{"id":"79b2808e-3d5c-4eb1-97f2-819a3e88ea4e","arxiv_id":"2601.09977","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In a symmetric three-photon Fourier interferometer, engineered super-Poissonian light with g(2)≈1.9 and g(3)≈3.6 maximizes visibility at ≈0.61, surpassing the magnitude of the single-photon value.","lead":"A theory result shows that adding controlled intensity noise to light can sharpen, rather than blur, three-photon interference. In a symmetric three-port Fourier circuit, engineered super-Poissonian light achieves higher interference visibility on the paper's metric than ideal single photons, hinting that classical light may calibrate such circuits more efficiently.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Noise-enhancement claim is tied to the Eq. (2) visibility normalization; under the alternative product-normalized visibility the classical advantage inverts (≈0.27 vs ≈0.67 for single photons).","rationale":"The reader's weakest_assumption identified exactly the load-bearing issue: the classical noise advantage is an artifact of the visibility normalization in Eq. (2). The paper explicitly discusses the alternative normalization, but the central abstract and conclusion make an unqualified claim that statistical noise enhances interference. If the claim were restricted to V as defined in Eq. (2), it would be correct but much weaker. The proposed concrete check directly tests whether the enhancement survives a reasonable alternative normalization; the predicted outcome (reversal) would demonstrate that the headline conclusion is measure-dependent. I concur with the reader that this warrants a conditional verdict rather than rejection: the algebra is internally consistent, the classical bound in Eq. (10) follows from Cauchy–Schwarz, and the authors are transparent about their chosen definition. However, the broad wording 'statistical-noise-enhanced multi-photon interference' is not robust to normalization choices, which is precisely why the paper should be accepted only with clarification or qualification.","tokens_in":13410,"tokens_out":6975,"duration_ms":79032,"concrete_test":"Recompute the DFT coincidence probabilities with the alternative visibility V_alt = 1 − P_id/(⟨n1⟩⟨n2⟩⟨n3⟩) for the optimized classical mixture (g2≈1.9, g3≈3.6), single photons, and laser light, all at equal mean photon number. If V_alt(classical) ≈ 0.27 falls below V_alt(single) ≈ 0.67, the noise-enhancement claim is an artifact of the Eq. (2) denominator and the paper should qualify its headline to the chosen visibility measure.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result—classical super-Poissonian light surpassing |V_single|=0.5—rests entirely on V=1−P_id/P_dist defined in Eq. (2). This ratio does not merely cancel a common noise factor: P_dist itself contains the g(3) and g(2) terms injected by classical noise, so the large V_cl≈0.61 is substantially produced by dividing by a noise-inflated denominator. The paper acknowledges the alternative normalization 1−P_id/(⟨n1⟩⟨n2⟩⟨n3⟩) and dismisses it, but that dismissal is a physical choice rather than a derivation. For the optimized classical mixture (g2≈1.9, g3≈3.6), this alternative gives V_alt≈1−0.733≈0.27, while single-photon inputs give V_alt≈1−0.333≈0.67—the hierarchy reverses. The headline 'statistical-noise-enhanced' therefore describes a property of the chosen measure, not an invariant feature of the interference. Additionally, comparing +0.61 with −0.5 depends on taking absolute values across opposite-sign visibilities; a signed comparison or a requirement of the same sign would weaken the 'surpassing' statement. The algebraic core of Eqs. (7)–(10) is internally consistent and the Cauchy–Schwarz bound is correct, so the issue is not a mathematical error but an overgeneralized interpretation tied to a non-unique normalization.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies multi-photon interference in symmetric linear-optical circuits, focusing on three-photon coincidence probabilities in the 3×3 discrete Fourier transform (DFT) circuit and in a one-parameter family of symmetric unitaries. The authors define the interference visibility as V=1−P_id/P_dist, where P_id and P_dist are the normalized threefold coincidence probabilities for fully indistinguishable and fully distinguishable inputs with the same photon statistics. For the DFT they derive V^(3) = (6g^(2)−1)/(g^(3)+6g^(2)+2), and using the classical inequality g^(3)≥(g^(2))^2 they maximize this visibility over classical states, obtaining V≈0.61 for a modulated-laser mixture with g^(2)≈1.9, g^(3)≈3.6, which exceeds |V|=−0.5 for single-photon inputs. They also analyze partial mode overlap and a phase-tunable circuit, showing a phase-dependent reordering of visibilities and claiming a form of statistical complementarity between quantum and classical advantages.","tokens_in":13748,"tokens_out":13796,"duration_ms":142306,"significance":"The algebraic core of the paper is sound: I checked the DFT coincidence probabilities, the Cauchy–Schwarz bound, and the visibility optimization, and they are internally consistent. The paper supplies a full supplemental derivation and gives explicit, parameter-free formulas. The result is a useful counterexample to the intuition that single-photon inputs always maximize multi-photon interference contrast, and it identifies a practical regime where classical super-Poissonian light gives a larger HOM-like visibility magnitude than single photons, which could be relevant for circuit calibration. The significance is, however, conditional on the chosen visibility measure: the enhancement is a property of the ratio in Eq. (2) rather than an invariant of raw coincidence rates.","major_comments":[{"comment":"The central claim that engineered super-Poissonian noise enhances interference is tied to the visibility definition V=1−P_id/P_dist. The alternative normalization 1−P_id/(⟨n1⟩⟨n2⟩⟨n3⟩), which the paper dismisses in one sentence, reverses the ordering: for the optimized classical mixture (g2≈1.9, g3≈3.6) it yields ≈0.27, while single photons yield ≈0.67. The dismissal that this alternative 'introduces a trivial background scaling factor g(N)' is qualitative and not a derivation; the alternative does not simply reintroduce a common factor, and it removes the distinguishable baseline. Please quantify this comparison and either justify Eq. (2) as the unique physically relevant visibility or explicitly qualify the title/abstract claims as measure-dependent.","section":"Introduction and Eq. (2)"},{"comment":"The 'surpassing' of the single-photon signature uses the absolute value: V_cl≈+0.61 vs V_single=−0.5. A negative visibility is a coincidence peak, a positive one a dip; the two have opposite physical interpretation. The authors should state explicitly that the comparison is of contrast magnitudes and should discuss whether the practical alignment advantage relies on the signed or unsigned visibility. As stated, the claim could be misread as classical light achieving a larger positive dip than the single-photon negative dip, which is true, but the significance of comparing across signs needs an argument.","section":"Noise-Enhanced visibility in 3-port beamsplitter, Eq. (9)–(10)"},{"comment":"The paper claims that quantum and classical advantages are mutually exclusive resources, but immediately notes a narrow phase window in Fig. S2 where both Fock states (n≥3) and engineered noise surpass the laser benchmark. This contradicts the 'mutually exclusive' phrasing unless 'mutual exclusivity' is defined with respect to a particular phase or in a limit. Please provide a precise statement of the claimed complementarity and specify its domain, e.g., at the DFT point, or soften the claim to 'cannot be simultaneously large'.","section":"Conclusion / statistical complementarity"}],"minor_comments":[{"comment":"The sentence 'the upper bound converges to a finite value of 0.4 in the limit g(2)→∞' refers to the Gaussian-state bound, not the classical bound V_cl, which tends to zero. Please make the antecedent explicit.","section":"After Eq. (10)"},{"comment":"In the Fig. 4 caption, please clarify that the dashed blue curve is the absolute value of the single-photon visibility; the solid blue curve is the signed visibility.","section":"Circuit-dependent reordering of visibilities"},{"comment":"The sentence 'mixtures of Gaussian states in the regime g(2)>4/9 can violate the bound' should read 'mixed Gaussian states' and should note that Eq. (11) is a bound for pure Gaussian states.","section":"Eq. (11) discussion"},{"comment":"The term 'statistical complementarity' is evocative but not defined. If kept, please provide an operational definition, for example in terms of the signs of the visibility or the phase window where one resource outperforms the other.","section":"Conclusion"}],"recommendation":"major_revision","confidential_remarks":"The mathematical content is correct and the paper is clearly written. The main risk is that the headline result is a property of the HOM-like visibility normalization rather than an invariant measure; the authors do discuss the alternative normalization but not quantitatively. I believe a major revision that addresses the normalization dependence and softens the 'complementarity' claim would make the paper acceptable. There is no question of novelty or internal inconsistency."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper is worth a serious look. The algebraic core is clean and verifiable: the three-photon coincidence probabilities for the DFT tritter depend on g(2) and g(3) in a way that makes the visibility non-monotonic, and a classical super-Poissonian mixture can push it to ~0.61, above the single-photon magnitude of 0.5. I spot-checked the permanent calculation and the Supplement; everything is internally consistent. The derivation of the classical bound from g(3) ≥ (g(2))^2 is correct, and the mode-mismatch analysis is a nice addition. The result is genuinely new relative to the cited distinguishability literature, which focused on mode mismatch rather than photon statistics as a control parameter.\n\nThe main soft spot is the one the authors themselves flag: the enhancement lives in the visibility normalization of Eq. (2). Dividing by P_dist, which itself contains g(2) and g(3) terms, does some of the work. With the alternative normalization they mention, the hierarchy reverses. I don't think this is fatal—every visibility metric involves a choice, and theirs is physically motivated as isolating interference from trivial count-rate scaling—but the headline claim should be understood as 'enhanced relative to this particular measure,' not as an invariant feature of the interference. The paper says this, but more quietly than it might.\n\nThe bigger overreach is the 'statistical complementarity' claim. Their own numerical analysis finds only a narrow phase window where both Fock-state and noise inputs beat the laser baseline, and the mutual-exclusivity language is more assertive than the evidence. That is fixable by softening the phrasing and making the phase-dependence explicit. There's also a minor point that the optimized mixture is described as a modulated laser, but no experimental recipe is given; that's okay for a theory paper.\n\nThis deserves peer review. It is internally consistent, the math is solid, and the normalization caveat is a matter of interpretation rather than error. A good referee will push for a clearer discussion of what the visibility measure is and isn't capturing, and for a more careful statement of what complementarity means.","headline":"Solid, checkable theory with a real non-monotonic visibility result, but the 'noise enhancement' is normalization-dependent and the complementarity claim overreaches the evidence.","tokens_in":14215,"tokens_out":2157,"would_cite":true,"duration_ms":23519,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Photon-number noise, not single photons, can maximize three-photon interference visibility in the 3×3 DFT circuit.","keywords":["multi-photon interference","photon statistics","Hong-Ou-Mandel visibility","discrete Fourier transform circuit","super-Poissonian light","statistical complementarity","three-photon interference","intensity correlations"],"falsifier":"Send a modulated-laser source with g(2) ≈ 1.9 and g(3) ≈ 3.6 into a 3×3 DFT and measure the three-fold coincidence visibility; if V(3) does not reach about 0.61, or if using the alternative normalization by ⟨n1⟩⟨n2⟩⟨n3⟩ reverses the ordering relative to single photons, the paper's central claim would be refuted.","tokens_in":13300,"feed_emoji":"🔆","tokens_out":6727,"duration_ms":65077,"temperature":0.7,"pith_summary":"The paper claims that the usual intuition—statistical noise always degrades multi-photon interference—breaks down for three-photon interference. Working with the normalized visibility V = 1 − P_id/P_dist, it shows that in a 3×3 discrete-Fourier-transform circuit the three-photon coincidence contrast is a non-monotonic function of the intensity correlations g(n). Classical light obeying g(3) = (g(2))^2 reaches a maximum visibility of about 0.61 near g(2) ≈ 1.9 and g(3) ≈ 3.6, which exceeds the magnitude 0.5 of ideal single-photon inputs. The paper also shows that as the circuit phase is tuned, the visibility ordering among sub-Poissonian, Poissonian, and super-Poissonian light inverts, implying that quantum and classical advantages are mutually exclusive resources—a statistical complementarity. A sympathetic reader would care because it changes what counts as a good light source for multi-photon interferometry and suggests a practical alignment tool.","feed_headline":"Super-Poissonian light tops single photons in 3-port interference","feed_subtitle":"Tuning photon-number fluctuations to g(2)≈1.9 lifts three-photon contrast past the single-photon level.","key_machinery":"The load-bearing object is the normalized interference visibility V(N) = 1 − P_id/P_dist, where P_id and P_dist are the N-fold coincidence probabilities for fully indistinguishable and fully distinguishable inputs of the same source. For the 3×3 discrete Fourier transform circuit—the balanced three-port unitary—the paper derives V(3) = (6g(2) − 1)/(g(3) + 6g(2) + 2) from the normalized intensity-correlation functions g(2) and g(3), then applies the classical Cauchy–Schwarz bound g(3) ≥ (g(2))^2 to locate the maximum classical value. A phase-tunable symmetric unitary with parameter φ parametrizes a family of circuits including the DFT and sweeps the visibility ordering.","core_discovery":"In the paper's own terms, the central discovery is that photon statistics is not a one-way resource for multiphoton interference. For one-photon-per-port input to the 3×3 DFT, the coincidence probability develops an anti-bunching peak and the visibility is V(3) = −0.5. For engineered super-Poissonian input with g(2) ≈ 1.9 and g(3) ≈ 3.6, the same circuit yields V(3) ≈ +0.61, so the magnitude of the classical-noise signature exceeds the single-photon signature. The mechanism is an interplay between the g(3) term, which only enters the indistinguishable-input probability, and the g(2) terms, which survive in the distinguishable-input probability; the chosen visibility normalization cancels the","pith_inferences":["Editorial inference: the claimed 0.61 > 0.5 comparison relies on the paper's visibility normalization and on taking absolute values of opposite-sign visibilities; under a raw-coincidence normalization, the classical advantage would likely shrink or reverse.","Editorial inference: the result suggests a direct experimental protocol—intensity-modulate a laser to target g(2) ≈ 1.9 and g(3) ≈ 3.6, then record three-fold coincidences in a tritter; the predicted sharp fringe is a clean test of the normalization scheme.","Editorial inference: the phase-dependence hints that statistical complementarity may be a general property of symmetric multiports; one could test whether the same trade-off appears in four-photon suppressed-output circuits, a direction the paper leaves open.","Editorial inference: because V(3) is negative for single photons and positive for classical light, comparing magnitudes may conflate a sign flip with an enhancement; a signed visibility or a separate measure of bunching versus anti-bunching would clarify the resource comparison."],"forward_implications":["Two-photon HOM visibility decreases monotonically with g(2), but three-photon DFT visibility does not: it rises with statistical noise up to an optimum, then falls.","A classical light source generated by a modulated laser, with g(3) = (g(2))^2, can yield V ≈ 0.61 in the DFT, beating the |V| = 0.5 single-photon signature.","Phase tuning of the symmetric circuit inverts the visibility hierarchy: single-photon, Poissonian, and super-Poissonian inputs each win in different phase regions, with the laser benchmark never lowest.","For calibration and alignment of three-photon circuits, super-Poissonian light offers both higher count rates and a sharper interference fringe, avoiding the trade-off that single photons impose.","Quantum (sub-Poissonian) and classical (engineered super-Poissonian) advantages are mutually exclusive: improving one suppresses the other, a form of statistical complementarity."],"fun_headline_variants":["Super-Poissonian light boosts 3-photon interference","Classical noise tops single photons in 3-port setup","Noisy light enhances multiphoton visibility","Super-Poissonian statistics lift three-photon contrast","Statistical fluctuations improve 3-photon interference"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The whole noise-enhancement conclusion rests on defining visibility as V = 1 − P_id/P_dist, which cancels the trivial statistical bunching of the source; if one instead normalizes by the product of average output counts, the g(2) background reappears and the classical advantage shrinks or inverts.","fun_headline_variants_meta":{"raw":{"variants":["Super-Poissonian light boosts 3-photon interference","Classical noise tops single photons in 3-port setup","Noisy light enhances multiphoton visibility","Super-Poissonian statistics lift three-photon contrast","Statistical fluctuations improve 3-photon interference"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00015,"raw_usage":{"total_tokens":1002,"prompt_tokens":679,"completion_tokens":323,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":423,"completion_tokens_details":{"reasoning_tokens":247}},"tokens_in":423,"tokens_out":323,"duration_ms":3881,"temperature":1.0,"reasoning_tokens":247,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T10:26:03.811873+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Send a modulated-laser source with g(2) ≈ 1.9 and g(3) ≈ 3.6 into a 3×3 DFT and measure the three-fold coincidence visibility; if V(3) does not reach about 0.61, or if using the alternative normalization by ⟨n1⟩⟨n2⟩⟨n3⟩ reverses the ordering relative to single photons, the paper's central claim would be refuted.","supporting_citations":[],"review_version":1}