{"id":"604c1ada-60f8-4ff0-a84b-da503a960692","arxiv_id":"2502.03780","paper_version":2,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper claims that at hybrid exceptional-diabolic points in a four-mode bosonic system, the quantum Fisher information diverges as theta^{-4}, giving a quadratic error scaling delta-theta proportional to theta^2 that doubles the scaling exponent of ordinary singularities.","lead":"This paper analyzes a four-cavity sensor tuned to a special singularity where two kinds of degeneracy meet, and claims that a tiny common frequency shift can be estimated with an error that shrinks quadratically with the shift. It matters because such a scaling, if valid, would be a qualitatively stronger precision boost than ordinary exceptional-point sensing schemes.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central θ^{-4} QFI scaling is asserted, not derived: substituting the HED response Eq. (43) into the paper's own Gaussian formula Eq. (29) yields a different leading power unless the direct-transmission term in V_out is shown to cancel, which the paper does not do.","rationale":"The paper's program is clear: use the pole order of the singular response function to predict the QFI scaling at HED points. The structural work—the four-mode model, the input-output formalism, and the Sain–Massey expansion in App. C—is plausibly relevant. What is not established is the link between the response-function pole and the QFI exponent. The reader correctly flags Eq. (40) as imported from Ref. [24], and I agree that is the fragile premise. However, the reader's specific scalar counterexample (F~θ^{-2} for any s when mean and covariance scale as θ^{-s} and θ^{-2s}) does not apply directly to this model: the output covariance in Eq. (23) contains a θ-independent direct-transmission term V_in, so V^{-1} does not scale as θ^{2s}. A direct expansion of Eq. (29) for the stated HED response instead indicates Fθ~θ^{-6}, unless the θ^{-2} term in V_out cancels; the paper neither proves such a cancellation nor displays the calculation. This strengthens rather than weakens the objection: the claimed θ^{-4} scaling is not merely unproven, it appears inconsistent with the paper's own equations. No machine-checked proof or reproducible code accompanies the numerical Fig. 3, so the discrepancy cannot be resolved from the manuscript. The experimental-feasibility section is qualitative and does not bear on the scaling exponent. Keeping the reader's REJECT is therefore appropriate; if the authors supply a direct derivation of Eq. (45) from Eq. (29) and reproducible numerics, a CONDITIONAL acceptance could be reconsidered.","tokens_in":18924,"tokens_out":13573,"duration_ms":141849,"concrete_test":"Directly compute Eq. (29) for the HED case: fix γ=1, κ=1, η_l=0.5, nA=nB=1, n=I, and use Eq. (43) (X0, X1 from App. C) in Eqs. (22)–(23). Evaluate F_θ for θ=10^{-1},...,10^{-6} with exact rational/symbolic arithmetic or high-precision numerics, and fit the log-log slope. If the slope is not -4—in particular if it is ≈-6 as the two-line expansion above indicates—then Eq. (45) and the abstract's twofold claim are wrong. Repeat identically for the non-HED case Eq. (47) (expected slope -2 per Eq. (49)); if both slopes shift, the imported law Eq. (40) is the culprit. Make the reproducing script available.","verdict_should_be":"REJECT","load_bearing_attack":"Equation (45) is the load-bearing step: it converts the Sain–Massey pole order s=2 of the response function into the claimed QFI scaling θ^{-4}. The paper never computes Eq. (29) for the actual output state; it imports Eq. (40) from Ref. [24]. That import is not innocent here. Using the paper's own Eqs. (22)–(23) with the stated Gθ=θ^{-2}(X0+θX1), the output mean and covariance are S_out,θ = S_in − √κ θ^{-2}X0 S_in + O(θ^{-1}) and V_out,θ = V_in − √κ θ^{-2}(X0V_in + V_inX0^T) + O(θ^{-4}). Because V_in is invertible (nA=nB=1 in Fig. 3), V_out^{-1} = V_in^{-1}+O(θ^{-2}), so dS/dθ ~ θ^{-3} and dV/dθ ~ θ^{-3}. Substitution into Eq. (29) then gives Fθ ~ θ^{-6} from both the mean and covariance contributions, not θ^{-4}; only a special cancellation of the θ^{-2} term in V_out could change this, and no such cancellation is shown or derivable from the displayed X0. Thus the abstract's twofold improvement is an unverified imported consequence, and Fig. 3—claimed to be a direct numerical evaluation of Eqs. (29)/(32)—is not reproducible from the manuscript because no code or data accompany it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper considers a four-mode bosonic system with balanced gain and loss, coupled by inter-mode couplings g and J, described by the non-Hermitian dynamical generator H in Eq. (13). It studies the estimation of a small linear perturbation theta to the common cavity frequency, encoded as H_theta = H - theta.n with n = I. The response function G_theta = J(theta.n - H)^{-1} becomes singular when H is singular; for the HED point g = gamma, J = 0, the paper finds via the Sain-Massey expansion a pole of order s = 2 (Eqs. (43)-(44)). Using a scaling law imported from Ref. [24] (Eq. (40)), the paper concludes that the quantum Fisher information scales as F_theta ~ theta^{-4}, giving an estimation error delta_theta ~ theta^2 (Eqs. (45)-(46)), and that heterodyne detection achieves the same scaling (Eq. (51)). This is contrasted with non-HED singular points where F_theta ~ theta^{-2} and delta_theta ~ theta. The paper argues for a 'twofold improvement' at HED points and discusses experimental feasibility and imperfections.","tokens_in":19182,"tokens_out":11609,"duration_ms":104922,"significance":"If the claimed scaling were correct, the paper would provide a concrete bosonic model in which hybrid exceptional-diabolic points give a superlinear enhancement in parameter estimation, extending the authors' previous two-mode results to a four-mode system with a diabolic subspace. The model setup and the input-output formalism are clearly presented, and the Sain-Massey computation of the pole order in Appendix C is explicit and appears internally consistent. The identification of heterodyne detection as an optimal measurement for this Gaussian model is also a useful contribution. However, the central quantitative claim rests on an unproven and, as I show in the major comments, incorrect application of the imported scaling law. The significance is therefore conditional: the qualitative idea of HED-enhanced sensing may survive, but the specific twofold-improvement result is not established by the manuscript.","major_comments":[{"comment":"The scaling law F_theta = theta^{-2s}[b0+O(theta)] in Eq. (40) is imported from Ref. [24] and is applied with s=2 at the HED point, but it is inconsistent with the manuscript's own formulas (22), (23), and (29). With G_theta = theta^{-2}X0 + theta^{-1}X1 from Eq. (43) and V_in = (2n_A+1)I, Eq. (23) gives V_out,theta = V_in - sqrt(kappa) theta^{-2}(X0 V_in + V_in X0^T) + O(theta^{-1}). The matrix X0 in Eq. (C12) is not skew-symmetric (for example, the (1,2) and (2,1) elements of X0+X0^T equal 2*gamma), so the theta^{-2} term does not cancel. Consequently dV_out/dtheta ~ theta^{-3}, V_out^{-1} = V_in^{-1} + O(theta^{-2}), and substitution into Eq. (29) yields F_theta ~ theta^{-6} from the covariance term (and, for S_in different from zero, from the mean term as well). This contradicts Eq. (45). The paper provides no cancellation argument that would restore F_theta ~ theta^{-4}. Since Eq. (45) is the load-bearing step for the claimed twofold improvement, the central result is unsupported.","section":"Sec. III C, Eq. (40); Sec. IV, Eq. (45)"},{"comment":"Equation (41) defines b0 without derivation, and the expression is not the coefficient obtained by inserting Eqs. (22)-(23) into Eq. (29) for the present four-mode model. It appears to be taken from Ref. [24], where the sensor configuration (and in particular the role of the direct transmission term I - Kprobe G_theta) differs. Because b0 determines the leading QFI coefficient in both the HED case (Eq. (45)) and the non-HED case (Eq. (49)), the authors must provide a self-contained derivation of Eq. (41) in the current model or show explicitly how Eq. (29) reduces to it. As written, the step from the response-function expansion (39) to the QFI scaling (40) is an assumption, not a theorem.","section":"Sec. III C, Eq. (41)"},{"comment":"The numerical results in Fig. 3 are described as a direct evaluation of Eqs. (29) and (32) and are said to confirm Eqs. (46), (50), and (52). However, no code, data, or numerical parameters beyond those in the caption are provided, so the figure cannot be independently reproduced or checked. Given that the analytic scaling predictions are themselves in question (see above), the figure cannot serve as independent support for the claimed scaling. The authors should make the numerical routines available or provide a transparent analytic calculation that resolves the discrepancy between the claimed theta^{-4} scaling and the direct expansion of Eq. (29).","section":"Fig. 3"},{"comment":"Appendix C correctly shows, via the Sain-Massey rank condition, that the response function at the HED point has a pole of order s=2 and computes X0 and X1. This is a useful and apparently valid computation. However, the pole order of G_theta does not by itself determine the scaling of the quantum Fisher information for the output state; the connection is the imported Eq. (40), which is not established here (see above). The appendix therefore does not bridge the gap between the singular response and the claimed estimation error scaling.","section":"Appendix C"}],"minor_comments":[{"comment":"The heading 'ENHANCED PRECESION' contains a typo; it should read 'ENHANCED PRECISION'.","section":"Sec. IV heading"},{"comment":"The text refers to a 'non-HEP singularity' in the paragraph after Eq. (46); this should be 'non-HED' for consistency with the abstract and the rest of the paper.","section":"Sec. IV"},{"comment":"Equation (5) as written, delta^2(theta) >= 1/F_theta >= 1/F_theta, repeats the same symbol and is not meaningful; the authors should distinguish the quantum and classical Fisher information with different notation if that is the intent.","section":"Eq. (5)"},{"comment":"In Eq. (33), the expression 'theta n theta H^{-1}' contains an extra theta after n; it should presumably be 'theta n H^{-1}'.","section":"Eq. (33)"},{"comment":"The table lists delta_Q theta = 1/F_theta and delta_C theta = 1/F_theta, but the correct Cramer-Rao relation is delta = 1/sqrt(F), as used in Eqs. (46) and (50); the table should be corrected.","section":"Table I"}],"recommendation":"reject","confidential_remarks":"The paper's main quantitative claim depends heavily on the authors' own PRL [24], but the key lemma is not reproduced or verified for the four-mode model. The numerical figure is not accompanied by code or data, which is particularly problematic given the central role of the scaling plot. The topic is within the scope of a quantum-sensing journal, but the execution does not meet the standard of a refereed publication in its present form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper has a clear structure and a genuinely new object: a four-mode bosonic sensor where a second-order EP and a second-order DP coincide, and the claim that the DP subspace does not spoil the quadratic error scaling. The pole-order computation in Appendix C is the strongest part; the Sain–Massey ranks check out and s=2 is correct for n=I at the HED point. The authors also engage honestly with the no-advantage literature and do not oversell EPs as a free lunch.\n\nThe trouble is the load-bearing step. Equation (45) is asserted, not derived. The paper never computes Eq. (29) for the actual output state; it imports Eq. (40) from Ref. [24] and applies it with s=2. But that import is not innocent here. Substituting the response expansion G ~ theta^-2 X0 into the paper's own equations (22)–(23) gives S_out = S_in + O(theta^-2) and V_out = V_in + O(theta^-2). Since V_in is invertible, the derivatives are O(theta^-3) for both mean and covariance. Feeding that into the standard Gaussian QFI formula (29) yields F_theta ~ theta^-6 from the mean term and the covariance term alike, not theta^-4. The only way to recover theta^-4 would be a cancellation of the leading theta^-2 term in V_out, and no such cancellation is shown or visible from the displayed X0. So the abstract's 'twofold improvement' is unverified and, on a direct substitution, appears wrong.\n\nFig. 3, which claims to numerically confirm the scaling, is not reproducible from the manuscript: no code or data are provided, and the parameter range is limited to n=I and exact resonance. Those would be minor issues if the analytics were right; here they compound the central problem.\n\nYes, this paper deserves a serious referee — the HED concept is interesting, and catching this kind of error is exactly what peer review is for. But as it stands, the main result does not hold. The authors should re-derive the QFI directly from their Gaussian formulas before resubmitting.","headline":"The four-mode HED model is real and the pole-order calculation is careful, but the claimed theta^-4 QFI scaling does not follow from the paper's own equations — it looks like an inherited error from Ref. [24].","tokens_in":19789,"tokens_out":3376,"would_cite":false,"duration_ms":33929,"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":"At hybrid exceptional-diabolic points, a four-mode bosonic sensor estimates a frequency shift with error scaling as θ², twice as fast as at ordinary singular points.","keywords":["quantum sensing","exceptional points","diabolic points","hybrid exceptional-diabolic singularities","quantum Fisher information","non-Hermitian systems","Gaussian state estimation","heterodyne detection"],"falsifier":"Compute the exact quantum Fisher information in Eq. (29) for the four-mode model at $g=\\gamma$, $J=0$, $n=I$ over a range of small $\\theta$, using the full response matrix rather than its truncated Laurent expansion. If the log-log slope of $\\delta\\theta$ versus $\\theta$ near the HED point is 2, the claim stands; if the slope is 1, the pole-order scaling law is not transferring to the hybrid point.","tokens_in":18657,"feed_emoji":"🎯","tokens_out":12359,"duration_ms":111445,"temperature":0.7,"pith_summary":"This paper argues that a four-mode bosonic cavity with balanced gain and loss can estimate a small linear perturbation with an error that shrinks quadratically in the perturbation strength, provided the system is tuned to a hybrid exceptional-diabolic (HED) point—the parameter point where an exceptional curve meets a diabolic plane and the dynamical generator becomes singular. At such a point, the response function develops a pole of order two, which the authors translate, through the quantum Fisher information, into a $\\theta^{-4}$ divergence of the Fisher information and a $\\delta\\theta\\propto\\theta^{2}$ estimation error. Away from HED points, the same singular generator gives only a first-order pole and a linear $\\delta\\theta\\propto\\theta$ error, so the HED point doubles the scaling exponent. The paper also shows that heterodyne detection reproduces this optimal scaling, making the enhancement accessible to a standard continuous-variable measurement.","feed_headline":"Hybrid exceptional-diabolic points double sensor precision","feed_subtitle":"Near the four-mode cavity's hybrid point, estimation error shrinks as θ² instead of θ, a twofold gain in scaling.","key_machinery":"The load-bearing object is the resonant response function, the symplectic representation of $(\\theta n-H)^{-1}$: all $\\theta$-dependence of the output amplitude and covariance matrix enters through this matrix, so its pole structure controls the Fisher information. The authors expand the inverse of the singular matrix $\\theta n-H$ with the Sain-Massey method, which fixes the pole order $s$ by a rank condition on augmented matrices and provides the coefficients $X_0,X_1,\\ldots$; at the HED point these are $X_0=H|_{g=\\gamma}$ and $X_1=I$, giving $s=2$. A scaling law from the earlier two-mode analysis, $F_{\\theta}=\\theta^{-2s}[b_0+O(\\theta)]$, then converts the pole order into the estimation error. The HED point is the parameter point $g=\\gamma$, $J=0$ where the exceptional-plane condition, the diabolic-plane condition, and the singularity-surface condition coincide.","core_discovery":"The authors' central discovery is that hybrid exceptional-diabolic singularities are a distinct and stronger resource for singularity-enhanced quantum sensing than ordinary exceptional or diabolic points. For their Hamiltonian with balanced gain and loss, the singular surface $J=\\sqrt{g^{2}-\\gamma^{2}}$ contains the HED point $g=\\gamma$, $J=0$ as well as non-HED points such as $g=\\sqrt{2}\\gamma$. Using the Sain-Massey expansion of the resonant response function with perturbation matrix $n=I$, they find a second-order pole at the HED point, $G_{\\theta}=\\theta^{-2}(H|_{g=\\gamma}+\\theta I)$ in the symplectic representation, and only a first-order pole away from it. Feeding this expansion into the Gaussian-state quantum Fisher information gives $F_{\\theta}=\\theta^{-4}[b'_0+O(\\theta)]$ at the HED point, hence $\\delta\\theta\\propto\\theta^{2}$, while at the non-HED singular point it gives $F_{\\theta}=\\theta^{-2}[b''_0+O(\\theta)]$, hence $\\delta\\theta\\propto\\theta$. The same two scalings are obtained for the classical Fisher information under heterodyne detection.","pith_inferences":["If the pole-order-to-Fisher scaling law holds generally, a higher-order hybrid singularity in a larger bosonic lattice would predict even steeper error scaling, $\\delta\\theta\\propto\\theta^{s}$, although detector inefficiency and the breakdown of the Laurent expansion at finite $\\theta$ would cap the practical gain.","The exact Gaussian quantum Fisher information at the HED point could be evaluated without the Laurent ansatz; a direct calculation of the derivative terms would reveal whether the $\\theta^{-4}$ divergence comes from the amplitude response, the covariance response, or both, a distinction that matters for designing the measurement.","A concrete experimental test would be to tune $g$ through $\\gamma$ at $J=0$ in a photonic or circuit-QED array and watch the log-log slope of the estimation error change from 1 to 2.","Since heterodyne detection saturates the bound, a simpler homodyne measurement on the most responsive quadrature may achieve the same scaling with fewer resources."],"forward_implications":["A sensor tuned to the HED point $(g,J)=(\\gamma,0)$ estimates the common-mode frequency shift with error $\\delta\\theta\\propto\\theta^{2}$, so precision improves quadratically as the perturbation shrinks.","The quadratic scaling survives the presence of the second-order diabolic subspace: the orthogonal subspace co-existing with the exceptional point does not reduce the pole order.","Heterodyne detection reaches the same scaling as the quantum Cramér–Rao bound, so the predicted enhancement is attainable with a standard measurement rather than an idealized optimal one.","The pole-order criterion implies that simultaneous satisfaction of the exceptional, diabolic, and singularity conditions is needed for the twofold scaling gain; breaking any one condition lowers the enhancement.","The result carries the two-mode singularity-enhanced sensing analysis over to a four-mode architecture, showing the enhancement is not limited to minimal two-mode models."],"supporting_citations":[{"why":"Supplies the singular-generator scaling law that converts a response pole of order $s$ into a $\\theta^{-2s}$ divergence of the quantum Fisher information, and the two-mode framework this paper extends.","marker":"[24]"},{"why":"Provides the Sain-Massey expansion used to compute the pole order $s=2$ and the coefficients $X_0$ and $X_1$ at the HED point.","marker":"[42]"},{"why":"Derives the Gaussian-state quantum Fisher information formula used to calculate precision from the output amplitude and covariance matrix.","marker":"[39]"},{"why":"Gives the rank criterion for the order of a pole of a matrix function, used in the appendix to certify the second-order pole by rank differences.","marker":"[43]"},{"why":"Supplies the direct-sum decomposition of the Hamiltonian used to identify the exceptional, diabolic, and HED curves in parameter space.","marker":"[8]"}],"fun_headline_variants":["HED points slash sensing error scaling in four-mode bosonic system","Four-mode HED point delivers quadratic quantum sensing boost","HED singularity doubles quantum sensing precision scale","Hybrid singular point gives quadratic error scaling in sensing","Twofold gain: HED points refine quantum sensing limit"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on the earlier result that a response pole of order $s$ makes the quantum Fisher information diverge as $F_{\\theta}=\\theta^{-2s}$; if that result fails at the hybrid point, the claimed $\\delta\\theta\\propto\\theta^{2}$ error scaling does not follow.","fun_headline_variants_meta":{"raw":{"variants":["HED points slash sensing error scaling in four-mode bosonic system","Four-mode HED point delivers quadratic quantum sensing boost","HED singularity doubles quantum sensing precision scale","Hybrid singular point gives quadratic error scaling in sensing","Twofold gain: HED points refine quantum sensing limit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000434,"raw_usage":{"total_tokens":2173,"prompt_tokens":869,"completion_tokens":1304,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":485,"completion_tokens_details":{"reasoning_tokens":1226}},"tokens_in":485,"tokens_out":1304,"duration_ms":9367,"temperature":1.0,"reasoning_tokens":1226,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T00:50:07.991922+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the exact quantum Fisher information in Eq. (29) for the four-mode model at $g=\\gamma$, $J=0$, $n=I$ over a range of small $\\theta$, using the full response matrix rather than its truncated Laurent expansion. If the log-log slope of $\\delta\\theta$ versus $\\theta$ near the HED point is 2, the claim stands; if the slope is 1, the pole-order scaling law is not transferring to the hybrid point.","supporting_citations":[{"cited_title":"Quantum noise theory of exceptional point amplify- ing sensors,","cited_arxiv_id":null,"evidence_quote":"Supplies the singular-generator scaling law that converts a response pole of order $s$ into a $\\theta^{-2s}$ divergence of the quantum Fisher information, and the two-mode framework this paper extends."},{"cited_title":"Quantum Fisher information for states in exponential form,","cited_arxiv_id":null,"evidence_quote":"Provides the Sain-Massey expansion used to compute the pole order $s=2$ and the coefficients $X_0$ and $X_1$ at the HED point."},{"cited_title":"Gaussian quantum infor- mation,","cited_arxiv_id":null,"evidence_quote":"Derives the Gaussian-state quantum Fisher information formula used to calculate precision from the output amplitude and covariance matrix."},{"cited_title":"Statistical distance and the geometry of quantum states,","cited_arxiv_id":null,"evidence_quote":"Gives the rank criterion for the order of a pole of a matrix function, used in the appendix to certify the second-order pole by rank differences."},{"cited_title":"Diabolical points in the spec- tra of triangles,","cited_arxiv_id":null,"evidence_quote":"Supplies the direct-sum decomposition of the Hamiltonian used to identify the exceptional, diabolic, and HED curves in parameter space."}],"review_version":1}