REVIEW 3 minor 60 references
Weight of informativeness, state exclusion games and excludible information
T0 review · 0 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read The weight of informativeness of a measurement equals its optimal advantage in state exclusion games.
desk verdict This paper proves the weight-exclusion dual of the robustness-discrimination result and is sound enough to referee. 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
The central object is the weight of informativeness, $\mathrm{WoI}(M)=\min\{w : M_a = wN_a + (1-w)q(a)\mathbb{1}\}$ with $N$ a POVM and $q$ a probability distribution, which has the closed form $\mathrm{WoI}(M)=1-\sum_a \lambda_{\min}(M_a)$. Its dual semidefinite programme yields the extremal ensemble that saturates the bound in Result 1. The second engine is the quantum-to-classical channel $\Lambda_M$ and the conditional exclusion entropy $H_{-\infty}(X|G)$, which is manipulated into $-\log P^{\mathrm{Q}}_{\mathrm{err}}(\mathcal{E}, M)$ by optimising over classical post-processings. The simulability order $N_x=\sum_a p(x|a)M_a$ connects the game's minimisation to the resource order, and minimax reasoning is used to prove the completeness result.
What would settle it
Take a POVM $M$, compute $\mathrm{WoI}(M)=1-\sum_a \lambda_{\min}(M_a)$, and obtain the optimal ensemble from the dual SDP in Appendix B; if the ratio $P^{\mathrm{Q}}_{\mathrm{err}}(\mathcal{E}^*,M)/P^{\mathrm{C}}_{\mathrm{err}}(\mathcal{E}^*)$ is not exactly $1-\mathrm{WoI}(M)$, Result 1 is false. Similarly, one can compute the excludible information of $\Lambda_M$ directly from the minimisation in Appendix C and compare it with $-\log[1-\mathrm{WoI}(M)]$.
Extended reading notes
Core claim
Result 1 is the exact identity $\min_{\mathcal{E}} P^{\mathrm{Q}}_{\mathrm{err}}(\mathcal{E}, M)/P^{\mathrm{C}}_{\mathrm{err}}(\mathcal{E}) = 1 - \mathrm{WoI}(M)$, where the left side is the smallest ratio, over all ensembles, between the best exclusion-error probability achievable using the measurement $M$ and the best classical error probability, and the right side is a number computed from $M$ alone. Result 2 adds that the single-shot excludible information of the quantum-to-classical channel $\Lambda_M(\rho)=\sum_a |a\rangle\langle a| \operatorname{Tr}(M_a\rho)$ equals $-\log[1-\mathrm{WoI}(M)]$. Result 3 states that $M$ can simulate $N$ if and only if $P^{\mathrm{Q}}_{\mathrm{err}}(\mathcal{E}, M) \le P^{\mathrm{Q}}_{\mathrm{err}}(\mathcal{E}, N)$ for every ensemble $\mathcal{E}$. The paper's claim is that these form one coherent correspondence: a weight-based resource measure, an exclusion task, and a single-shot information quantity are the same object from three viewpoints.
Load-bearing premise
The proof assumes that the best a player can do with a fixed measurement in an exclusion game is to classically process its outcomes, and that the relevant optimisation problems can be swapped in the standard convex way; if some strategy outside this description performed better, the exact equality with the weight of informativeness would fail.
Editorial extensions
If this is right
- For any measurement and any state exclusion game, the measurement can reduce the classical error probability by at most a factor $1-\mathrm{WoI}(M)$, and there is a game where this bound is attained.
- The single-shot excludible information of the channel induced by a measurement is exactly $-\log[1-\mathrm{WoI}(M)]$, so a purely geometric quantifier computes a communication-theoretic quantity for exclusion tasks.
- A measurement $M$ can simulate a measurement $N$ exactly when $M$ is never worse than $N$ in any state exclusion game, giving a second complete set of monotones for measurement simulation.
- The weight of informativeness can be evaluated efficiently: it is the solution of an SDP and has the explicit eigenvalue formula $1-\sum_a \lambda_{\min}(M_a)$.
Reading between the lines
- The authors conjecture that the weight-exclusion correspondence is generic across quantum resource theories; if that holds, weight-based quantifiers in other resource theories would acquire operational interpretations as optimal advantages in exclusion tasks rather than remaining purely geometric measures.
- Because $\mathrm{WoI}$ has a closed form, Result 1 turns an optimisation over all ensembles into an eigenvalue computation, which could make exclusion advantages easy to estimate for large or noisy POVMs.
- The if-and-only-if in Result 3 implies that any failure of simulation is witnessed by some exclusion game, but the proof does not identify a finite set of games that would certify simulation; finding such a finite witnessing set would turn the criterion into a practical test.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper introduces the weight of informativeness (WoI) as a resource quantifier for the convex quantum resource theory of measurement informativeness, proves its basic properties, and establishes two main operational results. First, for any measurement M, the optimal advantage over classical strategies in state exclusion games, quantified by the ratio of quantum to classical error probabilities minimized over all ensembles, equals 1 - WoI(M). Second, the single-shot excludible information of the quantum-to-classical channel associated with M is -log[1 - WoI(M)], extending the known robustness-based three-way correspondence to weight-based quantifiers and exclusion tasks. The paper also proves that error probabilities in state exclusion games form a complete set of monotones for measurement simulation, giving a necessary and sufficient condition for M to simulate N in terms of these probabilities. The central proofs are analytic, using an explicit dual SDP for the WoI, strong duality, and a minimax argument.
Significance. If the results hold, the paper makes a significant conceptual contribution to quantum resource theories: it provides the first operational interpretation of a weight-based quantifier, establishes a parallel three-way correspondence between resource quantifiers, operational tasks, and single-shot information-theoretic quantities, and identifies a second complete set of monotones for measurement simulation. The proofs are mathematically sound and self-contained: the dual SDP in Appendix B is explicit, the reduction in Appendix C is valid for quantum-to-classical channels, and the monotonicity proof in Appendix D uses a correct minimax step. There is no circularity: the state exclusion game and excludible information are defined independently of WoI, and the equalities are proven rather than assumed. The conjecture that the weight-exclusion correspondence holds for arbitrary resource theories is clearly stated and appropriately supported by a forthcoming companion result.
minor comments (3)
- [Section II, Lemma (iii)] The monotonicity statement in the main text reads "N≼ M→ WoI(N)≤ WoI(N)", which is a typo: the second WoI should be evaluated on M. Appendix A states the correct inequality WoI(M') ≤ WoI(M). Please correct the main-text statement.
- [Appendix C, proof of Result 2] The proof silently restricts the decoding POVM to the computational basis by setting D_g = |g><g|, then optimizes over classical post-processings p(x|g). This reduction is valid for a quantum-to-classical channel because any decoding POVM D_g enters only through its diagonal entries <a|D_g|a>, which form a conditional probability distribution, but this justification is not stated. Since Eq. (C6) and hence Eq. (8) rely on this reduction, please add a brief explanation. Also, the Kronecker delta immediately after "Choosing D_g = |g><g|" should read δ_{a,g}, not δ_a^x.
- [Appendix D, sufficiency proof] In the sufficiency proof, the argument treats M' as a k-outcome measurement in Eq. (D3), but M' has l' outcomes. The assumption (9) holds for all ensembles, so one can take k = l' (or pad/coarse-grain accordingly) to make the comparison well-defined. The text should state this explicitly so that the subsequent minimax step is unambiguous.
Circularity Check
No significant circularity: WoI, exclusion games, and excludible information are defined independently, and the equalities are proven via SDP duality and minimax rather than assumed.
full rationale
Result 1 is proven directly from Definitions 2 and 3: the lower bound uses the defining decomposition of WoI, and the upper bound uses the dual SDP of the explicit form WoI = 1 - sum_a lambda_min(M_a), constructing the adversarial ensemble from the dual variables. Neither direction assumes the equality being proved; the state exclusion game and the weight are defined independently. Result 2 follows from Result 1 after reducing the decoding optimization for the c-q channel Lambda_M to post-processing of M, which is valid because any decoding POVM on the classical output only enters through its diagonal entries, which form a conditional probability distribution; no fitted parameter is renamed as a prediction. Result 3 is a direct minimax argument over the simulability polytope, with no importation of a uniqueness theorem. The only author-overlap citation is [23] (Skrzypczyk-Linden), used for context, notation, and the parallel robustness/discrimination/accessible-information correspondence; none of the proofs in this work invoke [23] as a premise. The unstated decoding reduction in Appendix C and the outcome-count handling in Appendix D are presentation gaps, not circular steps. I therefore find no significant circularity.
Assumptions & free parameters
assumptions (4)
- domain assumption Uninformative measurements {q(a)1} form the free set, and classical strategies in the exclusion game are exactly uninformative measurements followed by classical post-processing.
- standard math Strong duality holds for the SDP (A14) and its dual (B6).
- domain assumption For quantum-to-classical channels, arbitrary decoding POVMs can be reduced to computational-basis measurement plus classical post-processing.
- standard math The minimax theorem applies to the exchange in Eq. (D6).
Cite this review
Pith. "Pith review of Weight of informativeness, state exclusion games and excludible information." pith.science (2026). https://pith.science/paper/VBTNZW7X
@misc{pith2026190810347,
author = {Pith},
title = {Pith review of: Weight of informativeness, state exclusion games and excludible information},
year = {2026},
howpublished = {\url{https://pith.science/paper/VBTNZW7X}},
note = {Machine review of arXiv:1908.10347}
}
read the original abstract
We consider the quantum resource theory of measurement informativeness and introduce a weight-based quantifier of informativeness. We show that this quantifier has operational significance from the perspective of quantum state exclusion, by showing that it precisely captures the advantage a measurement provides in minimising the error in this game. We furthermore introduce information theoretic quantities related to exclusion, in particular the notion of excludible information of a quantum channel, and show that for the case of quantum-to-classical channels it is determined precisely by the weight of informativeness. This establishes a three-way correspondence which sits in parallel to the recently discovered correspondence in quantum resource theories between robustness-based quantifiers, discrimination games, and accessible information. We conjecture that the new correspondence between a weight-based quantifier and an exclusion-based task found in this work is a generic correspondence that holds in the context of quantum resource theories.
Figures
Reference graph
Works this paper leans on
-
[23]
Skrzypczyk and N
P. Skrzypczyk and N. Linden, Phys. Rev. Lett. 122, 140403 (2019)
2019
-
[1]
Chitambar and G
E. Chitambar and G. Gour, Rev. Mod. Phys. 91, 025001 (2019)
2019
-
[2]
C. H. Bennett, G. Brassard, S. Popescu, B. Schumacher, J. A. Smolin, and W. K. Wootters, Phys. Rev. Lett. 76, 722 (1996)
1996
-
[3]
C. H. Bennett, H. J. Bernstein, S. Popescu, and B. Schu- macher, Phys. Rev. A 53, 2046 (1996)
1996
-
[4]
M. Oszmaniec, L. Guerini, P. Wittek, and A. Ac´ ın, Phys. Rev. Lett. 119, 190501 (2017)
work page 2017
-
[5]
Oszmaniec and T
M. Oszmaniec and T. Biswas, Quantum 3, 133 (2019)
2019
-
[6]
A resource theory of quantum measure- ments,
T. Guff, N. A. McMahon, Y. R. Sanders, and A. Gilchrist, “A resource theory of quantum measure- ments,” (2019), arXiv:1902.08490
arXiv 2019
-
[7]
J. I. de Vicente, Journal of Physics A: Mathematical and Theoretical 47, 424017 (2014)
2014
Show all 60 references
-
[8]
Duarte and B
C. Duarte and B. Amaral, Journal of Mathematical Physics 59, 062202 (2018)
2018
-
[9]
Resource theory of contextuality,
B. Amaral, “Resource theory of contextuality,” (2019), arXiv:1904.04182
2019 arXiv
-
[10]
Gallego and L
R. Gallego and L. Aolita, Phys. Rev. X 5, 041008 (2015)
2015
-
[11]
Resource theories of quan- tum channels and the universal role of resource erasure,
Z.-W. Liu and A. Winter, “Resource theories of quan- tum channels and the universal role of resource erasure,” (2019), arXiv:1904.04201
2019 arXiv
-
[12]
Operational resource theory of quantum channels,
Y. Liu and X. Yuan, “Operational resource theory of quantum channels,” (2019), arXiv:1904.02680
2019 arXiv
-
[13]
Baumgratz, M
T. Baumgratz, M. Cramer, and M. B. Plenio, Phys. Rev. Lett. 113, 140401 (2014)
2014
-
[14]
Winter and D
A. Winter and D. Yang, Phys. Rev. Lett. 116, 120404 (2016)
2016
-
[15]
F. G. S. L. Brand˜ ao, M. Horodecki, J. Oppenheim, J. M. Renes, and R. W. Spekkens, Phys. Rev. Lett. 111, 250404 (2013)
2013
-
[16]
Theurer, N
T. Theurer, N. Killoran, D. Egloff, and M. B. Plenio, Phys. Rev. Lett. 119, 230401 (2017)
2017
-
[17]
Streltsov, H
A. Streltsov, H. Kampermann, S. Wlk, M. Gessner, and D. Bruß, New Journal of Physics 20, 053058 (2018)
2018
-
[18]
Takagi and Q
R. Takagi and Q. Zhuang, Phys. Rev. A 97, 062337 (2018)
2018
-
[19]
Convex resource theory of non-markovianity,
S. Bhattacharya, B. Bhattacharya, and A. S. Majumdar, “Convex resource theory of non-markovianity,” (2018), arXiv:1803.06881
2018 arXiv
-
[20]
Operational resource theory of non- markovianity,
E. Wakakuwa, “Operational resource theory of non- markovianity,” (2017), arXiv:1709.07248
2017 arXiv
-
[21]
Gour and R
G. Gour and R. W. Spekkens, New Journal of Physics 10, 033023 (2008)
2008
-
[24]
Vidal and R
G. Vidal and R. Tarrach, Phys. Rev. A 59, 141 (1999)
1999
-
[25]
Steiner, Phys
M. Steiner, Phys. Rev. A 67, 054305 (2003)
2003
-
[26]
Piani and J
M. Piani and J. Watrous, Phys. Rev. Lett. 114, 060404 (2015)
2015
-
[27]
Piani, M
M. Piani, M. Cianciaruso, T. R. Bromley, C. Napoli, N. Johnston, and G. Adesso, Phys. Rev. A 93, 042107 (2016)
2016
-
[28]
Napoli, T
C. Napoli, T. R. Bromley, M. Cianciaruso, M. Piani, N. Johnston, and G. Adesso, Phys. Rev. Lett. 116, 150502 (2016)
2016
-
[29]
ˇSupi´ c, P
I. ˇSupi´ c, P. Skrzypczyk, and D. Cavalcanti, Phys. Rev. A 99, 032334 (2019)
2019
-
[30]
The opera- tional advantages provided by non-classical teleporta- tion,
P. Lipka-Bartosik and P. Skrzypczyk, “The opera- tional advantages provided by non-classical teleporta- tion,” (2019), arXiv:1908.05107
2019 arXiv
-
[31]
Howard and E
M. Howard and E. Campbell, Phys. Rev. Lett. 118, 090501 (2017)
2017
-
[32]
A. C. Elitzur, S. Popescu, and D. Rohrlich, Physics Let- ters A 162, 25 (1992)
1992
-
[33]
This quantifier has several different names such as: part, content, cost and weight
in the context of entanglement. This quantifier has several different names such as: part, content, cost and weight. In order to keep consistency with recent notation in the literature, we adopt weight in this work. Definition 3: (Weight of informativeness) The weight of informat...
-
[34]
Skrzypczyk, M
P. Skrzypczyk, M. Navascu´ es, and D. Cavalcanti, Phys. Rev. Lett. 112, 180404 (2014)
2014
-
[35]
M. F. Pusey, J. Opt. Soc. Am. B 32, A56 (2015)
2015
-
[36]
Lewenstein and A
M. Lewenstein and A. Sanpera, Phys. Rev. Lett.80, 2261 (1998)
1998
-
[37]
K. Bu, N. Anand, and U. Singh, Phys. Rev. A 97, 032342 (2018)
2018
-
[38]
Takagi, B
R. Takagi, B. Regula, K. Bu, Z.-W. Liu, and G. Adesso, Phys. Rev. Lett. 122, 140402 (2019)
2019
-
[39]
Cavalcanti and P
D. Cavalcanti and P. Skrzypczyk, Phys. Rev. A 93, 052112 (2016)
2016
-
[40]
Skrzypczyk, I
P. Skrzypczyk, I. ˇSupi´ c, and D. Cavalcanti, Phys. Rev. Lett. 122, 130403 (2019)
2019
-
[41]
Carmeli, T
C. Carmeli, T. Heinosaari, and A. Toigo, Phys. Rev. Lett. 122, 130402 (2019)
2019
-
[42]
General resource theo- ries in quantum mechanics and beyond: operational characterization via discrimination tasks,
R. Takagi and B. Regula, “General resource theo- ries in quantum mechanics and beyond: operational characterization via discrimination tasks,” (2019), arXiv:1901.08127
2019 arXiv
-
[43]
Bandyopadhyay, R
S. Bandyopadhyay, R. Jain, J. Oppenheim, and C. Perry, Phys. Rev. A 89, 022336 (2014)
2014
-
[44]
Guerini, J
L. Guerini, J. Bavaresco, M. T. Cunha, and A. Ac´ ın, Journal of Mathematical Physics 58, 092102 (2017)
2017
-
[45]
Operational characterization of incompatibil- ity of quantum channels with quantum state discrimina- tion,
J. Mori, “Operational characterization of incompatibil- ity of quantum channels with quantum state discrimina- tion,” (2019), arXiv:1906.09859
2019 arXiv
-
[46]
M. F. Pusey, J. Barrett, and T. Rudolph, Nature Physics 8, 475 (2012). 6
2012
-
[47]
or not-Post-Peierls compatibility (Post-Peierls in- compatibility) [48, 49]. We adopt an operational ap- proach here, so that this property guarantees that the game of state exclusion is won with probability one, or conclusive (or perfect) state exclusion [43, 50]. The game of...
-
[48]
Boyd and L
S. Boyd and L. Vandenberghe, Convex Optimization (Cambridge University Press, New York, NY, USA, 2004)
2004
-
[49]
C. M. Caves, C. A. Fuchs, and R. Schack, Phys. Rev. A 66, 062111 (2002)
2002
-
[50]
Heinosaari and O
T. Heinosaari and O. Kerppo, Journal of Physics A: Mathematical and Theoretical 51, 365303 (2018)
2018
-
[51]
A. F. Ducuara, C. E. Susa, and J. H. Reina, Journal of Physics A: Mathematical and Theoretical 50, 265302 (2017)
2017
-
[52]
Z.-W. Liu, C. Perry, Y. Zhu, D. E. Koh, and S. Aaronson, Phys. Rev. A 93, 012347 (2016)
2016
-
[53]
Molina, Quantum 3, 117 (2019)
A. Molina, Quantum 3, 117 (2019)
2019
-
[54]
Perry, R
C. Perry, R. Jain, and J. Oppenheim, Phys. Rev. Lett. 115, 030504 (2015)
2015
-
[56]
In preparation,
A. F. Ducuara and P. Skrzypczyk, “In preparation,” (2019). APPENDICES Appendix A: Lemma Lemma: (Properties of WoI) The weight of informa- tiveness (2) satisfies the following properties: (i) Faithfulness: WoI(M) = 0↔ M ={Ma =q(a)1}. (A1) (ii) Convexity: given two measurements M...
2019
-
[57]
We now consider the quantities for p∈ [0, 1]: pM1a + (1−p)M2a = =p [ (1−w∗ 1)q∗ 1(a)1 +w∗ 1N∗ 1a ] + + (1−p) [ (1−w∗ 2)q∗ 2(a)1 +w∗ 2N∗ 2a ]
satisfying: M1a = (1−w∗ 1)q∗ 1(a)1 +w∗ 1N∗ 1a, M2a = (1−w∗ 2)q∗ 2(a)1 +w∗ 2N∗ 2a. We now consider the quantities for p∈ [0, 1]: pM1a + (1−p)M2a = =p [ (1−w∗ 1)q∗ 1(a)1 +w∗ 1N∗ 1a ] + + (1−p) [ (1−w∗ 2)q∗ 2(a)1 +w∗ 2N∗ 2a ] . (A7) We now define the variables: ˜w =pw∗ 1 + (1−p)w∗...
-
[58]
(B1) Let us start with the weight of informativeness of a mea- surement as given by (2)
First part In this first part we prove that: [1− WoI(M)]P C err(E)≤P Q err(E, M), ∀E, M. (B1) Let us start with the weight of informativeness of a mea- surement as given by (2). Consider that the minimum is achieved with the triple (q∗, N∗,w∗) so that∀a: Ma− (1−w∗)q∗(a)1 =w∗Na≥...
-
[59]
(B3) 8 This will be done by considering the dual formulation of the primal SDP for the weight of informativeness [45]
Second part In this second part we prove that ∀M,∃E M such that: [1− WoI(M)]P C err ( E M) ≥P Q err(E M, M), ∀M. (B3) 8 This will be done by considering the dual formulation of the primal SDP for the weight of informativeness [45]. a. Deriving the dual SDP We start by addressi...
-
[60]
Necessary condition Let us address the necessary condition: M≽ M′ =⇒ P Q err(E, M)≤P Q err(E, M′) ∀E. (D1) Let us consider the probability of error in state exclusion: P Q err(E, M′) = min M′≽N′ k∑ x=1 Tr (N′ x ˜ρx), = min {p(x|b)} k∑ x=1 Tr l′ ∑ b=1 p(x|b)M′ b ˜ρx, = min {p(x...
-
[61]
(D2) 10 Let us start by assuming that the right-hand side is true
Sufficient condition We now address the sufficient condition: M≽ M′ ⇐= P Q err(E, M)≤P Q err(E, M′) ∀E. (D2) 10 Let us start by assuming that the right-hand side is true. We now want to prove that M≽ M′ which is equivalent to∑ aq(x|a)Ma = M′ x. Let us continue by considering the i...
-
[2016]
acknowledges support from a Royal Society URF (UHQT)
P.S. acknowledges support from a Royal Society URF (UHQT)
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.