{"id":"2ada5873-354f-4399-b733-314f684fcc09","arxiv_id":"1908.04182","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A cloning-based incompatibility measure is defined, proved faithful, and shown to be maximized exactly by mutually unbiased bases.","lead":"The paper defines a new measure of how incompatible quantum measurements are, based on how well their eigenstates can be cloned. It shows commuting observables have zero incompatibility, mutually unbiased bases are the most incompatible, and it finds the optimal cloning machine for two qubit measurements.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Abstract's 'iff MUBs' claim is contradicted by Theorem 5 for N>d+1; four tetrahedral qubit observables attain the bound without being MUBs.","rationale":"","tokens_in":17401,"tokens_out":31749,"duration_ms":295403,"concrete_test":"","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline result, as stated in the abstract, is that the upper bound on Q_c is attained if and only if the observables are mutually unbiased. This cannot hold for N>d+1, since at most d+1 mutually unbiased bases exist in dimension d. The paper's own Theorem 5 concedes this: for N>d+1 it asserts equality for a set of observables whose eigenstates form a unitarily invariant ensemble [20], not for MUBs. The issue is not a wording quibble. In d=2 with N=4, take four qubit observables whose Bloch axes form a regular tetrahedron. For any measurement direction r, one has sum_i (r.a_i)^2 = 4/3, so A_opt(S) = max_r [4 + sum_i (r.a_i)^2] = 16/3 and A_opt/(Nd) = (16/3)/8 = 2/3 = 2/(d+1). Substituting a = 2/3 into Eq. (16) and optimizing over parameters with p^2 > 2pq gives sup F_avg = 5/6 (the universal-cloner boundary), hence Q_c = 1/6, exactly the upper bound of Eq. (29) for N>d+1 — while no MUBs exist for N=4 in d=2. Thus the abstract's if-and-only-if statement is false as written; it requires the restriction N<=d+1. A secondary gap is that the proof of Theorem 5 for N<=d+1 cites the POVM bound of [17] for A_opt, although A_opt is a maximum over projective measurements; a simple Cauchy argument (choose B equal to one of the ensemble's own bases, giving A >= d + (N-1)*1 = N+d-1, with equality forcing every other basis to be unbiased) would supply the needed projective bound and uniqueness, but it is not included. Finally, the restriction p^2 > 2pq in Eq. (21) makes F_opt a supremum rather than an attained maximum when the optimum lies on the boundary p^2 = 2pq, so the word 'max' in Eqs. (19) and (21) is technically inaccurate.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a cloning-based incompatibility measure Q_c for a set of N observables on a d-dimensional Hilbert space, defined as Q_c(X) = 1 - F_opt(S), where S is the uniform ensemble of eigenstates and F_opt is the maximum average fidelity under a symmetric 1→2 QCM (Eqs. (19)-(23)). The authors derive an explicit formula for the average cloning fidelity (Eq. (16)), obtain the optimal cloner parameters for a general ensemble (Lemma 2, Appendix B), evaluate the measure for mutually unbiased bases (MUBs), prove an upper bound (Theorem 5), and work out the two-qubit example (Lemma 6).","tokens_in":17766,"tokens_out":5466,"duration_ms":54483,"significance":"The paper's explicit formulas in Appendix B and the two-qubit solution are careful and reproduce the known limits (MUB and commuting cases), and the faithfulness property (Lemma 1) gives a clean operational interpretation. If the proof gaps discussed below are fixed, the measure would provide a quantitative, operationally meaningful link between no-cloning and incompatibility, with MUBs singled out for N≤d+1. The connection to cloning attacks in QKD is also plausible and potentially useful. However, the headline claim as stated in the abstract is too strong and needs correction.","major_comments":[{"comment":"The abstract's statement that the upper bound is attained 'if and only if the observables are mutually unbiased' is contradicted by the paper's own Theorem 5 for N>d+1. Since at most d+1 MUBs exist in dimension d, no set with N>d+1 can be MUBs, yet Eq. (29) states the bound is attained for a unitarily invariant ensemble. A concrete counterexample in d=2, N=4 is given by four tetrahedral qubit observables, whose Bloch vectors satisfy sum_i (r.a_i)^2 = 4/3 for every direction r; this gives A_opt/(Nd)=2/3 and Q_c=1/6, saturating Eq. (29) without the observables being MUBs. The abstract, introduction, and conclusions must be amended to state the restriction N≤d+1 for the iff claim.","section":"Abstract and Theorem 5"},{"comment":"The restriction p^2>2pq is introduced with only a heuristic discussion, not a proof that the global optimum of the average cloning fidelity lies in this regime. Since Q_c is defined via this restricted maximization in Eq. (23), it is not established as the true optimal cloning fidelity of the ensemble. Moreover, the proof of the upper bound in Eq. (30) explicitly relies on the sign of (p^2-2pq). Either a proof that the restriction is without loss of generality, or a clear redefinition of Q_c as a 'restricted-cloning incompatibility measure', is needed for the operational claims to be sound.","section":"Sec. II C, Eq. (21)"},{"comment":"The equality condition 'iff MUBs' rests on the assertion that the lower bound (N+d-1)/Nd for the measurement-reconstruction fidelity is achieved only by MUBs, which is cited from [17] and used after Eq. (30). Lemma 7 only proves achievability by a projective measurement for MUBs; the uniqueness direction is not proved in this paper. Since the abstract's iff claim depends on this uniqueness, the gap is load-bearing. A simple Cauchy argument choosing the measurement basis as one of the ensemble's own bases would supply the needed projective bound and uniqueness, but it is not included.","section":"Theorem 5 proof, N≤d+1"},{"comment":"In the N>d+1 part of the proof, the lower bound A_opt(S)/Nd ≥ 2/(d+1) is imported from Fuchs-Sasaki [20], but that bound concerns the optimal fidelity over all measurements (including POVMs), while A_opt(S)/Nd is the maximum over projective measurements. The equality A_opt(S)/Nd ≡ max_B F_avg(S,M_B,A_B) from Prop. 1 does not imply that the projective restriction preserves the lower bound. This step needs a justification that the gap between the projective optimum and the POVM optimum does not violate the inequality, or the proof must use a different argument.","section":"Eq. (33)"}],"minor_comments":[{"comment":"The heading contains a typo: 'Quantifying incomptibility' should be 'Quantifying incompatibility'.","section":"Sec. II A heading"},{"comment":"The sentence 'BEfore proceeding to prove an upper bound' should read 'Before proceeding'.","section":"Sec. III A"},{"comment":"There are typos 'no-singaling principle' and 'propsed' that should be corrected.","section":"Introduction"},{"comment":"The symbol M is used in Eq. (25) and (26) but is not defined in the statement of Lemma 2; it should be set to M=Nd.","section":"Lemma 2"},{"comment":"The displayed formula for G(S,N,d) appears to be missing a division symbol between (A_opt(S)-Nd) and (Nd-2A_opt(S)); compare with Eq. (B5).","section":"Eq. (26)"},{"comment":"In the line for p(-)|b±⟩, an '=' sign is missing before the bracket expression.","section":"Eq. (40)"}],"recommendation":"major_revision","confidential_remarks":"The paper has a solid core, and the detailed calculations in Appendix B are a strength. However, the abstract's unqualified iff-MUBs claim is false for N>d+1, and the proof of Theorem 5 leans on external results from [17] and [20] in ways that need to be made explicit or replaced with self-contained arguments. The authors should be asked to correct the overclaim and to justify or rescope the restricted optimization in Eq. (21)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: the abstract overclaims. The statement that the upper bound on Q_c is attained iff the observables are MUBs is false for N>d+1, since at most d+1 MUBs exist but the paper's own Theorem 5 gives equality for a unitarily invariant ensemble. The concrete counterexample is a regular tetrahedron of four qubit observables: a direct calculation from Eq. (16) gives A_opt/(Nd)=2/3, and optimizing gives F_opt=5/6, so Q_c=1/6, exactly the upper bound in Eq. (29), with no MUBs in sight. The body is more careful than the abstract, but the abstract is the headline, and it is wrong as written.\n\nWhat is genuinely new and good: the cloning-based measure Q_c is new, and the explicit optimal cloner for a pair of qubit observables is a solid, checkable technical result. The faithfulness proof (Lemma 1) is clean, and the comparison with the measurement-and-reconstruction measure Q is useful. The paper builds on [17] in a sensible way; the reliance on that prior work is legitimate, not circular.\n\nThe soft spots are real but addressable. For N<=d+1, the tightness claim depends on the uniqueness of the ensemble attaining A_opt/(Nd)=(N+d-1)/(Nd) in the measurement-and-reconstruction problem. That uniqueness is not proven here; it is imported from [17], and the paper does not show that equality in A_opt over projective measurements forces MUBs. A short Cauchy argument, choosing the measurement basis as one of the ensemble's own bases, would likely supply the bound and the uniqueness, but it is not included. Second, the restriction p^2>2pq in Eq. (21) means F_opt is a supremum when the optimum lies on the boundary p^2=2pq. The universal cloner for N=d+1 MUBs sits exactly on that boundary, so the 'max' in Eqs. (19) and (21) is technically inaccurate. Third, there are minor typos: Eq. (36) has (d-3)/(2(d+1)) where it should be (d+3)/(2(d+1)), and Corollary 4 gives p=2/(d+1) where the unitarity condition requires p=sqrt(2/(d+1)).\n\nWho is this for? People working on operational measures of incompatibility, and anyone interested in cloning attacks in QKD. It does not open a new program; it refines an existing one. As it stands, I would not cite it without caveats. But the underlying approach is viable and the qubit result is worth having.\n\nRecommendation: send it to peer review. It needs a serious referee, and the authors should fix the abstract, prove or explicitly cite the uniqueness for the N<=d+1 case, and clarify the supremum issue. With those changes, this would be a reasonable paper.","headline":"A promising cloning-based incompatibility measure with a nice qubit cloner, but the abstract's 'iff MUBs' claim is false for N>d+1 and the proof of tightness for N<=d+1 leans on an unproven uniqueness result.","tokens_in":18385,"tokens_out":8264,"would_cite":false,"duration_ms":77353,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Mutually unbiased bases are the hardest observables to clone","keywords":["quantum cloning","incompatibility measure","mutually unbiased bases","symmetric quantum cloning machine","no-cloning theorem","eigenstate ensemble","qubit observables","quantum key distribution"],"falsifier":"Search all sets of $N=3$ observables in $d=4$ for one whose eigenstate ensemble gives $A_{opt}(S)=N+d-1=6$ but is not mutually unbiased; if such a set exists, its optimal cloning fidelity matches the MUB value and the 'if and only if' fails. Alternatively, optimize two-qubit cloning beyond the paper's $p^2>2pq$ restriction for a pair with Bloch overlap $0.5$; a higher fidelity would show $Q_c$ is not the true optimal cloning fidelity.","tokens_in":17128,"feed_emoji":"⚛️","tokens_out":15205,"duration_ms":151945,"temperature":0.7,"pith_summary":"This paper proposes that the mutual incompatibility of quantum observables can be measured by how badly their eigenstates can be cloned. For a set of observables, take the uniform mixture of all their eigenstates and ask for the highest average fidelity a symmetric 1-to-2 quantum cloning machine can achieve; the measure $Q_c$ is one minus that fidelity, so larger incompatibility means lower cloning fidelity. The paper proves $Q_c$ is faithful: it is zero exactly when the observables commute. Its main result is a tight upper bound: for sets of at most $d+1$ observables on a $d$-dimensional system, $Q_c$ is maximized only by mutually unbiased bases, the bases in which every vector of one basis has equal overlap with every vector of the other. This gives the no-cloning principle a quantitative role in identifying which measurements are maximally incompatible.","feed_headline":"Mutually unbiased bases resist cloning the hardest","feed_subtitle":"A new measure links no-cloning to measurement incompatibility; mutually unbiased bases set the maximum.","key_machinery":"The central object is the symmetric 1-to-2 quantum cloning machine, a unitary transformation parameterized by $p$ and $q$ with $p^2 + 2(d-1)q^2 = 1$, which sends a basis state to a weighted sum of a copied term and symmetrized cross terms. For an input ensemble $S$, the average clone fidelity depends on the cloning basis $B$ only through $A(S,B)$, the sum of participation ratios of the ensemble states in the basis; maximizing over $B$ is equivalent to choosing the best projective measurement-and-reconstruction strategy. The upper bound is proved by comparing $A(S,B)$ with its mutually-unbiased-bases value $A_{opt}(S_{MUB}) = N+d-1$, using an imported uniqueness result, and for $N>d+1$ by falling back on the universal-cloner fidelity bound.","core_discovery":"For a set $X$ of $N$ observables on a $d$-dimensional Hilbert space, let $S$ be the uniform ensemble of their eigenstates and let $F_{opt}(S)$ be the maximum average fidelity achieved by a symmetric 1-to-2 quantum cloning machine, optimized over cloning basis and parameters in the regime $p^2 > 2pq$. The paper defines $Q_c(X) = 1 - F_{opt}(S)$ and claims three things: it is faithful, vanishing only for commuting observables; for sets of at most $d+1$ observables it is bounded above by the value for $N$ mutually unbiased bases, with equality only for mutually unbiased sets; and for more than $d+1$ observables it is bounded by the universal-cloner fidelity. It also gives an explicit optimal cloner for a pair of qubit observables, with optimal cloning basis along the sum and difference of their Bloch vectors, and shows that a complete set of $d+1$ mutually unbiased bases is cloned optimally by the universal symmetric cloner.","pith_inferences":["The paper does not discuss generalized measurements, but if its logic carries over, a cloning-based measure for POVMs should vanish exactly when the effects are jointly commuting; a natural test is whether the upper bound is again attained by mutually unbiased bases.","The explicit two-qubit solution predicts that the optimal cloner depends only on the absolute overlap between the two Bloch vectors, so a tabletop experiment with photon polarization qubits could map that fidelity curve and check the formula.","Because the bound for more than $d+1$ observables is attained by a unitarily invariant ensemble, maximal cloning-based incompatibility may occur in every dimension even when a complete set of mutually unbiased bases does not exist, extending the paper's MUB result beyond the usual MUB dimensions."],"forward_implications":["A complete set of $d+1$ mutually unbiased bases is cloned optimally by the universal symmetric cloner, so the cloning fidelity for such an ensemble is the universal value $(d+3)/(2(d+1))$.","No set of at most $d+1$ observables can be more incompatible than the same number of mutually unbiased bases, so wherever such bases exist they are the maximally incompatible measurements.","For more than $d+1$ observables, the cloning-based bound saturates at the universal-cloner level, meaning that adding extra observables beyond $d+1$ cannot increase the incompatibility as measured here.","Because the optimal cloning fidelity is always at least the best measurement-and-reconstruction fidelity for mutually unbiased ensembles, a cloning attack outperforms an intercept-and-resend attack on such signal states, a fact relevant to quantum key distribution."],"supporting_citations":[{"why":"Supplies the measurement-and-reconstruction incompatibility measure Q, its lower bound F_opt >= (N+d-1)/(Nd), and the uniqueness of MUB attainment that the upper-bound proof uses.","marker":"[17]"},{"why":"Defines the symmetric 1-to-2 quantum cloning machine transformation, the unitarity condition p^2 + 2(d-1)q^2 = 1, and the universal cloner parameters used throughout.","marker":"[7]"},{"why":"Provides the general lower bound F_opt(S) >= 2/(d+1) and the unitarily invariant ensemble that saturates the bound for N > d+1.","marker":"[20]"},{"why":"The no-cloning theorem: perfect cloning is possible only for orthogonal states, backing the faithfulness result in Lemma 1.","marker":"[1]"},{"why":"The no-broadcasting theorem for non-orthogonal ensembles, used in the converse direction of Lemma 1.","marker":"[2]"},{"why":"Establishes optimality of the universal symmetric cloner in any dimension, used to identify the optimal cloner for a complete set of mutually unbiased bases.","marker":"[9]"},{"why":"Shows an optimal symmetric cloner can be formed from optimal asymmetric cloners, justifying the restriction to symmetric cloning machines.","marker":"[19]"}],"fun_headline_variants":["Cloning fidelity quantifies measurement incompatibility","MUBs are the hardest to clone, new measure shows","No-cloning principle linked to incompatible observables","Optimal cloning bound reached by mutually unbiased bases","A cloning-based measure of quantum incompatibility"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 'if and only if' statement rests on an imported uniqueness result: only a mutually unbiased ensemble attains the minimal measurement-and-reconstruction fidelity $(N+d-1)/(Nd)$ among sets of $N$ observables; if another ensemble also attains that minimum, the upper bound could be reached without MUBs.","fun_headline_variants_meta":{"raw":{"variants":["Cloning fidelity quantifies measurement incompatibility","MUBs are the hardest to clone, new measure shows","No-cloning principle linked to incompatible observables","Optimal cloning bound reached by mutually unbiased bases","A cloning-based measure of quantum incompatibility"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000325,"raw_usage":{"total_tokens":1810,"prompt_tokens":919,"completion_tokens":891,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":535,"completion_tokens_details":{"reasoning_tokens":819}},"tokens_in":535,"tokens_out":891,"duration_ms":9995,"temperature":1.0,"reasoning_tokens":819,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:51:07.582136+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search all sets of $N=3$ observables in $d=4$ for one whose eigenstate ensemble gives $A_{opt}(S)=N+d-1=6$ but is not mutually unbiased; if such a set exists, its optimal cloning fidelity matches the MUB value and the 'if and only if' fails. Alternatively, optimize two-qubit cloning beyond the paper's $p^2>2pq$ restriction for a pair with Bloch overlap $0.5$; a higher fidelity would show $Q_c$ is not the true optimal cloning fidelity.","supporting_citations":[{"cited_title":"Bandyopadhyay, and P","cited_arxiv_id":null,"evidence_quote":"Supplies the measurement-and-reconstruction incompatibility measure Q, its lower bound F_opt >= (N+d-1)/(Nd), and the uniqueness of MUB attainment that the upper-bound proof uses."},{"cited_title":"Buz’ek & M","cited_arxiv_id":null,"evidence_quote":"Defines the symmetric 1-to-2 quantum cloning machine transformation, the unitarity condition p^2 + 2(d-1)q^2 = 1, and the universal cloner parameters used throughout."},{"cited_title":"Fuchs, and M","cited_arxiv_id":null,"evidence_quote":"Provides the general lower bound F_opt(S) >= 2/(d+1) and the unitarily invariant ensemble that saturates the bound for N > d+1."},{"cited_title":"Wootters and W.H","cited_arxiv_id":null,"evidence_quote":"The no-cloning theorem: perfect cloning is possible only for orthogonal states, backing the faithfulness result in Lemma 1."},{"cited_title":"Barnum et al","cited_arxiv_id":null,"evidence_quote":"The no-broadcasting theorem for non-orthogonal ensembles, used in the converse direction of Lemma 1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows an optimal symmetric cloner can be formed from optimal asymmetric cloners, justifying the restriction to symmetric cloning machines."}],"review_version":1}