REVIEW 2 major objections 6 minor 1 cited by
Facets of Non-locality and Advantage in Entanglement-Assisted Classical Communication Tasks
T0 review · 2 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read This paper proves an exact equivalence between violation of a task-tailored Bell inequality and quantum advantage in the corresponding classical-communication task, and constructs communication tasks in which any non-local facet…
desk verdict A solid wire-cutting equivalence and a genuinely new wire-reading tool, but the advertised qutrit-over-qubit advantage is not proven because the qubit upper bound only covers rank-1 projective measurements. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
Three mechanisms carry the argument. (1) Wire-cutting: for a task with channel $T$ and payoff weights $w$, define the Bell functional $B_{S,T}(P)=\sum_{m,\tau,\tau',n} w^m_n\, T(\tau'|\tau)\, P(\tau,n|m,\tau')$; because the channel couples Alice's sent message to Bob's received message, the payoff equals the Bell functional, giving the violation–advantage equivalence. (2) Wire-reading: the classical message can be read without disturbance, so the payoff may include the received message $\tau'$ as an observable; Proposition 1 reduces such tasks to ordinary tasks with a penalty term, and this visibility is what lets non-local assistance show up when the original payoff is already saturable by shared randomness. (3) Non-local facets: a face of the no-signalling polytope containing no local extremal point; Lemma 2 asserts that on such a facet, for any function $L_B$ there is an Alice input $x$ such that $P(a,b=L_B(y)|x,y)=0$ for every output $a$ on some non-empty set of $y$'s. That zero-probability structure lets Alice encode so that Bob's output always avoids the forbidden image, giving algebraic maximum payoff.
What would settle it
Take a specific non-local facet such as the I3322 facet with the two extremal correlations given in Section 3.2.2 and enumerate all $2^3=8$ functions $L_B:[3]\to[2]$, computing the sets $\Phi_{x,a}^{L_B}$ for each extremal correlation; if for some $L_B$ every input $x$ has some output $a$ with $P(a,L_B(y)|x,y)>0$ for all $y$, then Lemma 2 fails and the claimed algebraic-maximum payoff for the corresponding facet-tailored task cannot hold.
Extended reading notes
Core claim
The central claim is Theorem 1: for a communication task $C_{M,N}[T,\{w\}]$ with local bound $s_{\Lambda}$, a no-signalling correlation $P$ violates the associated Bell inequality $B_{S,T}(P)>s_{\Lambda}$ if and only if using $P$ as assistance to the channel gives payoff $S(N)>s_{\Lambda}$; the same equality holds between the maximum quantum payoff and the maximum quantum violation of the associated Bell inequality. The proof runs through the “wire-cutting” identity $S(N)=B_{S,T}(P)$, where the classical wire is cut at the channel and the received message becomes Bob's input to the shared correlation. The paper then introduces wire-reading, in which the payoff also records the received classical message; this reveals non-classical assistance in scenarios where the message-independent payoff shows no advantage. Using wire-reading, the authors define tasks $CS[d,k]$ and facet-tailored tasks $CS[\{P^{*}_{NL}\},d,k]$ in a receiver-without-input prepare-and-measure scenario, prove a tight shared-randomness bound for each, and show that any correlation on a non-local facet of the no-signalling polytope attains the algebraic maximum payoff. Numerically, two-qubit entangled states give payoffs above the local bound for several $(d,k)$, and for $CS[2,3]$ a two-qutrit strategy with non-projective measurements beats the optimal two-qubit strategy, with matching NPA level-2 upper bound.
Load-bearing premise
The paper's results rest on Lemma 2's geometric claim that on a non-local facet—a face of the no-signalling polytope containing no local extreme point—every assignment of a forbidden output for each of Bob's inputs has some Alice input under which those forbidden outputs have exactly zero probability; if a facet in the standard sense contains local vertices, or if that zero-probability structure fails, the optimal-payoff proofs for the constructed task families do not go through.
Editorial extensions
If this is right
- For every correlation-assisted bounded communication task, Bell violation and communication advantage are two readings of the same linear functional: the maximum quantum-assisted payoff equals the maximum quantum violation of the associated Bell inequality.
- The NPA hierarchy and the Navascués–Vértesi method therefore become tools for bounding entanglement-assisted communication advantage, not just Bell violations.
- Wire-reading makes the classical message part of the score, so tasks that look classically optimal without wire-reading can still certify non-local assistance: the four-input, one-bit example goes from no visible advantage to local bound $3/4$.
- In $CS[d,k]$, any correlation on a non-local facet achieves payoff $1$, and its mixtures with white noise are advantageous whenever the noise fraction stays below $1/k^{d-1}$; for the facet-tailored tasks the corresponding threshold is $p>1/2$.
- Quantum states provide explicit advantages in these tasks, and $CS[2,3]$ certifies that a two-qutrit entangled state with non-projective measurements outperforms two-qubit entanglement, so the tasks are sensitive to the local dimension of the shared state.
Reading between the lines
- Beyond the paper: the wire-reading construction suggests a general recipe—promote any classical message to an observable in the payoff—that could turn other classically saturable communication tasks into non-locality witnesses, including tasks with noisy channels or multiple rounds.
- Beyond the paper: the facet-tailoring method indicates that each face of the no-signalling polytope can be probed by a dedicated communication task; enumerating all facets of small polytopes would yield a catalogue of communication tasks whose optimal payoff singles out that facet.
- Beyond the paper: the qutrit-over-qubit separation suggests communication payoffs may serve as device-independent dimension witnesses; one testable extension is to check whether the $CS[2,3]$ payoff gap persists under noise and under finite measurement precision.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper introduces two techniques, wire-cutting and wire-reading, for studying correlation-assisted classical communication tasks in a prepare-and-measure scenario. Theorem 1 states that, for any correlation-assisted bounded classical communication task, a no-signalling correlation violates the associated Bell inequality if and only if it provides an advantage over shared randomness in the task, with an analogous equivalence for quantum correlations. The paper then constructs two families of tasks, CS[d,k] and CS[{P*_NL},d,k], in which correlations on the authors' 'non-local facets' achieve the algebraic maximum payoff while shared randomness is strictly suboptimal. Numerical see-saw and NPA results are reported, including the headline claim that the task CS[d=2,k=3] exhibits a qutrit-over-qubit entanglement advantage and enables device-independent certification of local dimension.
Significance. If the main claims hold, Theorem 1 gives a clean operational equivalence between Bell-violation and communication advantage, and wire-reading is a genuinely interesting conceptual addition that can certify non-classical assistance in scenarios where the usual payoff would not. The task families are elegant and the local bounds are derived exactly and without fitted parameters, which is a strength. The advertised dimensional-advantage example is, however, the least secure part of the paper: it rests on numerical upper bounds that currently do not cover all two-qubit strategies. The paper also uses a nonstandard notion of 'non-local facet' whose geometric status is not fully clarified. These issues are fixable within the manuscript's scope, so the appropriate decision is major revision rather than rejection.
major comments (2)
- [Section 3.3, Tables 2-3] The claimed qutrit-over-qubit advantage for CS[d=2,k=3] is not established for all two-qubit strategies. The value 0.93491 is described as an upper bound 'on the quantum payoff with two-qubit entanglement and rank 1 projectors' from the Navascués-Vértesi method [31]. A general POVM on a qubit can be Naimark-dilated to a projective measurement on a larger local Hilbert space, so a bound restricted to rank-1 projectors on two qubits does not rule out two-qubit POVM strategies exceeding 0.93491. The see-saw lower bounds in Table 2 also appear to be over projective qubit strategies. Consequently, the statements in Section 3.3 that 'assistance from two-qubit entangled states leads to a lower payoff' and that CS[2,3] enables device-independent certification of local dimension are stronger than what is proven. The authors should either extend the bound to arbitrary POVMs on two-qubit states or explicitly restrict the dimensional-advantage claim to rank-1 projective measurements.
- [Section 2.1, Lemma 2 and Appendix D] The paper's 'non-local facet' is not a facet in the standard polytope sense, and the proofs of Theorems 3 and 5 depend on this nonstandard notion. The definition before Lemma 2 should be stated as an explicit assumption (for example, a face of the no-signalling polytope containing no local extreme points) rather than being presented as a standard geometric object. The examples, in particular Face{P_NL^(1), P_NL^(2)} for I3322 in Section 3.2.2, should be verified to be faces or facets of the relevant no-signalling polytope under that definition. Without this, the claim that 'any non-local facet' leads to optimal payoff in the first task family is not connected to the usual geometry of no-signalling polytopes. In addition, the proof of Lemma 2 should state explicitly that the weight p in the decomposition P_NL = p P_L + (1-p) P~ is chosen as a sufficiently small positive number, bounded by the minimum positive probability among the finitely many relevant entries.
minor comments (6)
- [Appendix D] The decomposition step in the proof of Lemma 2 is stated as 'Clearly' but requires the finite-minimum argument; please spell out that p is chosen to be the minimum of the relevant positive probabilities.
- [Appendix F] The notation N(tau', n != m_{tau'} | m) is used without definition; it should be written as sum_{n != m_{tau'}} N(tau', n | m).
- [Proposition 1, Appendix B] The infinite penalty -infinity in the payoff (25) is handled informally; please add a short argument that the optimal payoff is recovered in the limit Theta -> infinity.
- [Section 2.1] There is a typo in the definition of the non-local facet: 'some son-local facet' should read 'some non-local facet'.
- [Table 1] The relation between the task in Table 1 and the task CS[2,2] in Section 3.1 is not immediately transparent; please spell out the mapping between the four inputs and the pairs m=(m1,m2).
- [Section 3.3] The numerical claims in Tables 2 and 3 are not reproducible from the text; please describe the see-saw and NPA implementations (solver, convergence criteria, measurement parameterizations) or make code available.
Circularity Check
No significant circularity: the central theorem is a transparent construction and the numerical claims rest on independent computational methods.
full rationale
The paper's derivation chain is self-contained and does not reduce to its own inputs. Theorem 1's wire-cutting equivalence is explicitly a construction: the Bell functional B_{S,T} in Eq. (8) is defined from the task's payoff coefficients and the channel T, so the identity S(N)=B_{S,T}(P) in Eqs. (11)-(12) is a substitution rather than a fitted prediction. The paper states that the inequality is 'tailored' to the task, so this is a proof technique, not a disguised empirical claim. Lemma 1 computes the local bound independently from the shared-randomness characterization, and Theorems 2-5 rely on explicit protocols built from the zero-pattern Lemma 2. The second task family is transparently constructed around a given non-local facet, which is an existence construction rather than a hidden circular step. The numerical claims in Section 3.3 use standard NPA hierarchies, see-saw lower bounds, and the independent Navascués-Vértesi method; no parameter is fitted to the payoff being predicted. The self-citations to [3] and [4] are contextual and non-load-bearing. A legitimate correctness concern remains that the two-qubit upper bound 0.93491 is stated for rank-1 projective measurements, so the advertised qutrit-over-qubit and device-independent-dimension-certification claims may be stronger than what is proven for general two-qubit POVMs, but that is a correctness risk rather than a circularity.
Assumptions & free parameters
assumptions (4)
- domain assumption Assistance by a no-signalling correlation can be modeled by a single use with local pre- and post-processing and a classical channel wiring (Eq. (3)).
- domain assumption A classical message can be read without disturbing the communication.
- standard math The NPA hierarchy and the Navascués-Vértesi method provide valid upper bounds on quantum correlations.
- ad hoc to paper Non-local facets, defined as faces of the no-signalling polytope containing only non-local extreme points, exist for the scenarios considered.
Cite this review
Pith. "Pith review of Facets of Non-locality and Advantage in Entanglement-Assisted Classical Communication Tasks." pith.science (2026). https://pith.science/paper/TVBFN3LN
@misc{pith2026250710830,
author = {Pith},
title = {Pith review of: Facets of Non-locality and Advantage in Entanglement-Assisted Classical Communication Tasks},
year = {2026},
howpublished = {\url{https://pith.science/paper/TVBFN3LN}},
note = {Machine review of arXiv:2507.10830}
}
read the original abstract
We reveal key connections between non-locality and advantage in correlation-assisted classical communication. First, using the wire-cutting technique, we provide a Bell inequality tailored to any correlation-assisted bounded classical communication task. The violation of this inequality by a quantum correlation is equivalent to its quantum-assisted advantage in the corresponding communication task. Next, we introduce wire-reading, which leverages the readability of classical messages to demonstrate advantageous assistance of non-local correlations in setups where no such advantage can be otherwise observed. Building on this, we introduce families of classical communication tasks in a Bob-without-input prepare-and-measure scenario, where non-local correlation enhances bounded classical communication while shared randomness assistance yields strictly suboptimal payoff. For the first family of tasks, assistance from any non-local facet leads to optimal payoff, while each task in the second family is tailored to a non-local facet. We reveal quantum advantage in these tasks, including qutrit over qubit entanglement advantage.
Figures
Forward citations
Cited by 1 Pith paper
-
On graph products and multi-word-representability
μ is preserved under Cartesian and rooted products, bounded under other products, and used to prove a sublinear τ(n) bound and a negative answer to the Word-representable Bipartition problem for n≥2593.
Reference graph
Works this paper leans on
-
[26]
On en- tanglement assistance to a noiseless classical channel
Péter E. Frenkel and Mihály Weiner. “On en- tanglement assistance to a noiseless classical channel”. Quantum6, 662 (2022)
work page 2022
-
[31]
Boundingthesetoffinitedimensionalquan- tum correlations
Miguel Navascués and Tamás Vértesi. “Boundingthesetoffinitedimensionalquan- tum correlations”. Phys. Rev. Lett. 115, 020501 (2015)
work page 2015
-
[1]
On the Einstein Podolsky Rosen paradox
J. S. Bell. “On the Einstein Podolsky Rosen paradox”. Physics Physique Fizika 1, 195– 200 (1964)
work page 1964
-
[2]
Nicolas Brunner, Daniel Cavalcanti, Ste- fano Pironio, Valerio Scarani, and Stephanie 14 Wehner. “Bell nonlocality”. Rev. Mod. Phys. 86, 419–478 (2014)
work page 2014
-
[3]
Random access codes and nonlocal resources
Anubhav Chaturvedi, Marcin Pawlowski, and Karol Horodecki. “Random access codes and nonlocal resources”. Phys. Rev. A 96, 022125 (2017)
work page 2017
-
[4]
Mutually unbiased balanced functions and generalized ran- dom access codes
Vaisakh M, Ram krishna Patra, Mukta Jan- pandit, Samrat Sen, Manik Banik, and Anubhav Chaturvedi. “Mutually unbiased balanced functions and generalized ran- dom access codes”. Phys. Rev. A 104, 012420 (2021)
work page 2021
-
[5]
Multiparty quan- tum communication complexity
Harry Buhrman, Wim van Dam, Peter Høyer, and Alain Tapp. “Multiparty quan- tum communication complexity”. Phys. Rev. A 60, 2737–2741 (1999)
work page 1999
-
[6]
Complementarity be- tween entanglement-assisted and quantum distributed random access code
Alley Hameedi, Debashis Saha, Piotr Mironowicz, Marcin Pawłowski, and Mo- hamed Bourennane. “Complementarity be- tween entanglement-assisted and quantum distributed random access code”. Phys. Rev. A 95, 052345 (2017)
work page 2017
Show all 36 references
-
[7]
Quantum communica- tion complexity protocol with two entangled qutrits
Časlav Brukner, Marek Żukowski, and Anton Zeilinger. “Quantum communica- tion complexity protocol with two entangled qutrits”. Phys. Rev. Lett.89, 197901 (2002)
2002
-
[8]
Quantum communication complexity protocols based on higher-dimensional entangled systems
ČASLAV BRUKNER, TOMASZ PA- TEREK, and MAREK ŻUKOWSKI. “Quantum communication complexity protocols based on higher-dimensional entangled systems”. International Journal of Quantum Information01, 519–525 (2003)
2003
-
[9]
Bell’s inequali- ties and quantum communication complex- ity
Časlav Brukner, Marek Żukowski, Jian-Wei Pan, and Anton Zeilinger. “Bell’s inequali- ties and quantum communication complex- ity”. Phys. Rev. Lett.92, 127901 (2004)
2004
-
[10]
Quantum entanglement and com- munication complexity
Harry Buhrman, Richard Cleve, and Wim van Dam. “Quantum entanglement and com- munication complexity”. SIAM Journal on Computing 30, 1829–1841 (2001)
2001
-
[11]
Substi- tuting quantum entanglement for communi- cation
Richard Cleve and Harry Buhrman. “Substi- tuting quantum entanglement for communi- cation”. Phys. Rev. A56, 1201–1204 (1997)
1997
-
[12]
Im- proving zero-error classical communication with entanglement
Toby S. Cubitt, Debbie Leung, William Matthews, and Andreas Winter. “Im- proving zero-error classical communication with entanglement”. Phys. Rev. Lett.104, 230503 (2010)
2010
-
[13]
Zero-error channel capacity and simulation assisted by non-local correlations
Toby S. Cubitt, Debbie Leung, William Matthews, and Andreas Winter. “Zero-error channel capacity and simulation assisted by non-local correlations”. IEEE Trans- actions on Information Theory 57, 5509– 5523 (2011)
2011
-
[14]
Entanglement can increase asymptotic rates of zero-error classical communication over classical channels
Debbie Leung, Laura Mancinska, William Matthews, Maris Ozols, and Aidan Roy. “Entanglement can increase asymptotic rates of zero-error classical communication over classical channels”. Communications in Mathematical Physics311, 97–111 (2012)
2012
-
[15]
Entanglement-based quantum communica- tion complexity beyond bell nonlocality
Joseph Ho, George Moreno, Samuraí Brito, Francesco Graffitti, Christopher L. Morrison, Ranieri Nery, Alexander Pick- ston, Massimiliano Proietti, Rafael Rabelo, Alessandro Fedrizzi, and Rafael Chaves. “Entanglement-based quantum communica- tion complexity beyond bell nonlocali...
2022
-
[16]
Entanglement-assisted random access codes
Marcin Pawłowski and Marek Żukowski. “Entanglement-assisted random access codes”. Phys. Rev. A81, 042326 (2010)
2010
-
[17]
Higher-dimensional communication com- plexity problems: Classical protocols versus quantum ones based on bell’s theorem or prepare-transmit-measure schemes
Armin Tavakoli and Marek Żukowski. “Higher-dimensional communication com- plexity problems: Classical protocols versus quantum ones based on bell’s theorem or prepare-transmit-measure schemes”. Phys. Rev. A95, 042305 (2017)
2017
-
[18]
Does violation of a Bell inequality always imply quantum advantage in a communication complexity problem?
Armin Tavakoli, Marek Żukowski, and Časlav Brukner. “Does violation of a Bell inequality always imply quantum advantage in a communication complexity problem?”. Quantum 4, 316 (2020)
2020
-
[19]
Correlations in entanglement-assisted prepare-and-measure scenarios
Armin Tavakoli, Jef Pauwels, Erik Wood- head, and Stefano Pironio. “Correlations in entanglement-assisted prepare-and-measure scenarios”. PRX Quantum2, 040357 (2021)
2021
-
[20]
Contextuality in entanglement-assisted one- shot classical communication
Shiv Akshar Yadavalli and Ravi Kunjwal. “Contextuality in entanglement-assisted one- shot classical communication”. Quantum6, 839 (2022)
2022
-
[21]
Entanglement in prepare-and-measure scenarios: many ques- tions, afewanswers
Jef Pauwels, Armin Tavakoli, Erik Wood- head, and Stefano Pironio. “Entanglement in prepare-and-measure scenarios: many ques- tions, afewanswers”. NewJournalofPhysics 24, 063015 (2022)
2022
-
[22]
Adaptive advantage in entanglement-assisted communications
Jef Pauwels, Stefano Pironio, Em- manuel Zambrini Cruzeiro, and Armin 15 Tavakoli. “Adaptive advantage in entanglement-assisted communications”. Phys. Rev. Lett.129, 120504 (2022)
2022
-
[23]
Communication tasks with infinite quantum-classical separa- tion
Christopher Perry, Rahul Jain, and Jonathan Oppenheim. “Communication tasks with infinite quantum-classical separa- tion”. Phys. Rev. Lett.115, 030504 (2015)
2015
-
[24]
Nonlocality- assisted enhancement of error-free commu- nication in noisy classical channels
Kunika Agarwal, Sahil Gopalkrishna Naik, Ananya Chakraborty, Samrat Sen, Pratik Ghosal, Biswajit Paul, Manik Banik, and Ram Krishna Patra. “Nonlocality- assisted enhancement of error-free commu- nication in noisy classical channels” (2025). arXiv:2412.04779
2025 arXiv
-
[25]
Advantage of hardy’s nonlocal correlation in reverse zero-error channel coding
Mir Alimuddin, Ananya Chakraborty, Govind Lal Sidhardh, Ram Krishna Patra, Samrat Sen, Snehasish Roy Chowdhury, Sahil Gopalkrishna Naik, and Manik Banik. “Advantage of hardy’s nonlocal correlation in reverse zero-error channel coding”. Phys. Rev. A108, 052430 (2023)
2023
-
[27]
Classi- cal information storage in an n-level quan- tum system
Péter E Frenkel and Mihály Weiner. “Classi- cal information storage in an n-level quan- tum system”. Communications in Mathe- matical Physics340, 563–574 (2015)
2015
-
[28]
Quantifying Bell: the Resource Theory of Nonclassicality of Common-Cause Boxes
Elie Wolfe, David Schmid, Ana Belén Sainz, Ravi Kunjwal, and Robert W. Spekkens. “Quantifying Bell: the Resource Theory of Nonclassicality of Common-Cause Boxes”. Quantum 4, 280 (2020)
2020
-
[29]
The type-independent resource theory of local operations and shared ran- domness
David Schmid, Denis Rosset, and Francesco Buscemi. “The type-independent resource theory of local operations and shared ran- domness”. Quantum4, 262 (2020)
2020
-
[30]
A convergent hierarchy of semidefinite programs characterizing the set of quantum correlations
Miguel Navascués, Stefano Pironio, and An- tonio Acín. “A convergent hierarchy of semidefinite programs characterizing the set of quantum correlations”. New Journal of Physics 10, 073013 (2008)
2008
-
[32]
Bounding the set of quan- tum correlations
Miguel Navascués, Stefano Pironio, and An- tonio Acín. “Bounding the set of quan- tum correlations”. Phys. Rev. Lett. 98, 010401 (2007)
2007
-
[33]
Post-quantum steering is a stronger- than-quantum resource for information pro- cessing
Paulo J. Cavalcanti, John H. Selby, Jamie Sikora, Thomas D. Galley, and Ana Belén Sainz. “Post-quantum steering is a stronger- than-quantum resource for information pro- cessing”. npj Quantum Information8 (2022)
2022
-
[34]
Quantum violations in the Instrumental scenario and their relations to the Bell scenario
Thomas Van Himbeeck, Jonatan Bohr Brask, Stefano Pironio, Ravis- hankar Ramanathan, Ana Belén Sainz, and Elie Wolfe. “Quantum violations in the Instrumental scenario and their relations to the Bell scenario”. Quantum3, 186 (2019)
2019
-
[35]
Bipartite postquantum steering in generalized scenar- ios
Ana Belén Sainz, Matty J. Hoban, Paul Skrzypczyk, and Leandro Aolita. “Bipartite postquantum steering in generalized scenar- ios”. Phys. Rev. Lett.125, 050404 (2020). A Bob to Alice signalling correlation lead to invalid conditional probabilities in (3) Say the correlationP ={...
2020
-
[36]
L Proof of Corollary 4 In CS[{P∗(1) NL ,P∗(2) NL},d = 3,k = 2], say the parties use correlationP∈Face{P∗(1) NL ,P∗(2) NL} as assistance to the communication channel
This payoff is higher thansΛ = 3 4 if p + 1−p 2 > 3 4 =⇒ p> 1 2. L Proof of Corollary 4 In CS[{P∗(1) NL ,P∗(2) NL},d = 3,k = 2], say the parties use correlationP∈Face{P∗(1) NL ,P∗(2) NL} as assistance to the communication channel. The following protocol for the parties gives a...
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.