{"id":"45eafd55-9e06-4588-89ec-ccff101f8ecb","arxiv_id":"1908.04093","paper_version":2,"verdict":"REJECT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"A certified-answer discrimination scheme for ordered quantum states is cast as an SDP, but the proposed analytical lower bound rests on an invalid feasibility argument.","lead":"This paper introduces a quantum state discrimination protocol for ordered hypotheses where any wrong answer is guaranteed to be within a chosen distance of the true position, and it formulates the optimal measurement as a semidefinite program. It also claims an analytical lower bound on the success probability and applies both to quantum change point and anomaly detection.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section III's lower-bound ansatz violates the positivity constraint it is derived from: for rank-one ME blocks Eq. (7) forces the central diagonal to zero, so Eq. (11) is not a valid lower bound.","rationale":"I read the paper as making two independent contributions: the efficient SDP (5) for certified-answer discrimination, and an analytical lower bound (11) requiring only the minimum-error solution. The SDP construction in Section II appears coherent. The central advertised result, however, is the lower bound, and the derivation in Section III has a genuine gap: a necessary principal-minor condition is relaxed to Eq. (8), then saturating the relaxed inequality is asserted to give a feasible ansatz for SDP (6). No off-diagonal completion is constructed, and the necessary condition is insufficient. The rank-one structure used in the QCP application makes the failure concrete: the determinant condition (7) forces the central diagonal to be exactly zero whenever the corresponding off-diagonal element of ZME is nonzero, so the positive bound from Eq. (9) cannot be achieved. This is an internal inconsistency in the proof, not a disagreement with the surrounding consensus. The QSAD symmetry argument and numerical SDP results may survive, but they do not repair the claimed general theorem. The reader's verdict of REJECT is therefore appropriate, and my stress-test does not motivate changing it.","tokens_in":11426,"tokens_out":6699,"duration_ms":68416,"concrete_test":"Run a small feasibility SDP for the QCP with n=3, c=0.6, Delta=1: form G_{ij}=c^{|i-j|}, compute S=sqrt(G), set ZME_i=|s_i>= |x|^2. For a dyad block |s_i>= ab, hence z <= 0. Since z >= 0, this forces z = 0. Thus a nonzero positive central element, including the value prescribed by Eq. (9), violates the very principal minor used to derive Eq. (8). The logical error is that Eq. (8) is only a necessary consequence of (7) obtained via AM-GM; saturating (8) does not guarantee the determinant condition (7). No off-diagonal completion is supplied, and in the rank-one case no completion can help because the violation already appears in a principal minor involving only the diagonal entry z. Consequently the general lower bound (11) and its QCP specialization (21) are unsupported. The SDP formulation (5) and the numerical results may still stand, but the analytical lower-bound theorem is the casualty.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces 'certified answer discrimination' for ordered sets of linearly independent pure states, in which a POVM must never return an answer more than Delta sites away from the true hypothesis. It gives a block-structured SDP, Eq. (5), and claims a general lower bound on the Delta-scheme success probability that depends only on the minimum-error solution, Eq. (11). The framework is then specialized to the quantum change point and quantum state anomaly detection problems, with numerical comparisons and an analytic bound for the change point, Eq. (21).","tokens_in":11686,"tokens_out":9664,"duration_ms":96767,"significance":"If the lower-bound theorem were valid, the paper would provide a useful interpolation between unambiguous and minimum-error discrimination for ordered hypotheses, and the QCP example would quantify how much success probability is lost by enforcing certified answers. The SDP reformulation in Section II is a coherent and potentially useful contribution, and the numerical demonstrations in Section IV illustrate the behavior of the exact SDP. However, the central analytical result of Section III is not established: the proposed ansatz is infeasible in exactly the rank-one case used in the applications. The advertised general bound and its QCP specialization are therefore unsupported.","major_comments":[{"comment":"The proposed lower-bound ansatz is not a feasible point of SDP (6). Feasibility of Z requires Z_ME - Z_Delta >= 0, so every 2-by-2 principal minor must be positive semidefinite. For a rank-one block Z_ME_i = |s_i>= |x|^2. Since |x|^2 = ab for the dyad block, this forces z <= 0, and with z >= 0 it forces z = 0. Thus any nonzero central diagonal element, including the value prescribed by Eq. (9), violates the very positivity condition from which Eq. (8) was derived. Eq. (8) is only a necessary consequence of Eq. (7) obtained through AM-GM; saturating it does not guarantee the determinant condition (7), and no off-diagonal completion can repair the violation because the violation occurs in a principal minor whose only Z_Delta entry is the diagonal element z. Consequently Eq. (11), and its QCP specialization Eq. (21), are unsupported.","section":"Section III, Eqs. (8)-(11)"},{"comment":"Even if the lower-bound theorem were valid for the exact ME solution, the QCP application replaces Z_ME by the square-root measurement of Ref. [13] without proving a one-sided bound. The text says that the ME solution 'can be very well approximated' by the square-root measurement, but a certified lower bound requires a controlled direction of approximation, for example Z_ME - Z_SRM >= 0 or a norm estimate with a definite sign. Without such a statement, Eq. (21) is at best an estimate for the square-root measurement rather than a certified lower bound on P_s^Delta.","section":"Section IV, Eqs. (13) and (21)"}],"minor_comments":[{"comment":"The phrase 'linear independent estates' should read 'linearly independent states'.","section":"Section II"},{"comment":"The sentence 'we can have have a one-site error' contains a duplicated word.","section":"Introduction"},{"comment":"The bound is written as |mu_hat(r,c)| <= M0(c) e^{k log c}; the index r on the left and k in the exponent should be the same symbol.","section":"Appendix A, Eqs. (18) and (A1)"},{"comment":"The claim that for Delta < floor(n/2) the protocol is equivalent to unambiguous discrimination is asserted rather than proved; a derivation of this symmetry reduction should be supplied, since Eq. (30) relies on it.","section":"Section IV, QSAD"},{"comment":"The contour argument should specify the branch cuts of the square root in mu(z,c) and justify taking the limit epsilon -> 0 inside the integral; the proof as written is too terse on these points.","section":"Appendix A"}],"recommendation":"reject","confidential_remarks":"The central lower-bound theorem is invalid because the diagonal ansatz of Section III is infeasible for the rank-one blocks used in the applications. This is a load-bearing error that cannot be fixed by local editing; the advertised bound and its QCP and QSAD specializations would need to be either removed or replaced by a genuinely feasible ansatz."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The SDP part of this paper is real work. The CAD problem -- ordered hypotheses where no answer may deviate from the truth by more than a distance Δ -- is a natural interpolation between unambiguous and minimum-error discrimination, and the reformulation (5) in terms of the Gram matrix is clean, efficient, and likely correct. The numerical results for the quantum change point (sharp jump in success probability at Δ=1) and for QSAD are also plausible and potentially useful. I would happily cite the SDP contribution if the paper were revised.\n\nThe problem is Section III. The claimed lower bound (11) is built on an infeasible ansatz. The authors correctly note that a 2x2 principal minor of Z_ME − Z_Δ gives the necessary condition (a−z)b ≥ |x|² (Eq. 7), and then they use AM-GM to get a weaker inequality (8). They then choose the central diagonal z to saturate (8) and declare victory. But saturation of the weaker inequality does not respect the original determinant condition. For the rank-one blocks used in the QCP and QSAD applications, |x|² = ab, so (7) reduces to (a−z)b ≥ ab, i.e. z ≤ 0. Since Z ≥ 0 forces z ≥ 0, the only feasible value is z = 0. The positive z prescribed by (9) violates the very minor that produced (8). No off-diagonal completion can repair this, because the violation is already present in a 2x2 principal minor. Consequently Eq. (11), the QCP bound (21), and the QSAD bounds (28)/(30) are unsupported. The Fourier-analysis appendix is fine but addresses a different question; an exponential decay of S_{k,l} does not rescue an infeasible witness.\n\nSo the paper has one load-bearing flaw in its central advertised result. The SDP formulation and numerics may stand, but the analytical lower bound is a casualty. As is, the paper is not acceptable. That said, the underlying problem is well motivated and the SDP part deserves a serious referee rather than a desk rejection. A revision that removes or fixes the lower-bound claim, perhaps replacing it with numerical evidence or a different proof strategy, could make this a useful contribution.\n\nA serious editor should send this to peer review, because the error is nontrivial and the SDP contribution is worth salvaging. For a reading group, it could be instructive as a case study in infeasible SDP ansatze, but I would not cite it in its current form.","headline":"The SDP formulation for certified-answer discrimination is a genuine contribution, but the advertised general lower bound does not survive contact with the positivity constraints it is derived from.","tokens_in":12270,"tokens_out":3687,"would_cite":false,"duration_ms":36726,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.-a"],"model":"deepseek-v4-flash","headline":"Ordered chains of linearly independent pure quantum states admit certified-answer discrimination as a compact semidefinite program, and a lower bound on success probability follows from the minimum-error solution alone.","keywords":["quantum state discrimination","certified answers","semidefinite programming","unambiguous discrimination","minimum error discrimination","quantum change point","quantum state anomaly detection","Gram matrix"],"falsifier":"Take the quantum change point with $n=4$ and overlap $c=0.9$, compute $Z^{\\mathrm{ME}}$ from the square-root measurement, set $\\Delta=1$, and test whether the diagonal entries from Eq. (9) admit any positive-semidefinite completion satisfying $\\Phi_{1,\\Delta}[Z]\\le Z^{\\mathrm{ME}}$; if no completion exists, or if the value $\\tilde{P}_s$ from (11) exceeds the optimum of the SDP (5), the bound is refuted.","tokens_in":11119,"feed_emoji":"🎯","tokens_out":13199,"duration_ms":116584,"temperature":0.7,"pith_summary":"This paper introduces certified-answer discrimination (CAD) for multi-hypothesis quantum detection in which the candidate states are linearly independent pure states with a natural chain ordering. The task is to maximize the probability of identifying the true state under the hard constraint that no answer may differ from the true position by more than a chosen distance $\\Delta$; this sits between unambiguous discrimination ($\\Delta=0$) and minimum-error discrimination ($\\Delta=n-1$). The paper's central claim is that this optimization is a compact semidefinite program whose data is only the Gram matrix of the states, together with a general lower bound on the certified success probability that can be read off from the solution of the unconstrained minimum-error problem alone. The payoff is a quantitative certificate: for any allowed error distance, one knows how much success probability is sacrificed for answer quality without running a new optimization. Specialized to the quantum change point, the bound predicts an exponential return to the minimum-error value as $\\Delta$ grows, and numerically a single allowed unit of error more than doubles the success probability.","feed_headline":"Certified-answer quantum discrimination fits in one SDP","feed_subtitle":"A bound from the minimum-error solution alone predicts the cost of answers that stay close to the truth.","key_machinery":"The argument rests on three pieces. First, the Gram-matrix transformation: with $R=\\sum_i|\\Psi_i\\rangle\\langle i|$, the POVM constraints become $\\Phi_\\Delta[Z]\\le G$, so all physics of the ordered hypotheses is encoded in the Gram matrix $G$ and in the linear map $\\Phi_\\Delta=\\Phi_2\\circ\\Phi_{1,\\Delta}$, which embeds the block-diagonal variable $Z$ into an $n\\times n$ matrix; this shrinks the SDP from $n^2$ variables to $[n(2\\Delta+1)-\\Delta(\\Delta+3)]$. Second, the lower-bound construction: starting from the minimum-error solution $Z^{\\mathrm{ME}}$, the paper chooses the diagonal entries of a feasible ansatz by saturating the $2\\times 2$ principal-minor inequality between a diagonal entry and the entry $\\Delta+1$ steps away, which yields the formula (11). Third, for the quantum change point, the quantitative bound (21) follows from an exponential-decay estimate for the elements of $S=\\sqrt{G}$, obtained by bounding the Fourier coefficients of $(1-2c\\cos\\theta+c^2)^{-1/2}$ through a contour-integral argument; this decay is what converts the abstract bound (11) into the explicit exponential approach to the minimum-error value.","core_discovery":"In the paper's own terms, the central discovery is that the optimal $\\Delta$-scheme for $n$ linearly independent ordered pure states is given by the semidefinite program (5): maximize $\\frac{1}{n}\\mathrm{Tr}[ZA]$ subject to $\\Phi_\\Delta[Z]\\le G$, $Z\\ge 0$, where $G=R^\\dagger R$ is the Gram matrix built from $R=\\sum_i |\\Psi_i\\rangle\\langle i|$, and $\\Phi_\\Delta$ is a linear map that encodes which answers at distance greater than $\\Delta$ are forbidden. The Gram matrix carries all discrimination properties of the hypotheses, and the SDP variable has only $O(n\\Delta)$ free entries instead of $n^2$. The second claim is the lower bound (11), which expresses the certified success probability as the minimum-error value $P_s^{\\mathrm{ME}}$ minus a sum of correction terms $H_i(\\Delta)$ that depend only on the minimum-error solution $Z^{\\mathrm{ME}}$; the corrections are obtained by saturating $2\\times 2$ principal-minor inequalities in the relaxed constraint $\\Phi_{1,\\Delta}[Z]\\le Z^{\\mathrm{ME}}$. For the quantum change point with overlap $c$, exponential decay of the square-root matrix $S=\\sqrt{G}$ turns this into $\\tilde{P}_s\\ge (1-2c\\,e^{\\Delta\\log c}+c^2e^{2\\Delta\\log c})\\,P_s^{\\mathrm{ME}}$, and for quantum state anomaly detection it gives a linear interpolation between the unambiguous and minimum-error success probabilities once $\\Delta\\ge\\lfloor n/2\\rfloor$.","pith_inferences":["Not drawn in the paper: the saturation ansatz that underlies the bound could be tested numerically on random ordered Gram matrices; finding a case with no positive-semidefinite completion, or a value of $\\tilde{P}_s$ above the true SDP optimum, would pinpoint exactly when the bound breaks.","If the bound holds generally, the natural reading is that the price of certification is controlled by the tail of the minimum-error outcome distribution: ordered problems whose ME measurement already concentrates its errors within distance $\\Delta$ pay almost nothing for certified answers, which is precisely what the QCP numerics show.","The same SDP machinery transfers directly to non-symmetric tolerances (different allowed distances forward and backward), with the bound acquiring separate $H_i(\\Delta_+,\\Delta_-)$ terms; this is a concrete, testable generalization the paper only sketches.","For linearly dependent or noisy ordered states the Gram-matrix route fails, but the CAD constraints can be written directly on the POVM elements; whether an analogous minimum-error-only bound survives there is open."],"forward_implications":["CAD supplies a tunable interpolation between unambiguous discrimination ($\\Delta=0$) and minimum-error discrimination ($\\Delta=n-1$); for the quantum change point the minimum-error regime is already reached at $\\Delta\\approx 8$, so the most relevant operating range uses a small SDP.","The bound (11) means that solving the minimum-error problem once gives, for free, a certificate of how much success probability is lost when answers must stay within distance $\\Delta$ of the truth; no new optimization is needed.","For the quantum change point, the certified success probability approaches the minimum-error value at least as fast as $(1-2c\\,e^{\\Delta\\log c}+c^2e^{2\\Delta\\log c})$; numerically, allowing one error unit raises success from $0.27$ to $0.50$ for $n=25$, $c=0.6$, while $90\\%$ of answers remain within one position.","For quantum state anomaly detection, every $\\Delta<\\lfloor n/2\\rfloor$ is provably equivalent to unambiguous discrimination, and the lower bound interpolates linearly between the unambiguous and minimum-error values above that threshold, matching the symmetry of the problem.","Because the SDP dimension is $O(n\\Delta)$ rather than $n^2$, certified discrimination remains numerically tractable precisely in the regime where certification is meaningful."],"supporting_citations":[{"why":"Supplies the minimum-error discrimination framework whose success probability the bound (11) is measured against.","marker":"[1]"},{"why":"Establishes that unambiguous discrimination requires linear independence, the condition that justifies the Gram-matrix transformation.","marker":"[2]"},{"why":"The earlier interpolation scheme between unambiguous and minimum-error discrimination that the certified-answer scheme is contrasted with.","marker":"[5]"},{"why":"Defines the quantum change point problem and provides the square-root-measurement approximation with exponentially decaying elements used in bound (21).","marker":"[13]"},{"why":"Gives the exact quantum change point solutions and the symmetry-based SDP ansatz that motivates the lower-bound construction.","marker":"[14]"},{"why":"Shows how unambiguous discrimination is formulated as a semidefinite program, the template for the certified-answer SDP.","marker":"[19]"},{"why":"Introduces the square-root ('pretty good') measurement used as the stand-in for the optimal minimum-error solution in the QCP bound.","marker":"[24]"},{"why":"Establishes optimality of the square-root measurement for the QSAD Gram matrix, making the bound exact in that regime.","marker":"[25]"}],"fun_headline_variants":["Certified quantum discrimination solved by one SDP","Bounded-error answers via a single SDP","One SDP certifies quantum state discrimination","SDP bounds error for ordered quantum states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The lower-bound proof assumes that a matrix whose diagonal entries are chosen to saturate the relaxed $2\\times 2$ positivity inequalities can always be completed into a positive-semidefinite block matrix that still satisfies the measurement constraint, but the paper gives no such completion and the $2\\times 2$ condition is only necessary, not sufficient.","fun_headline_variants_meta":{"raw":{"variants":["Certified quantum discrimination solved by one SDP","Bounded-error answers via a single SDP","One SDP certifies quantum state discrimination","SDP bounds error for ordered quantum states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000484,"raw_usage":{"total_tokens":2397,"prompt_tokens":962,"completion_tokens":1435,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":578,"completion_tokens_details":{"reasoning_tokens":1378}},"tokens_in":578,"tokens_out":1435,"duration_ms":14021,"temperature":1.0,"reasoning_tokens":1378,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:54:35.608171+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the quantum change point with $n=4$ and overlap $c=0.9$, compute $Z^{\\mathrm{ME}}$ from the square-root measurement, set $\\Delta=1$, and test whether the diagonal entries from Eq. (9) admit any positive-semidefinite completion satisfying $\\Phi_{1,\\Delta}[Z]\\le Z^{\\mathrm{ME}}$; if no completion exists, or if the value $\\tilde{P}_s$ from (11) exceeds the optimum of the SDP (5), the bound is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the minimum-error discrimination framework whose success probability the bound (11) is measured against."},{"cited_title":"Chefles, Unambiguous discrimination between linearly independent quantum states, Physics Letters A 239, 339 (1998)","cited_arxiv_id":null,"evidence_quote":"Establishes that unambiguous discrimination requires linear independence, the condition that justifies the Gram-matrix transformation."},{"cited_title":"Bagan, R","cited_arxiv_id":null,"evidence_quote":"The earlier interpolation scheme between unambiguous and minimum-error discrimination that the certified-answer scheme is contrasted with."},{"cited_title":"Sent´ ıs, E","cited_arxiv_id":null,"evidence_quote":"Defines the quantum change point problem and provides the square-root-measurement approximation with exponentially decaying elements used in bound (21)."},{"cited_title":"Sent´ ıs, J","cited_arxiv_id":null,"evidence_quote":"Gives the exact quantum change point solutions and the symmetry-based SDP ansatz that motivates the lower-bound construction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows how unambiguous discrimination is formulated as a semidefinite program, the template for the certified-answer SDP."},{"cited_title":"Hausladen, R","cited_arxiv_id":null,"evidence_quote":"Introduces the square-root ('pretty good') measurement used as the stand-in for the optimal minimum-error solution in the QCP bound."},{"cited_title":"Dalla Pozza and G","cited_arxiv_id":null,"evidence_quote":"Establishes optimality of the square-root measurement for the QSAD Gram matrix, making the bound exact in that regime."}],"review_version":1}