{"id":"394a8853-081b-4a47-a866-e907eaf3fbf5","arxiv_id":"1908.01265","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"New two-operator heat trace invariants are defined and their first two short-time asymptotic coefficients are computed explicitly.","lead":"This paper introduces new spectral invariants of a pair of elliptic operators on a manifold, built from the heat traces of both operators together, and proves they have short-time asymptotic expansions whose coefficients encode the geometry of both operators. The result gives mathematicians and quantum field theorists a new set of observables that depend on eigenvalues and eigenfunctions, not just eigenvalues, potentially distinguishing manifolds that have the same spectrum.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of Theorem 1 does not control derivatives of the heat-kernel remainders inside the geodesic ball, so the term-by-term Laplace expansion is not fully justified as written.","rationale":"The reader identified the uniform heat-kernel estimates behind Lemma 6 as the weakest assumption; I agree and refine the concern: the real gap is not the off-diagonal exponential decay, which is standard, but the control of heat-kernel remainders and their derivatives inside B_r when the product is fed into Laplace's method. This step underpins the existence of the whole asymptotic expansion, and therefore also the explicit coefficients in Theorems 2 and 3. The paper's internal consistency checks (equal-operator reduction, commuting case) are real supporting evidence, and the missing estimates are likely obtainable by standard parametrix or Duhamel arguments on compact manifolds, so the concern does not justify rejection. It does justify the moderate-confidence conditional verdict already given: the central theorem is plausible but not completely proved as written, and the Dirac coefficient c1 is additionally under-verified because the final reduction from (7.88) to (1.50) is omitted. No change to the reader's verdict is needed.","tokens_in":40134,"tokens_out":39936,"duration_ms":405748,"concrete_test":"Truncate the heat-kernel expansion at order N for a compact test case (e.g., a flat torus with two constant metrics/potentials), keep the remainder R_N, and prove or derive the uniform bound |∇_x^m ∇_y^{m'} R_N(t;x,y)| ≤ C_{N,m,m'} t^{N+1-n/2-(m+m')/2} e^{-c σ(x,y)/t} on B_r×B_r for all m,m' up to 2N+n. Insert this bound into the Laplace expansion of the remainder integral and verify that the contribution is O(ε^N) for every N. If the needed derivative-order control cannot be established, Theorem 1 needs additional hypotheses before the term-by-term Laplace expansion is valid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 6.2 replaces U± by the infinite heat-kernel expansion (6.2) inside B_r and then applies Lemma 5 term-by-term to the product. A rigorous proof must truncate at order N, writing U± = U_{N,±} + R_{N,±}, and then show that the remainder ∫ exp[-Σ/(2εts)] R_{N,+}(εt)R_{N,-}(εs) contributes only at order ε^N after the Laplace method. The only estimate stated, (6.21), controls the full heat kernel outside B_r; it says nothing about R_N near the diagonal. Lemma 5 is not a statement about C^0-small errors: its coefficients are derivatives of the integrand at the diagonal. A remainder that is pointwise O(t^N) but whose first 2k derivatives grow like t^{-m/2} can contribute at order ε^{N-m/2}, potentially contaminating low-order coefficients. The required C^m off-diagonal heat-kernel remainder estimates are standard, and compactness makes their existence plausible, but they are neither stated nor proved in the paper. Without them, Theorem 1's expansion (1.20) is not established as written, and the same gap propagates through Lemma 8 to Theorem 3 and the explicit coefficients b1 and c1. This is a rigor gap rather than evidence that the theorem is false, but it is the most load-bearing unsupported step in the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces new relative spectral invariants of pairs of Laplace- and Dirac-type operators on compact Riemannian manifolds without boundary, namely the traces Ψ(t,s) and Φ(t,s) built from differences of heat semigroups, together with the 'combined heat traces' X(t,s)=Tr e^{-tL_+}e^{-sL_-} and Y(t,s)=Tr D_+ e^{-tD_+^2}D_- e^{-sD_-^2}. The central claim (Theorem 1) is that as ε→0 these combined traces admit asymptotic expansions in powers of ε whose coefficients are integrals of local invariants constructed polynomially from the two operators' symbols. The first two coefficients are computed explicitly for both the Laplace case (Theorem 2) and the Dirac case (Theorem 3), and the invariants are related to the Bogolyubov invariant of quantum field theory. The proofs rely on the Laplace method applied to products of heat kernel expansions, with detailed derivations of the leading coefficients and consistency checks in the case of equal or commuting operators.","tokens_in":40375,"tokens_out":20708,"duration_ms":179955,"significance":"If established, the paper would provide a new family of spectral invariants that depend not only on the eigenvalues but also on the eigensections of a pair of operators, and would give explicit leading asymptotics; this is a potentially useful contribution to spectral geometry and quantum field theory. The paper is largely self-contained, presents long explicit coefficient formulas, and includes nontrivial sanity checks (equal-operator and commuting limits) that support the correctness of the computations. Its main weakness is that the proof of Theorem 1 is not rigorous as written: the term-by-term Laplace expansion of a product of two heat kernel asymptotic series is not justified by the estimates supplied. This gap is likely fixable with standard heat kernel remainder estimates, so the central claim is plausibly true, but the current manuscript does not demonstrate it.","major_comments":[{"comment":"In deriving (1.20), the proof substitutes the full asymptotic heat kernel expansion (6.2) into the integral (6.18) and then applies Lemma 5 term-by-term to each Λ_m in (6.25). This interchange is not justified. Lemma 5 requires the integrand φ to be smooth, and its coefficients (5.42)–(5.45) depend on derivatives of φ at the diagonal; a remainder that is merely pointwise small can still contaminate low-order coefficients if its derivatives grow like ε^{-k/2}. The only off-diagonal estimate given, (6.21), controls the full heat kernel outside B_r(x'), not the remainder R_N inside the ball near the diagonal. Therefore the asymptotic expansion (1.20) and the definitions of B_k and b_k are not established as written. A rigorous proof must truncate the heat kernel expansion at finite order N, control the C^m norms of the remainders inside B_r(x') near the diagonal (for example by standard derivative estimates for heat kernels on compact manifolds), and show the remainder contributes only at order ε^{N-C}; this is missing.","section":"§6.2 (proof of Theorem 1, Laplace case)"},{"comment":"The same rigor gap appears in the Dirac case, and the text explicitly states 'we will omit some details'. Lemma 8 is asserted without proving the required derivative estimates for the remainders of D_±U_± inside B_r(x'); the estimate (6.39) applies only outside the geodesic ball. Since the expansion (1.21) and the explicit coefficients c_k in Theorem 3 depend on Lemma 8, the Dirac part of Theorem 1 and Theorem 3 are not fully proved as written. The required truncation of the heat kernel expansion must be applied to the differentiated heat kernels as well, with uniform C^m remainder estimates near the diagonal; this is a load-bearing step, not a cosmetic omission.","section":"§6.3 (proof of Theorem 1, Dirac case)"}],"minor_comments":[{"comment":"The abstract contains line-break artifacts such as 'smoo th' and 'eigenval ues'; these should be cleaned before publication.","section":"Abstract and §1"},{"comment":"The ε-scaling in (6.24)–(6.25) is not transparent: the reader may wonder how the (4πεts)^{-n/2} prefactor is compensated. A sentence noting that the Laplace integral over the geodesic ball supplies a factor ε^{n/2} would help.","section":"§6.2, Eqs. (6.24)–(6.25)"},{"comment":"In (6.50) the series is written with ε^{m-2} and later C_{-1} is shown to vanish; it would be clearer to state immediately that the effective expansion starts with ε^{-1}.","section":"§6.3, Eq. (6.50)"},{"comment":"The claim that the new invariants 'contain much more information about geometry' is heuristic; the dependence on eigensections is explicit in (3.19), but no evidence is given that the invariants actually distinguish non-isometric isospectral geometries.","section":"§1 and §3.2"},{"comment":"The Bogolyubov motivation is brief and defers details to [7]; the formulas (2.62)–(2.63) would benefit from at least a sketch of the inversion of the relations (2.59)–(2.61).","section":"§2"}],"recommendation":"major_revision","confidential_remarks":"The paper's core idea is novel and the explicit coefficient computations are impressive. The missing uniform remainder estimates are a standard technical step, but they are essential for the proof of Theorem 1. I recommend requiring this addition before publication; if the author cannot supply the estimates, the theorem should be restated with the necessary hypotheses or proved by a more explicit parametrix construction. The companion paper [7] should also be checked for the same issue."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe headline: this paper actually does something new. The combined heat traces X(t,s) and Y(t,s), and the relative invariants Ψ and Φ, are genuinely two-operator invariants; earlier work, including Avramidi's own, only handled single-operator spectral invariants. He proves they have a short-time asymptotic expansion in ε and computes the first two coefficients. The theorems are plausible and, based on the detailed intermediate steps, almost certainly correct.\n\nWhat is good: the derivation is self-contained and checks out in the special cases where you can compare with the classical heat trace (equal operators, L+ = L− + constant, commuting operators). Those are meaningful consistency tests, not decoration. The explicit formulas for b0, b1 and c0 are believable; c1 is enormous but the route to it is laid out. The citation pattern is fine; the reliance on standard heat kernel coefficients is normal and not circular.\n\nThe real soft spot is exactly where the stress-test note lands. In Section 6.2 the proof of Theorem 1 assumes you can replace each heat kernel by its full asymptotic series inside the geodesic ball and apply the Laplace method (Lemma 5) term-by-term. That is only justified if you control the remainder and its derivatives near the diagonal. The only estimate the paper states, (6.21), controls the full kernel outside the ball; it says nothing about the remainder inside. The needed C^m off-diagonal estimates are standard and compactness makes their existence plausible, but they are absent. This is a fillable rigor gap, not a sign of a false theorem, but it is load-bearing: the same gap flows through Lemma 8 into Theorem 3.\n\nTwo smaller issues. The Dirac part explicitly omits details ('we will omit some details'), and c1 in (1.50) is a monster; I would not want to certify it without a computer algebra check, though the consistency checks help. And the abstract's phrase 'much more information about geometry' is stronger than anything the paper actually quantifies; a sharper statement about which geometric data appear in b1 and c1 would be better.\n\nWho should read this: people who do heat kernel asymptotics, spectral geometry with pairs of operators, and QFT particle creation. It deserves serious peer review. My recommendation is to send it out, but with a referee request that the expansion theorem be tightened—either a proper remainder estimate or an explicit statement that the expansion is formal and standard methods make it rigorous. If the author supplies that, the paper is in decent shape. If not, it's still a useful contribution, but the proof as written falls short of the claim.","headline":"New two-operator heat-trace invariants with a plausible but not fully rigorous asymptotic proof; first coefficients computed.","tokens_in":40922,"tokens_out":3660,"would_cite":false,"duration_ms":38167,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["58J35","58J50","35P20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that combined heat traces of a pair of Laplace or Dirac type operators on a compact manifold admit short-time asymptotic expansions whose coefficients are integrals of local invariants built from both operators, and it…","keywords":["relative spectral invariant","combined heat trace","heat kernel expansion","Laplace type operator","Dirac type operator","Ruse-Synge function","local invariants","Bogolyubov invariant"],"falsifier":"On a flat torus, take $L_+=-\\Delta$ and $L_-=-(\\nabla+iA)^2$ with a constant connection one-form $A$, and compute $X(\\varepsilon t,\\varepsilon s)=\\sum_{k,j} e^{-\\varepsilon(t\\lambda_k+s\\mu_j)}|(\\phi_j^-,\\phi_k^+)|^2$ exactly from the Fourier basis; compare the coefficients of $\\varepsilon^0$ and $\\varepsilon^1$ with formulas (1.42)-(1.43). A mismatch in the coefficient of $\\varepsilon^1$ would refute Theorem 2, while matching on several $t,s$ ratios would support it.","tokens_in":39901,"feed_emoji":"📐","tokens_out":7773,"duration_ms":72882,"temperature":0.7,"pith_summary":"The paper introduces new spectral quantities attached to a pair of elliptic operators, namely Laplace-type operators $L_\\pm$ and Dirac-type operators $D_\\pm$ on a compact manifold, via the combined heat traces $X(t,s)=\\operatorname{Tr} e^{-tL_+}e^{-sL_-}$ and $Y(t,s)=\\operatorname{Tr} D_+ e^{-tD_+^2}D_- e^{-sD_-^2}$. It proves that, as $\\varepsilon\\to 0$, these traces have full asymptotic expansions in powers of $\\varepsilon$ whose coefficients are integrals of local scalar invariants built from the symbols of both operators. The first two coefficients are computed explicitly for both traces. Because the combined traces involve products of the two heat semigroups, they depend on overlap inner products of eigensections of the two operators, so they carry spectral information that the individual heat traces lose. The motivation is that such invariants control particle creation in quantum field theory and may sharpen the geometric information obtainable from spectra.","feed_headline":"Two-operator heat traces expand into local invariants","feed_subtitle":"Combined traces of Laplace and Dirac pairs encode eigenfunction overlaps; first coefficients are explicit.","key_machinery":"The machinery is the Laplace method applied to the product of two heat kernels. The combined trace is written as an integral over $M\\times M$ of the product $U_+(t;x,x')U_-(s;x',x)$, and for $\\varepsilon\\to 0$ the off-diagonal part is shown to be exponentially small using uniform heat kernel bounds. On a common geodesic ball the exponent is the phase $\\Sigma(t,s;x,x')=s\\sigma_+(x,x')+t\\sigma_-(x,x')$ built from the two Ruse-Synge functions, half the squared geodesic distances for the two metrics; its unique nondegenerate critical point is the diagonal, with Hessian the dual metric $G_{ij}=s g^+_{ij}+t g^-_{ij}$. A Morse-lemma reduction converts the integral to a Gaussian average, and the coefficients come from evaluating covariant Taylor coefficients of the heat kernel coefficients, Van Vleck-Morette determinants, and parallel transports at the diagonal. The Ruse-Synge/Laplace-method lemma is the engine that turns the product of two heat kernel expansions into local invariant polynomials.","core_discovery":"The central claim is Theorem 1: for Laplace-type operators $L_\\pm$ and Dirac-type operators $D_\\pm$ on a compact manifold without boundary, there are asymptotic expansions $X(\\varepsilon t,\\varepsilon s)\\sim (4\\pi\\varepsilon)^{-n/2}\\sum_{k\\ge 0}\\varepsilon^k B_k(t,s)$ and $Y(\\varepsilon t,\\varepsilon s)\\sim (4\\pi\\varepsilon)^{-n/2}\\sum_{k\\ge 0}\\varepsilon^{k-1} C_k(t,s)$, where $B_k=\\int g^{1/2} b_k$ and $C_k=\\int g^{1/2} c_k$, with $b_k,c_k$ local invariants built polynomially from covariant derivatives, taken with respect to a time-dependent metric and connection, of the two metrics, the connection-difference vectors $C_\\pm$, and the potentials $Q_\\pm$ and $S_\\pm$; the invariants are symmetric under $(t,L_+)\\leftrightarrow(s,L_-)$ and homogeneous in $t,s$. Theorems 2 and 3 identify the first two coefficients: $b_0=\\operatorname{tr}I$, $b_1$ as in (1.43), $c_0=\\tfrac12 g^{ij}\\operatorname{tr}(\\gamma_i^+\\gamma_j^-)$, and $c_1$ as in (1.50). The corollaries express the relative spectral invariants $\\Psi$ and $\\Phi$, which vanish when the two operators coincide, as combinations of the classical heat trace coefficients and the new $B_k,C_k$.","pith_inferences":["Editorial inference: Because the combined traces carry overlap factors between eigensections of two operators, they may distinguish manifolds that are isospectral for every single Laplace operator; a testable project is to search for such a pair among known isospectral non-isometric manifolds.","Editorial inference: The same phase and Gaussian reduction should extend to products of more than two heat kernels or to higher-order pseudodifferential symbols, producing multi-operator local invariants; this is a natural next step not pursued in the paper.","Editorial inference: The explicit formula for $c_1$ involves commutators $[\\gamma_p^+,\\gamma_q^-]$ and connection-difference terms; in a concrete spinor model these may be interpreted physically as mixing or curvature coupling between the two Dirac fields, offering a direct route to test the asymptotics numerically.","Editorial inference: One could define relative zeta functions from $\\Psi$ and $\\Phi$ and use the short-time expansion to derive functional relations or determinant ratios for two operators; the paper introduces these zeta functions but does not analyze their analytic continuation."],"forward_implications":["The relative spectral invariants $\\Psi(t,s)$ and $\\Phi(t,s)$ defined in (1.3)-(1.4) have asymptotic expansions whose coefficients are explicit combinations of the classical heat invariants $A_k$ and the new local invariants $B_k,C_k$.","The invariants $X$ and $Y$ depend on the eigensections through the overlap factors $|(\\phi_j^-,\\phi_k^+)|^2$, so they contain strictly more spectral data than the individual heat traces; isospectral pairs of operators are not automatically indistinguishable by these invariants.","When the two operators coincide, the general formulas reduce to consistency identities relating $B_k,C_k$ to the classical heat coefficients $A_k$, and the relative invariants $\\Psi$ and $\\Phi$ vanish.","When the two Laplace operators differ only by an additive constant, or when $D_+=D_-+M$ with $M$ anticommuting and $M^2$ scalar, the combined traces factor through classical heat traces, giving exact closed forms.","The Bogolyubov invariants for bosons and fermions are expressed as double integrals of $\\Psi$ and $\\Phi$, so the new short-time asymptotics provide an expansion of particle creation in the in-out formalism."],"supporting_citations":[{"why":"Supplies the covariant heat-kernel technique used to compute heat kernel coefficients and coincidence limits.","marker":"[1]"},{"why":"Provides the Ruse-Synge function and Van Vleck-Morette determinant identities used throughout Section 4.","marker":"[2]"},{"why":"Gives the heat kernel expansion, Gaussian integral identities, and Lemma 5 style asymptotics on which the main theorems rely.","marker":"[5]"},{"why":"Supplies the standard theory of heat kernels and Dirac operators that defines the Dirac-type setup and the potential $Q_\\pm$.","marker":"[10]"},{"why":"Provides the classical heat trace asymptotics and the first coefficients $A_0,A_1$ that the relative invariants are compared with.","marker":"[14]"},{"why":"Supplies the uniform heat kernel estimates used to prove that off-diagonal contributions are exponentially small.","marker":"[15]"},{"why":"Provides the symmetrization identities used to simplify the products of derivatives of the phase in Lemma 5 and the explicit coefficients.","marker":"[17]"}],"fun_headline_variants":["Paired heat traces expose relative spectral invariants","Two-operator asymptotics define new geometric invariants","Relative invariants from paired Laplace-Dirac heat expansions","Heat trace pairs reveal local invariants of geometry"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole asymptotic expansion rests on the assumption that away from the diagonal the combined heat kernel is exponentially small uniformly over the largest geodesic balls valid for both metrics; if that uniform decay failed, the coefficients might cease to be local.","fun_headline_variants_meta":{"raw":{"variants":["Paired heat traces expose relative spectral invariants","Two-operator asymptotics define new geometric invariants","Relative invariants from paired Laplace-Dirac heat expansions","Heat trace pairs reveal local invariants of geometry"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000473,"raw_usage":{"total_tokens":2352,"prompt_tokens":950,"completion_tokens":1402,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":566,"completion_tokens_details":{"reasoning_tokens":1340}},"tokens_in":566,"tokens_out":1402,"duration_ms":11220,"temperature":1.0,"reasoning_tokens":1340,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:18:22.594758+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a flat torus, take $L_+=-\\Delta$ and $L_-=-(\\nabla+iA)^2$ with a constant connection one-form $A$, and compute $X(\\varepsilon t,\\varepsilon s)=\\sum_{k,j} e^{-\\varepsilon(t\\lambda_k+s\\mu_j)}|(\\phi_j^-,\\phi_k^+)|^2$ exactly from the Fourier basis; compare the coefficients of $\\varepsilon^0$ and $\\varepsilon^1$ with formulas (1.42)-(1.43). A mismatch in the coefficient of $\\varepsilon^1$ would refute Theorem 2, while matching on several $t,s$ ratios would support it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the covariant heat-kernel technique used to compute heat kernel coefficients and coincidence limits."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the heat kernel expansion, Gaussian integral identities, and Lemma 5 style asymptotics on which the main theorems rely."},{"cited_title":"Berline, E","cited_arxiv_id":null,"evidence_quote":"Supplies the standard theory of heat kernels and Dirac operators that defines the Dirac-type setup and the potential $Q_\\pm$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the classical heat trace asymptotics and the first coefficients $A_0,A_1$ that the relative invariants are compared with."},{"cited_title":"Grigor’yan, Heat Kernel and Analysis on Manifolds, AMS, International Press, 2009","cited_arxiv_id":null,"evidence_quote":"Supplies the uniform heat kernel estimates used to prove that off-diagonal contributions are exponentially small."},{"cited_title":"Novoseltsev, Spectral geometry of Riemannian submanifolds , PhD The- sis (New Mexico Tech, Socorro, 2005); arXiv:math /0507453 [math.SP]","cited_arxiv_id":null,"evidence_quote":"Provides the symmetrization identities used to simplify the products of derivatives of the phase in Lemma 5 and the explicit coefficients."}],"review_version":1}