{"id":"265f2d5f-73f4-4105-b22b-bfa6ee370ace","arxiv_id":"2507.00679","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":1,"one_line_summary":"Under a symmetry condition, an SDI security witness equals twice the sum of visibility and distinguishability, and a tightened bound lowers the certification threshold to D+V > 1.332.","lead":"This paper shows that a security witness for semi-device-independent quantum key distribution can be read directly from two classic interferometer measurements: fringe visibility and path distinguishability. The authors also propose a tighter security threshold and demonstrate the idea in a fiber interferometer using orbital-angular-momentum states of light.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed improved security threshold PB>0.833 rests on inequality (27), whose proof uses the false step 'sum_i m_i <= 1' for three unit Bloch vectors; without an added mutually-unbiased condition the enlarged secure region is unsupported.","rationale":"The reader's weakest assumption and my stress-test concern coincide: the only step that produces the improved security threshold is inequality (27), and that inequality is not proven by the text. The mistake is localized to Methods 1: for three arbitrary unit Bloch vectors, sum_i m_i <= 1 is false, and the displayed inequality is dimensionally inconsistent. The main mapping S=2(D+V), the symmetry and normalization conditions (19)-(20), and the classical bound D+V>1 remain plausible and are not undermined by this flaw. I also considered the weak-coherent-state dimension mismatch; that is a genuine experimental caveat, but it is less load-bearing than Eq. (27) because the theoretical construction is stated at the level of qubit preparations and the experiment is presented as a proof of principle rather than a full security demonstration. A corrected paper should either add the missing measurement constraint that would justify Eq. (27) or withdraw the PB>0.833 threshold; with that correction the conditional verdict remains appropriate.","tokens_in":11837,"tokens_out":6892,"duration_ms":85703,"concrete_test":"Check Eq. (27) by a numerical search over all qubit states n and all three binary measurement Bloch vectors m_i; the maximum of sum_i (n·m_i)^2 is attained, for example, at n=(1,0,0), m0=(1,0,0), m1=(1/sqrt(2),1/sqrt(2),0), m2=(1/sqrt(2),0,1/sqrt(2)), yielding 2 > 1. If such a triple is realizable by Bob's or Eve's measurements in the (4,2,2) prepare-and-measure scenario, Eq. (27) is false as stated; re-derive the security bound with the missing constraint (e.g., mutually unbiased measurement directions) or remove the PB>0.833 claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Methods 1 derives the tightened SDI security condition PB>0.833 from Eq. (27): [E(a0)]^2+[E(a1)]^2+[E(a0 xor a1)]^2 <= 1. The derivation sets E_i = n·m_i and claims sum_i (n·m_i)^2 <= |n|^2 sum_i m_i <= 1 because 'Since sum_i m_i <= 1'. This is not valid: the m_i are three unit Bloch vectors, so the vector sum can have norm up to 3, and the written inequality is dimensionally inconsistent; the correct bound is sum_i (n·m_i)^2 <= lambda_max(sum_i m_i m_i^T), which exceeds 1 unless the three measurement directions are pairwise orthogonal/mutually unbiased. Methods 1 states no such constraint. A concrete counterexample is n=(1,0,0), m0=(1,0,0), m1=(1/sqrt(2),1/sqrt(2),0), m2=(1/sqrt(2),0,1/sqrt(2)), for which the sum of squares is 2. Therefore Eq. (27), the upper bound (31), and the threshold PB>0.833 are unsupported as stated. The core mapping S=2(D+V) and the original threshold D+V>1.366 are not affected by this flaw, but the advertised enlargement of the secure parameter region does not follow from the presented argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a connection between wave-particle duality and semi-device-independent (SDI) security by expressing the (4,2,2) prepare-and-measure witness S in terms of input distinguishability D and interferometric visibility V. Under the symmetry relations (19) and normalization conditions (20), the authors derive max_{\\phi_x} S_{\\phi_x} = 2(D+V), so that measuring D and V is claimed to suffice for certifying non-classicality and key-rate positivity. They report a fiber-optical experiment with orbital-angular-momentum weak coherent states (mean photon number \\mu=0.2) that measures D, V, and S as a function of the tunable beam splitter phase, and they claim to validate the theoretical prediction. They also propose an improved security threshold P_B > 0.833 based on an inequality derived in Methods 1.","tokens_in":12188,"tokens_out":6068,"duration_ms":67556,"significance":"If correct, the mapping S=2(D+V) is a clean, operationally meaningful link between complementarity and SDI certification, and it could simplify experimental certification in prepare-and-measure quantum communication. The derivation of the mapping is explicit and internally consistent under the stated assumptions, and the experimental platform is described in detail with error bars, representing a genuine proof-of-principle of the interferometric technique. However, the improved security bound rests on an invalid inequality, and the weak-coherent-state implementation breaks the qubit-dimension premise of the SDI framework, so the certification claims currently outrun what the evidence supports.","major_comments":[{"comment":"The improved security condition P_B > 0.833 relies on inequality (27), whose proof is invalid. The step \"Since \\sum_i m_i \\le 1\" is false: the m_i are three unit Bloch vectors, and their vector sum can have norm up to 3. A concrete counterexample is n=(1,0,0), m0=(1,0,0), m1=(1/\\sqrt2,1/\\sqrt2,0), m2=(1/\\sqrt2,0,1/\\sqrt2), for which \\sum_i (n\\cdot m_i)^2 = 2 > 1. Equation (29) is also dimensionally inconsistent, as it places a vector sum on the right-hand side where a scalar bound is needed. Consequently, Eqs. (30)-(32) and the threshold P_B>0.833 are unsupported as stated. The core mapping S=2(D+V) and the original threshold D+V>1.366 from Ref. [32] are not affected by this flaw, but the advertised enlargement of the secure parameter region does not follow from the presented argument.","section":"Methods 1, Eqs. (27)-(32)"},{"comment":"The experiment uses weak coherent states with average photon number \\mu=0.2 per pulse, so the physical Hilbert space is infinite-dimensional and the emitted states are coherent states rather than qubits. The SDI framework certifies security under the assumption that the system dimension is bounded by 2. Multi-photon components violate this premise, and the manuscript provides no squash model, post-selection argument, or other justification that would map the experiment onto a qubit prepare-and-measure scenario. Therefore the data in Fig. 3 cannot, on their own, certify non-classicality or key-rate positivity in the SDI sense; at most they demonstrate the interferometric mapping under an additional, unproven assumption.","section":"Methods 3 / Experimental assessment"},{"comment":"The central identity max_{\\phi_x} S_{\\phi_x} = 2(D+V) is derived by algebraic rearrangement of the same detection probabilities that define D, V, and S under the symmetry and normalization conditions (19)-(20). The agreement between the measured S and 2(D+V) in Fig. 3 is therefore a consistency check of the interferometric model and of the symmetry constraints, rather than a test against an independent prediction. The abstract's phrase \"validating our theoretical predictions\" overstates the evidential value; the paper should explicitly state that the experimental results confirm the internal consistency of the derivation under the assumed symmetries, not independently validate the SDI relation.","section":"Results B, Eqs. (16) and (21)"}],"minor_comments":[{"comment":"There is a bracket typo in the third term: \"[(E(a0\\oplus a1)]2\" should read \"[E(a0\\oplus a1)]2\".","section":"Methods 1, Eq. (27)"},{"comment":"The dotted black line is described as marking \"the classical bound D+V>1\"; it would be clearer to write \"D+V=1\", since violation occurs for D+V>1.","section":"Figure 3 caption"},{"comment":"The abbreviations \"Bias\", \"IM\", and \"ATT\" in Fig. 2 are not defined in the caption or the main text; defining them would help readability.","section":"Results D, Fig. 2"},{"comment":"The assumption that P_{B,E}(a0)=P_{B,E}(a1) is introduced without discussion; while it may follow from uniformity of a0 and a1 plus a symmetry argument, the manuscript should state the justification explicitly.","section":"Methods 1, Eq. (22)-(26)"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a potentially valuable core result—the mapping S=2(D+V)—but it is presented as a certification of SDI security, which the current evidence does not support. The improved bound in Methods 1 is demonstrably false as stated; the authors should either remove it or supply a correct proof under clearly stated additional assumptions (e.g., mutually unbiased measurement directions). The weak-coherent-state experiment cannot certify a dimension bound without a squash model or an equivalent trusted-qubit argument. I recommend that the editor require the authors to (i) fix or remove the improved bound, (ii) explicitly address the dimension-breaking issue and either provide a valid qubit reduction or downgrade the certification claim to a proof-of-principle demonstration of the mapping, and (iii) temper the validation language in the abstract and conclusions. If these changes are made, the mapping result may be publishable in a quantum-information journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper's main result is the explicit expression of the SDI witness as S = 2(D+V) under the interferometric symmetries. That is a clean algebraic observation, and the experimental data support it: the points in Fig. 3 fall on the predicted curve, and the setup—a reconfigurable few-mode fiber interferometer with OAM states—is a genuine piece of work. Credit where due: the mapping is not present in the cited prior work, and the experiment does demonstrate the relation in a practical platform.\n\nThe soft spot is the improved security bound in Methods 1. The step 'Since sum_i m_i <= 1' for three unit Bloch vectors is simply false; the vector sum can have norm up to 3. Inequality (27) does not follow without an extra mutually-unbiased constraint, which the paper does not state. A concrete counterexample exists (e.g., n along x, m0 along x, m1 and m2 at 45 degrees in the yz planes), giving a sum of squares of 2. So the claimed PB > 0.833 threshold and the corresponding D+V > 1.332 line in Fig. 3 are unsupported. The original threshold D+V > 1.366 from Pawlowski and Brunner is not affected. This is a load-bearing flaw in the advertised 'improved bound', but it does not sink the rest of the paper.\n\nAnother concern is the experiment uses weak coherent states with mean photon number 0.2. That breaks the strict dimension-2 premise of SDI: multi-photon components are not negligible, so the 'security' certification is not rigorous as presented. It is a proof-of-principle demonstration of the mapping, not a valid SDI security certification under the stated assumptions. The circularity point is real but mild: Eq. (16) and (21) are algebraic rearrangements of the same probabilities that define D, V, and S, so the experiment is a consistency check rather than a test of an independent prediction. That is acceptable if framed as a reformulation, which the paper mostly does.\n\nBottom line: the paper is a useful reformulation with a decent experiment and a flawed ancillary claim. It deserves a serious referee, but only after the improved bound is fixed or removed and the WCS issue is addressed. I would not cite it in my own work in the next 12 months, but I would bring it to a reading group focused on SDI or complementarity.","headline":"The core mapping S=2(D+V) is a clean reformulation with solid experimental backing, but the improved security threshold rests on a false inequality and should not be accepted as stated.","tokens_in":12702,"tokens_out":2660,"would_cite":false,"duration_ms":34549,"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":"The paper shows that in the symmetric (4,2,2) prepare-and-measure scenario, the semi-device-independent witness equals exactly $S=2(D+V)$, so measuring visibility and input distinguishability certifies non-classicality and secure key rate.","keywords":["semi-device-independent security","wave-particle duality","entropic uncertainty relations","quantum key distribution","prepare-and-measure scenario","interferometric visibility","input distinguishability","orbital angular momentum"],"falsifier":"Take three identical measurement axes; then the geometric quantity inside inequality (27) equals 3, not at most 1, so the improved security threshold $P_B>0.833$ does not follow from the stated argument.","tokens_in":11673,"feed_emoji":"🔐","tokens_out":8620,"duration_ms":89612,"temperature":0.7,"pith_summary":"This paper aims to make wave-particle duality a practical certification tool for semi-device-independent security, where only the dimension of the quantum system is trusted. It derives a compact identity for the (4,2,2) prepare-and-measure scenario: once symmetry and normalization conditions are imposed, the SDI witness $S$ satisfies $\\max_{\\phi_x} S_{\\phi_x} = 2(D+V)$, with $D$ the input distinguishability and $V$ the interferometric visibility. Because both quantities are extracted from simple photon-counting statistics in a Mach-Zehnder interferometer, the identity turns raw interference data into a certificate of non-classicality and of positive key rate. The paper also presents a refined security bound, lowering the required threshold to $D+V > 1.332$, and reports a proof-of-principle experiment with orbital-angular-momentum states in a fiber interferometer that tracks the predicted curve. If the derivation holds, complementarity experiments become a direct route to quantum-cryptographic security statements.","feed_headline":"Two interferometer readings certify quantum key security","feed_subtitle":"A single witness built from visibility and distinguishability lets one interferometer certify secure key distribution.","key_machinery":"The carrying object is the pair $(D,V)$: input distinguishability $D=2p_{\\mathrm{guess}}(\\text{which-path})-1$ and interferometric visibility $V=2p_{\\mathrm{guess}}(\\text{output observable})-1$, both defined through optimal guessing probabilities in the spirit of entropic uncertainty. The mechanism is Bob's tunable beam splitter, which interpolates between the particle-like and wave-like measurements $M_0(\\phi_s,\\phi_x)$ and $M_1(\\phi_s,\\phi_x)$. That tunability, together with the parity-based encoding $(|0\\rangle, |1\\rangle, |+\\rangle, |-\\rangle)$, is what lets every term of the SDI witness be read as a distinguishability or a visibility, culminating in Eq. (21), $\\max_{\\phi_x} S_{\\phi_x}=2(D+V)$.","core_discovery":"The paper's central claim is that the SDI witness $S$ of Eq. (3) decomposes into operational wave-particle quantities. Starting from the parity-based encoding of Alice's four states and Bob's tunable beam-splitter measurements, the authors rewrite $S_{\\phi_x}$ as a sum of configuration-dependent distinguishabilities and visibilities (Eq. (16)). Imposing the interferometer symmetries (19) and the parity-oblivious normalization conditions (20) reduces that sum to the closed form $\\max_{\\phi_x} S_{\\phi_x} = 2(D+V)$. Since the classical bound is $S \\le 2$, non-classicality is certified exactly when $D+V > 1$, and key-rate positivity is certified when the stricter security thresholds are met. The reported experiment scans the tunable beam splitter from particle-like to wave-like measurements and observes that the data agree with $S/2 = D+V$, with the largest violation at $\\phi_s = \\pi/4$, where $D=V=\\sqrt{2}/2$.","pith_inferences":["A testable extension would be to derive analogous $D+V$ decompositions for prepare-and-measure scenarios with more than two inputs; the symmetry-encoding argument suggests the structure may persist, but the paper does not claim it.","The refined security threshold should be treated as conditional on the measurement axes satisfying the geometric inequality used in Methods 1; restricting the axes to mutually unbiased directions would make that condition true, and this restriction is implicit rather than stated.","Because $D$ and $V$ are estimated from raw count maxima and minima, the criterion offers a practical shortcut for monitoring SDI-QKD devices in real time; that operational convenience is an inference from the paper's mapping, not one of its theorems."],"forward_implications":["Semi-device-independent non-classicality certification reduces to checking $D+V>1$; neither maximal visibility alone nor maximal distinguishability alone suffices.","Positive key rate can be certified from interferometric data whenever $D+V$ exceeds the security threshold, $1.332$ under the paper's improved bound or $1.366$ under the original bound of Ref. [32].","A single reconfigurable interferometer can scan the tunable beam splitter across the whole particle-wave range and produce the security certificate without characterizing detectors or sources beyond their dimension.","The optimal operating point is balanced complementarity, $D=V=\\sqrt{2}/2$, reached at $\\phi_s=\\pi/4$, rather than an extreme wave or particle setting.","The proof-of-principle OAM experiment indicates that the relation survives in a realistic few-mode-fiber platform, so the certificate is obtainable with current technology."],"supporting_citations":[{"why":"Defines the SDI security framework, the (4,2,2) dimension witness S, the QRAC connection, and the original security threshold PB > 0.8415 that the paper refines.","marker":"[32]"},{"why":"Supplies the entropic-uncertainty definitions of input distinguishability D and interferometric visibility V used throughout the derivation.","marker":"[18]"},{"why":"Provides the Mach-Zehnder delayed-choice QKD protocol whose parity-based preparation scheme is adapted for the wave-particle game.","marker":"[27]"},{"why":"Earlier experimental demonstration of the entropic-uncertainty/wave-particle equivalence on the same OAM few-mode-fiber platform, which the present experiment builds on.","marker":"[25]"},{"why":"The photonic lantern acts as the input beam splitter that decomposes the OAM state into the two LP modes used as interferometric paths.","marker":"[33]"},{"why":"Establishes the decomposition of |OAM+1> into |LP11a> and |LP11b> that justifies the path encoding in the setup.","marker":"[34]"}],"fun_headline_variants":["Visibility and distinguishability certify key security","Wave-particle witness proves secure key rate","Two measurements lock down quantum key security","One witness from wave-particle duality secures keys"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The improved security threshold rests on a geometric inequality that the paper applies without stating a constraint under which it actually holds; for arbitrary triples of measurement axes in three dimensions the inequality can fail.","fun_headline_variants_meta":{"raw":{"variants":["Visibility and distinguishability certify key security","Wave-particle witness proves secure key rate","Two measurements lock down quantum key security","One witness from wave-particle duality secures keys"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000557,"raw_usage":{"total_tokens":2641,"prompt_tokens":928,"completion_tokens":1713,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":544,"completion_tokens_details":{"reasoning_tokens":1660}},"tokens_in":544,"tokens_out":1713,"duration_ms":14087,"temperature":1.0,"reasoning_tokens":1660,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:10:08.677218+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take three identical measurement axes; then the geometric quantity inside inequality (27) equals 3, not at most 1, so the improved security threshold $P_B>0.833$ does not follow from the stated argument.","supporting_citations":[{"cited_title":"Ac ´ın, N","cited_arxiv_id":null,"evidence_quote":"Defines the SDI security framework, the (4,2,2) dimension witness S, the QRAC connection, and the original security threshold PB > 0.8415 that the paper refines."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the entropic-uncertainty definitions of input distinguishability D and interferometric visibility V used throughout the derivation."},{"cited_title":"Mizutani, T","cited_arxiv_id":null,"evidence_quote":"Provides the Mach-Zehnder delayed-choice QKD protocol whose parity-based preparation scheme is adapted for the wave-particle game."},{"cited_title":"Scarani, H","cited_arxiv_id":null,"evidence_quote":"Earlier experimental demonstration of the entropic-uncertainty/wave-particle equivalence on the same OAM few-mode-fiber platform, which the present experiment builds on."},{"cited_title":"Scarani, N","cited_arxiv_id":null,"evidence_quote":"The photonic lantern acts as the input beam splitter that decomposes the OAM state into the two LP modes used as interferometric paths."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the decomposition of |OAM+1> into |LP11a> and |LP11b> that justifies the path encoding in the setup."}],"review_version":1}