Pith. sign in

REVIEW 3 major objections 4 minor 40 references

A prepare-transform-measure scenario with any number of sequential transformations admits a noncontextual ontological model exactly when its COPE tensor admits a nonnegative factorization satisfying rank equalities on every stage, and this

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

A rank-based linear-algebra criterion and decision algorithm certify generalized (non)contextuality in prepare-transform-measure scenarios with any number of sequential transformations.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection Solid multi-stage COPE extension with a genuinely new sequential-only contextuality example, but the 'full decision procedure' claim outruns the atomic-only noncontextuality notion it actually decides. the 3 major comments →

arxiv 2607.26139 v1 pith:KQ3XLV3O submitted 2026-07-28 quant-ph cs.CC

Linear Algebra of Generalized Contextuality in All Prepare-Transform-Measure Scenarios

Theodoros Yianni , Nyan Raess , Farid Shahandeh This is my paper
classification quant-ph cs.CC
keywords generalized contextualityprepare-transform-measure scenarioCOPE tensornoncontextual ontological modelnonnegative matrix factorizationrank constraintsoperational theoriesgeneralized probabilistic theories
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 extends the statistical, GPT-free COPE approach to contextuality from prepare–measure experiments to prepare–transform–measure experiments with any number of sequential transformation stages. Its central claim is that such a scenario admits a noncontextual ontological model exactly when its COPE tensor—the array of all conditional outcome probabilities—has a nonnegative factorization into response functions, stochastic transformations, and epistemic states whose ranks match the ranks of the corresponding tensor flattenings. This converts a foundational question into a checkable linear-algebraic condition, and the paper backs it with an algorithm and complexity bound: for a fixed minimal GPT dimension the problem is polynomial in the number of procedures, with complexity linear-exponential in that dimension. Along the way it constructs an operational theory that is noncontextual in every single-stage view but becomes contextual when transformations are sequenced, showing that sequential composition is a genuine source of contextuality.

Core claim

For a PT^1...T^{k-2}M scenario, the COPE k-tensor C contains all prepare/transform/measure probabilities. An ontological model is a nonnegative factorization C = R Γ^{k-2} ... Γ^1 E with column-stochastic Γ^l and epistemic states E and response functions R. The paper's central result, Theorem 19, says the model is noncontextual iff the factors satisfy rank(R)=rank(C[E]), rank(E)=rank(C[P]), and rank(G^l)=rank(C[T^l]) for every transformation stage l, where G^l is the row-flattened matrix of the stage-l transformations and the flattenings are the standard mode unfoldings of C. Theorems 13, 18, and 20 turn this into a feasibility linear program after first constructing universal extremal facto

What carries the argument

The central object is the COPE tensor—the complete table of conditional outcome probabilities ordered by preparation, transformation stages, and measurement—and its nonnegative rank-restricted factorization. For each set of atomic procedures, the relevant flattening of C and its factor must have equal rank; these equirank equalities are exactly the quotienting conditions that turn a generic nonnegative pre-GPT (a generalized probabilistic theory before quotienting operational equivalences) into a noncontextual one. The decision procedure builds on restricted nonnegative matrix factorization: every candidate NCOM can be compressed to an extremal response function and state matrix, after which

Load-bearing premise

The whole decision procedure assumes noncontextuality is required only for operationally equivalent atomic procedures; equivalences between composite procedures (such as preparation followed by transformation) are explicitly not enforced, and if one demands those too, the rank criteria become incomplete.

What would settle it

Run a brute-force search over small COPE tensors, enumerating all finite ontic spaces and comparing the LP-based verdict of Theorem 18 with the exhaustive existence of a noncontextual model; any tensor where an NCOM exists but the LP reports infeasible would refute the claimed completeness, and any tensor where the LP succeeds but the reconstructed factors fail to be column-stochastic would refute soundness.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • For any fixed minimal GPT dimension, deciding contextuality in prepare-transform-measure scenarios with arbitrarily many stages is polynomial in the number of preparations, transformations, and measurement outcomes; the exponential cost sits only in the GPT dimension.
  • Sequential structure matters: the paper constructs a theory whose single-stage statistics are noncontextual but whose two-stage statistics are contextual, so contextuality can be a property of how transformations compose, not of any stage in isolation.
  • The 8-state single-qubit stabilizer theory is certified transformation-contextual, matching existing results, while the discrete toy-bit theory is certified noncontextual at every stage count.
  • Whether a composite transformation is treated as atomic or as sequential changes the verdict: promoting composites to atomic procedures introduces operational equivalences that a noncontextual model must respect, flipping some scenarios from noncontextual to contextual.

Where Pith is reading between the lines

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

  • Editorial inference: the rank equalities are stated only for atomic procedures, so the decision procedure answers the atomic version; a theory that also demands noncontextual representation of composite equivalences (e.g., P_i T_j ≃ P_k T_l) would need extra rank constraints of the sort the paper sketches and might yield different verdicts.
  • Editorial inference: the complexity result suggests a natural benchmark for contextuality measures: quantifying sequential contextuality should inherit the same linear-exponential-in-GPT-dimension difficulty, and one could look for reductions to or from other hidden-variable feasibility problems.
  • Editorial inference: because the criterion is purely statistical, it can likely be turned into an approximate, noise-tolerant witness by relaxing the rank equalities or by checking the LP with finite samples, an extension the paper lists as future work.
  • Editorial inference: the sequential-only contextuality example implies that circuit-level quantum advantage, where computation is a composition of gates, could be sourced from contextuality that no single gate exhibits; the paper's framework gives a route to search for such gates.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper extends the COPE-matrix framework for prepare-measure scenarios to prepare-transform-measure (PTM) scenarios with an arbitrary number of sequential transformation stages. It defines a COPE tensor encoding all conditional outcome probabilities, shows that preGPT/GPT models correspond to tensor-train factorizations, and proves (Theorems 8, 16, 19) that a noncontextual ontological model (NCOM) exists if and only if the COPE tensor admits a nonnegative factorization whose factors satisfy rank equalities equal to the ranks of the relevant tensor unfoldings. It then gives linear-programming-based decision procedures (Theorems 13, 18, 20) with complexity polynomial in the number of procedures and exponential in the minimal GPT dimension. The framework is applied to Spekkens' toy theory, the 8-state single-qubit stabilizer subtheory, and a custom 2D example exhibiting contextuality only in sequential scenarios.

Significance. If the central claim holds, the paper provides a statistics-first, GPT-free characterization of (a version of) generalized contextuality in sequential scenarios, substantially generalizing prior PM-only work. The constructive LP procedures and explicit complexity bounds are valuable, and the 2D sequential-contextuality example is conceptually interesting. However, the significance is tempered by a load-bearing scope restriction: the noncontextuality conditions are enforced only for atomic procedures, not for composite procedures. Since standard Spekkens contextuality applies to all operationally equivalent procedures, the paper's decision procedure decides a strictly weaker notion. The authors are transparent about this in the body, but the abstract and introduction present the result as a full decision procedure for generalized contextuality, which is overstated.

major comments (3)
  1. [Sec. III.B (after Eq. (14)); Sec. VI.D] The rank criteria (Eq. (14), Theorems 8/16/19) enforce noncontextuality only for atomic procedures. Eq. (13) defines noncontextuality via atomic equivalences, and the paragraph after Eq. (14) explicitly declines to enforce composite equivalences such as P_i T^1_j ≃ P_k T^1_l. This is not the standard Spekkens notion, which applies to all operationally equivalent procedures. Section VI.D provides a concrete witness: for the first theory, the criteria certify an NCOM, yet the composite equivalence T_D' ≃ T_D [Eq. (59)] cannot be respected by any full NCOM, since Γ(H) is parity-reversing and Γ(D) is parity-preserving, forcing Γ(H)Γ(D)=Γ(D) to fail. Thus the decision procedure is incomplete for generalized contextuality as standardly defined, and the abstract's claim of a 'full decision procedure for contextuality' is overbroad. The paper must either explicitly scope all claims to atomic-pro
  2. [Sec. VI.C (around Eqs. (55)-(56))] The proof that the NCOM of the 2D toy theory is unique up to relabelling is stated as 'clear from the sparsity patterns'. This uniqueness is essential for the subsequent claim that the PT1T2M scenario S2 is contextual while PT1M is noncontextual; if other NCOMs existed, the impossibility argument for S2 would not follow. A detailed derivation is needed, e.g., by systematically solving the factorization equations C = R Γ E with the fixed R and E from Lemma 10, or by a formal polytope argument.
  3. [Sec. IV.A (after Theorem 15)] The claim that the preGPT constructed from the minimal tensor-train decomposition is the smallest possible GPT is argued by saying that otherwise there would exist a tensor-train decomposition smaller than the minimal one. This implicitly assumes that the minimal GPT dimension equals the maximum of the tensor-train ranks. That equivalence is standard, but it should be stated and proved or explicitly cited, since the paper's minimality claim is used to identify 'r' in the complexity statements.
minor comments (4)
  1. [Sec. II.B] The notation for the number of measurements n_M and total outcome events n_E is introduced but not used consistently; e.g., the COPE tensor is said to be n_E × n_T1 × n_P, though n_E already includes all outcomes across measurements. Clarify the relationship.
  2. [Example 6] The heading 'an NCOM' contains a typo ('an' should be 'A').
  3. [Eq. (17)] The flattening order in Y^1_{j,(r(m-1)+n)} := T^1_j_{m,n} is implicit; state explicitly how the row-flattening index is ordered to avoid ambiguity.
  4. [Sec. VI.D] In the second theory, the sets T'_1 = T'_2 are written with a prime, and Eq. (60) uses rank(C[T'1]) and rank(G'1). Ensure that the prime notation is defined consistently and does not conflict with the earlier use of T^1_j for atomic transformations.

Circularity Check

0 steps flagged

No significant circularity: the rank-characterization theorem is proved in-paper; the main caveat is the explicitly admitted atomic-only operational-equivalence scope.

full rationale

The central derivation is self-contained rather than circular. Theorems 5, 8, 16, and 19 reduce NCOM existence to rank equalities through Lemmas 6 and 7, which are proved in the paper from the definitions of operational equivalence (Eq. 13) and preGPT/GPT quotienting (Eq. 9). The COPE tensor is not fitted to the target result; the examples are consistency checks on known models, and no parameter is tuned to force the claimed verdicts. Self-citations to Refs. [23,24,34] supply the COPE formalism and a shear-matrix parameterization, but the multi-stage characterization is not assumed by those works, and no uniqueness theorem is imported to forbid alternatives. The paper explicitly flags the atomic-only scope after Eq. (14): 'these conditions certify that operational equivalences at each stage are respected at the ontological level; equivalences between composite procedures may not be respected.' Sec. VI D then exhibits a concrete separation between the atomic rank criteria and the standard notion applied to composite equivalences such as T_D' ≃ T_D. This makes the 'full decision procedure' wording overbroad for generalized contextuality in the standard sense, but it is a scope/correctness caveat, not a circular reduction of the paper's claims to its inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 0 invented entities

The central claim rests on the COPE tensor being complete, finite discrete ontic space, and the atomic-only equivalence convention; no free parameters are fitted and no new entities are invented.

axioms (5)
  • domain assumption COPE tensor contains the complete probabilistic structure of the operational theory (Assumption 1)
    Used throughout to identify operational equivalences from identical tensor slices (Corollary 1); if procedures outside the scenario could distinguish states/effects/transformations, rank criteria would be incomplete.
  • domain assumption Ontic space is finite-dimensional and discrete
    Sec. III.B.1; the equivalence to nonnegative preGPTs and the RNMF/LP construction rely on finite ontic points; continuous ontic models are not treated.
  • ad hoc to paper Noncontextuality is required only for operational equivalences between atomic procedures
    Eq. (13) and the discussion after Eq. (14) explicitly exclude equivalences between composite procedures; this bounds the decision procedure's scope.
  • domain assumption Operational equivalences/convex mixtures are respected by linear, convexity-preserving GPT representations
    Standard GPT formalism assumed in Sec. III.A; needed for rank equalities to imply quotienting.
  • standard math Tucker and tensor-train decompositions/RNMF exist; vertex enumeration complexity bounds
    Background results used in Lemmas 2, 10, 14 and complexity theorems.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Linear Algebra of Generalized Contextuality in All Prepare-Transform-Measure Scenarios." pith.science (2026). https://pith.science/paper/KQ3XLV3O

@misc{pith2026260726139,
  author       = {Pith},
  title        = {Pith review of: Linear Algebra of Generalized Contextuality in All Prepare-Transform-Measure Scenarios},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KQ3XLV3O}},
  note         = {Machine review of arXiv:2607.26139}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Generalized contextuality is a canonical distinguishing property of nonclassical generalized probabilistic theories, in particular quantum mechanics. Methods for certification and characterization of generalized contextuality of a given generalized probabilistic theory are well developed for prepare-measure and single-stage prepare-transform-measure scenarios. In a recent work [arXiv:2512.10000], a bottom-up, statistics-first linear-algebraic framework for contextuality in prepare-measure scenarios was introduced. We extend this approach to operational scenarios with sequential transformations with an arbitrary number of stages. We give a full decision procedure for contextuality of such scenarios within operational theories and analyze its computational complexity. In particular, our decision procedure has a complexity linearly exponential in the minimum generalized probabilistic theory (GPT) dimension, and polynomial in the number of procedures. We demonstrate our framework and approach through multiple examples, including Spekkens' toy theory and the 8-state single-qubit stabilizer theory. In particular, we construct an operational theory in which contextuality manifests itself only in the sequential structure of the transformations. Our findings thus shed new light on the significant role of compositional structures in the phenomenon of generalized contextuality.

Figures

Figures reproduced from arXiv: 2607.26139 by Farid Shahandeh, Nyan Raess, Theodoros Yianni.

Figure 1
Figure 1. Figure 1: FIG. 1. Diagram of a COPE 3-tensor. Each axis corresponds [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

40 extracted references · 4 linked inside Pith

  1. [1]

    These provide an ontological model of the COPE 4-tensor via the decomposition Cijkl =R i:Γ2jΓ1kE:l (21) analogous to Eq

    General Ontological Models Similarly to the previous section, an ontological model of this scenario is given by a response function matrixR, an epistemic state matrixE, and nonnegative stochas- tic transformation matrices{Γ 1k}and{Γ 2j}represent- ingT 1 andT 2, respectively. These provide an ontological model of the COPE 4-tensor via the decomposition Cij...

  2. [2]

    Therefore the new normalized ontological model is an NCOM

    In short, the transformation Γ 1i →MΓ 1iM ′ does not destroy any equivalences. Therefore the new normalized ontological model is an NCOM. 12 This above normal form allows us to set up the follow- ing linear program to decide contextuality in thePT 1M scenario. Theorem 13.It can be decided whether aPT 1Msce- nario in an operational theory withn P preparati...

  3. [3]

    In this section, we tackle the problem of determining whether such a factorization exists, and computing one if it does

    Complexity of computing NCOMs inPT1MScenario In the previous section we showed that an NCOM of a PTM scenario amounts to a nonnegative factorization of the COPE 3-tensor satisfying particular rank constraints. In this section, we tackle the problem of determining whether such a factorization exists, and computing one if it does. We also offer a complexity...

  4. [4]

    This is justified because, following the argument in Lemma 7, an NCOM of the PT1T2Mscenario is simply a nonnegative GPT model

    NCOMs ofPT 1T2MScenario Following the characterization of ontological models for thePT 1T2Mscenario, we use the GPT conditions of The- orem 15 to express the conditions that must be satisfied by an NCOM of the scenario. This is justified because, following the argument in Lemma 7, an NCOM of the PT1T2Mscenario is simply a nonnegative GPT model. Indeed, th...

  5. [5]

    E. S. Simon Kochen, Indiana Univ. Math. J.17, 59 (1968)

    1968

  6. [6]

    Abramsky and A

    S. Abramsky and A. Brandenburger, New Journal of Physics13, 113036 (2011)

    2011

  7. [7]

    E. N. Dzhafarov and J. V. Kujala, Journal of Mathemat- ical Psychology74, 11 (2016)

    2016

  8. [8]

    Cabello, Phys

    A. Cabello, Phys. Rev. Lett.101, 210401 (2008)

    2008

  9. [9]

    Cabello, S

    A. Cabello, S. Severini, and A. Winter, Phys. Rev. Lett. 112, 040401 (2014)

    2014

  10. [10]

    J. S. Bell, Physics Physique Fizika1, 195 (1964)

    1964

  11. [11]

    J. S. Bell, Reviews of Modern Physics38, 447 (1966)

    1966

  12. [12]

    Budroni, A

    C. Budroni, A. Cabello, O. G¨ uhne, M. Kleinmann, and J.-A. Larsson, Rev. Mod. Phys.94, 045007 (2022)

    2022

  13. [13]

    R. W. Spekkens, Physical Review A71(2005)

    2005

  14. [14]

    Kunjwal and R

    R. Kunjwal and R. W. Spekkens, Phys. Rev. Lett.115, 110403 (2015)

    2015

  15. [15]

    Bermejo-Vega, N

    J. Bermejo-Vega, N. Delfosse, D. E. Browne, C. Okay, and R. Raussendorf, Phys. Rev. Lett.119, 120505 (2017)

    2017

  16. [16]

    Quantum computational advantage im- plies contextuality,

    F. Shahandeh, “Quantum computational advantage im- plies contextuality,” (2021), arXiv:2112.00024 [quant- ph]

    Pith/arXiv arXiv 2021

  17. [17]

    Schmid, H

    D. Schmid, H. Du, J. H. Selby, and M. F. Pusey, Phys. Rev. Lett.129, 120403 (2022)

    2022

  18. [19]

    Saha and A

    D. Saha and A. Chaturvedi, Phys. Rev. A100, 022108 (2019)

    2019

  19. [20]

    Roy and A

    P. Roy and A. K. Pan, Journal of Physics A: Mathemat- ical and Theoretical57, 375303 (2024)

    2024

  20. [21]

    Ambainis, M

    A. Ambainis, M. Banik, A. Chaturvedi, D. Kravchenko, and A. Rai, (2016)

    2016

  21. [22]

    Howard, J

    M. Howard, J. Wallman, V. Veitch, and J. Emerson, Nature510, 351–355 (2014)

    2014

  22. [23]

    Mansfield and E

    S. Mansfield and E. Kashefi, Phys. Rev. Lett.121, 230401 (2018)

    2018

  23. [24]

    J. H. Selby, E. Wolfe, D. Schmid, A. B. Sainz, and V. P. Rossi, Physical Review Letters132(2024)

    2024

  24. [25]

    Gitton and M

    V. Gitton and M. P. Woods, Quantum6, 732 (2022)

    2022

  25. [26]

    Soltani, M

    S. Soltani, M. Erba, D. Schmid, and J. H. Selby, (2025)

    2025

  26. [27]

    Character- izing contextuality via rank separation with applications to cloning,

    F. Shahandeh, T. Yianni, and M. Doosti, “Character- izing contextuality via rank separation with applications to cloning,” (2024), arXiv:2406.19382 [quant-ph]

    arXiv 2024

  27. [28]

    A unified linear algebraic framework for physical models and gen- eralized contextuality,

    F. Shahandeh, T. Yianni, and M. Doosti, “A unified linear algebraic framework for physical models and gen- eralized contextuality,” (2025), arXiv:2512.10000 [quant- ph]

    arXiv 2025

  28. [29]

    Noncontextuality inequal- ities for prepare-transform-measure scenarios,

    D. Schmid, R. D. Baldij˜ ao, J. H. Selby, A. B. Sainz, and R. W. Spekkens, “Noncontextuality inequal- ities for prepare-transform-measure scenarios,” (2024), arXiv:2407.09624 [quant-ph]

    Pith/arXiv arXiv 2024

  29. [30]

    Schmid, J

    D. Schmid, J. H. Selby, M. F. Pusey, and R. W. Spekkens, Quantum8, 1283 (2024)

    2024

  30. [31]

    Lillystone, J

    P. Lillystone, J. J. Wallman, and J. Emerson, Physical Review Letters122, 140405 (2019)

    2019

  31. [32]

    M. A. O. Vasilescu and D. Terzopoulos, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformat- ics)2350, 447 (2002)

    2002

  32. [33]

    M. A. O. Vasilescu and D. Terzopoulos, Proceedings - 2005 IEEE Computer Society Conference on Com- puter Vision and Pattern Recognition, CVPR 2005I, 547 (2005)

    2005

  33. [34]

    The ontological identity of empir- ical indiscernibles: Leibniz’s methodological principle and its significance in the work of einstein,

    R. W. Spekkens, “The ontological identity of empir- ical indiscernibles: Leibniz’s methodological principle and its significance in the work of einstein,” (2019), arXiv:1909.04628 [physics.hist-ph]

    Pith/arXiv arXiv 2019

  34. [35]

    However, one can trivially restoreλ- mediation by enlarging the ontic space to include the space of recipes, in which case the epistemic states are no longer identical

    To violate the necessity direction, the measurement re- sponse functions would need to depend directly on the preparation recipes, thereby violating theλ-mediation assumption. However, one can trivially restoreλ- mediation by enlarging the ontic space to include the space of recipes, in which case the epistemic states are no longer identical

  35. [36]

    Gillis and F

    N. Gillis and F. Glineur, Linear Algebra and its Appli- cations437, 2685 (2012)

    2012

  36. [37]

    Khachiyan, E

    L. Khachiyan, E. Boros, K. Borys, K. Elbassioni, and V. Gurvich, Discrete & Computational Geometry39, 174 (2008)

    2008

  37. [38]

    Complexity of contextu- ality,

    T. Yianni and F. Shahandeh, “Complexity of contextu- ality,” (2025), arXiv:2506.09133 [quant-ph]

    Pith/arXiv arXiv 2025

  38. [39]

    I. V. Oseledets, https://doi.org/10.1137/09075228633, 2295 (2011)

    doi:10.1137/09075228633 2011

  39. [40]

    R. W. Spekkens, Physical Review A - Atomic, Molec- ular, and Optical Physics75(2005), 10.1103/Phys- RevA.75.032110

    doi:10.1103/phys- 2005

  40. [41]

    Hausmann, N

    L. Hausmann, N. Nurgalieva, and L. del Rio, (2021). 27 A. Flattenings and F actorizations In this Appendix we list some examples of COPE tensor flattenings as well as their factorizations, which can be interpreted as an ontological model. These pertain to thePT 1Mscenario. We useξ i, µi and Γ i to denote theith response function, epistemic state and trans...

    2021

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.