality in Arbitrary Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A3FWU4KZ}},
  note         = {Machine review of arXiv:2512.21962}
}
read the original abstract

Network nonlocality extends Bell nonlocality to settings with multiple independent sources and parties. Certifying it in quantum information processing tasks requires suitable witnesses. However, in contrast to local correlations, the set of network-local correlations is non-convex. This non-convexity makes certifying network nonlocality a highly non-trivial task. Existing approaches involve leveraging network-specific properties, or inflation-based methods whose constraints grow combinatorially in the number of local variables. In this work, we introduce a linear programming witness for network nonlocality built from five classes of linear constraints. These classes are network-agnostic, although the explicit forms of the constraints must be tailored to a specific network's structure. We use the procedure to construct network nonlocality witnesses for a family of ring networks and certify network nonlocality for a concrete example, relying only on observed probabilities and a tunable experimental parameter. Our work advances the search for efficient witnesses to certify network nonlocality across diverse quantum network architectures.

Figures

Figures reproduced from arXiv: 2512.21962 by the authors.

Figure 1
Figure 1. (a) Bipartite Scenario with one source S1 characterized by one LV λ1, distributing states to two parties, A and B. (b) Bilocal scenario with two sources S1 and S2, characterized by two LVs λ1 and λ2, distributing states to three parties A, B, and C. The parties receive inputs (x, y, z), determining their measurement settings and return measurement outcomes (a, b, c). witnesses for standard Bell nonlocality, as they … view at source ↗
Figure 2
Figure 2. D denotes the LV domain and O denotes the output space, which are related via the map F(D) 7→ O. S (1) p and S (2) p denote subregions in the LV space, Op = F(S (1) p ⊔ S(2) p ) denotes a subregion in output space, and λj and λ ′ j denote specific strategies in S (1) p and S (2) p respectively. λ1 → α, λ ¯ 2 → γ, ¯ (λ1, λ2) → β¯}), and wα¯β¯γ¯ indicates the probability of strategy α¯β¯γ¯ to occur. Note that Eqn. (8)… view at source ↗
Figure 3
Figure 3. Ring network with M sources, S1, . . . , SM, distributing tripartite single-photon W states to N parties, A1, . . . , AN . Each party is equipped with a tunable beamsplitter and a non-photon-number-resolving detector in each optical mode. Each source Sm has an associated LV λm. For this specific source-to-party configuration, odd-numbered sources signal to the next-nearest-neighbor to the left of their middle party … view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: 6-party ring network with 4 sources, characterized by their LVs, [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Numerical witness of network nonlocality. [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references

  1. [1]

    Linear programming ap- proach for demonstrating network nonlocality for arbitrary networks,

    S. Hayes-Shuptar, D. Bhatti, and D. Elkouss, “Linear programming ap- proach for demonstrating network nonlocality for arbitrary networks,” in Proceedings of the 2nd Workshop on Quantum Networks and Distributed Quantum Computing, 2025, pp. 21–27

  2. [2]

    Bell nonlocality,

    N. Brunner, D. Cavalcanti, S. Pironio, V . Scarani, and S. Wehner, “Bell nonlocality,”Reviews of modern physics, vol. 86, no. 2, pp. 419–478, 2014

  3. [3]

    On the einstein podolsky rosen paradox,

    J. S. Bell, “On the einstein podolsky rosen paradox,”Physics Physique Fizika, vol. 1, no. 3, p. 195, 1964

  4. [4]

    No signaling and quantum key distribution,

    J. Barrett, L. Hardy, and A. Kent, “No signaling and quantum key distribution,”Physical review letters, vol. 95, no. 1, p. 010503, 2005

  5. [5]

    Random numbers certified by bell’s theorem,

    S. Pironio, A. Ac ´ın, S. Massar, A. B. de La Giroday, D. N. Matsukevich, P. Maunz, S. Olmschenk, D. Hayes, L. Luo, T. A. Manninget al., “Random numbers certified by bell’s theorem,”Nature, vol. 464, no. 7291, pp. 1021–1024, 2010

  6. [6]

    Device-independent security of quantum cryptography against collec- tive attacks,

    A. Ac ´ın, N. Brunner, N. Gisin, S. Massar, S. Pironio, and V . Scarani, “Device-independent security of quantum cryptography against collec- tive attacks,”Physical Review Letters, vol. 98, no. 23, p. 230501, 2007

  7. [7]

    The quantum internet,

    H. J. Kimble, “The quantum internet,”Nature, vol. 453, no. 7198, pp. 1023–1030, 2008

  8. [8]

    Quantum internet: A vision for the road ahead,

    S. Wehner, D. Elkouss, and R. Hanson, “Quantum internet: A vision for the road ahead,”Science, vol. 362, no. 6412, p. eaam9288, 2018

Show all 32 references
  1. [9]

    “Event- ready-detectors

    M. ˙Zukowski, A. Zeilinger, M. A. Horne, and A. K. Ekert, ““Event- ready-detectors” Bell experiment via entanglement swapping,”Physical Review Letters, vol. 71, pp. 4287–4290, 1993

  2. [10]

    Bell nonlocality in networks,

    A. Tavakoli, A. Pozas-Kerstjens, M.-X. Luo, and M.-O. Renou, “Bell nonlocality in networks,”Reports on Progress in Physics, vol. 85, no. 5, p. 056001, 2022

  3. [11]

    Device- independent characterization of entanglement based on bell nonlocality,

    G. Chen, W.-H. Zhang, P. Yin, C.-F. Li, and G.-C. Guo, “Device- independent characterization of entanglement based on bell nonlocality,” Fundamental Research, vol. 1, no. 1, pp. 27–42, 2021

  4. [12]

    Quantifying bell nonlocality with the trace distance,

    S. G. d. A. Brito, B. Amaral, and R. Chaves, “Quantifying bell nonlocality with the trace distance,”Physical Review A, vol. 97, no. 2, p. 022111, 2018

  5. [13]

    Two-party bell inequalities derived from combinatorics via triangular elimination,

    D. Avis, H. Imai, T. Ito, and Y . Sasaki, “Two-party bell inequalities derived from combinatorics via triangular elimination,”Journal of Physics A: Mathematical and General, vol. 38, no. 50, p. 10971, 2005

  6. [14]

    Quantum inflation: A general approach to quantum causal compatibility,

    E. Wolfe, A. Pozas-Kerstjens, M. Grinberg, D. Rosset, A. Ac ´ın, and M. Navascu ´es, “Quantum inflation: A general approach to quantum causal compatibility,”Physical Review X, vol. 11, no. 2, p. 021043, 2021

  7. [15]

    Bounding the sets of classical and quantum correlations in networks,

    A. Pozas-Kerstjens, R. Rabelo, Ł. Rudnicki, R. Chaves, D. Cavalcanti, M. Navascu´es, and A. Ac´ın, “Bounding the sets of classical and quantum correlations in networks,”Physical review letters, vol. 123, no. 14, p. 140503, 2019

  8. [16]

    Single-photon nonlocality in quantum networks,

    P. Abiuso, T. Kriv ´achy, E.-C. Boghiu, M.-O. Renou, A. Pozas-Kerstjens, and A. Ac´ın, “Single-photon nonlocality in quantum networks,”Physical Review Research, vol. 4, no. 1, p. L012041, 2022

  9. [17]

    Semidefinite programming,

    L. Vandenberghe and S. Boyd, “Semidefinite programming,”SIAM review, vol. 38, no. 1, pp. 49–95, 1996

  10. [18]

    Proposed ex- periment to test local hidden-variable theories,

    J. F. Clauser, M. A. Horne, A. Shimony, and R. A. Holt, “Proposed ex- periment to test local hidden-variable theories,”Physical review letters, vol. 23, no. 15, p. 880, 1969

  11. [19]

    Geometry of the set of quantum correlations,

    K. T. Goh, J. Kaniewski, E. Wolfe, T. V ´ertesi, X. Wu, Y . Cai, Y .-C. Liang, and V . Scarani, “Geometry of the set of quantum correlations,” Physical Review A, vol. 97, no. 2, p. 022104, 2018

  12. [20]

    Violations of local realism by two entangled n- dimensional systems are stronger than for two qubits,

    D. Kaszlikowski, P. Gnaci ´nski, M. ˙Zukowski, W. Miklaszewski, and A. Zeilinger, “Violations of local realism by two entangled n- dimensional systems are stronger than for two qubits,”Physical Review Letters, vol. 85, no. 21, p. 4418, 2000

  13. [21]

    Bilocal versus non- bilocal correlations in entanglement-swapping experiments,

    C. Branciard, D. Rosset, N. Gisin, and S. Pironio, “Bilocal versus non- bilocal correlations in entanglement-swapping experiments,”Physical Review A—Atomic, Molecular, and Optical Physics, vol. 85, no. 3, p. 032119, 2012

  14. [22]

    Beyond bell’s theorem: correlation scenarios,

    T. Fritz, “Beyond bell’s theorem: correlation scenarios,”New Journal of Physics, vol. 14, no. 10, p. 103001, 2012

  15. [23]

    Some np-complete problems in quadratic and nonlinear programming,

    K. G. Murty and S. N. Kabadi, “Some np-complete problems in quadratic and nonlinear programming,” Tech. Rep., 1985

  16. [24]

    Memory attacks in network nonlocality and self-testing,

    M. Weilenmann, C. Budroni, and M. Navascues, “Memory attacks in network nonlocality and self-testing,”Quantum, vol. 9, p. 1735, 2025

  17. [25]

    Bell nonlocality in quantum networks with unreliable sources: Loophole-free postelection via self- testing,

    S. Boreiri, N. Brunner, and P. Sekatski, “Bell nonlocality in quantum networks with unreliable sources: Loophole-free postelection via self- testing,” 2025

  18. [26]

    Ecos: An socp solver for embedded systems,

    A. Domahidi, E. Chu, and S. Boyd, “Ecos: An socp solver for embedded systems,” in2013 European Control Conference (ECC). IEEE, 2013

  19. [27]

    Conic optimization via operator splitting and homogeneous self-dual embedding,

    B. O’Donoghue, E. Chu, N. Parikh, and S. Boyd, “Conic optimization via operator splitting and homogeneous self-dual embedding,”Journal of Optimization Theory and Applications, vol. 169, no. 3, pp. 1042–1068, 2016

  20. [28]

    Characterizing the nonlocal correlations created via entanglement swapping,

    C. Branciard, N. Gisin, and S. Pironio, “Characterizing the nonlocal correlations created via entanglement swapping,”Physical review letters, vol. 104, no. 17, p. 170401, 2010

  21. [29]

    The inflation technique completely solves the causal compatibility problem,

    M. Navascu ´es and E. Wolfe, “The inflation technique completely solves the causal compatibility problem,”Journal of Causal Inference, vol. 8, no. 1, pp. 70–91, 2020

  22. [30]

    Network nonlocality via rigidity of token counting and color matching,

    M.-O. Renou and S. Beigi, “Network nonlocality via rigidity of token counting and color matching,”Physical Review A, vol. 105, no. 2, 2022

  23. [31]

    Full network nonlocal- ity,

    A. Pozas-Kerstjens, N. Gisin, and A. Tavakoli, “Full network nonlocal- ity,”Physical review letters, vol. 128, no. 1, p. 010403, 2022

  24. [32]

    Genuine network quantum nonlocality and self-testing,

    I. ˇSupi´c, J.-D. Bancal, Y . Cai, and N. Brunner, “Genuine network quantum nonlocality and self-testing,”Physical Review A, vol. 105, no. 2, p. 022206, 2022. APPENDIXA PROOF OFTHEOREM1 We first claim the set of outcomes inO S compatible with a(m,n) are different from the set ...

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.