{"id":"08c6d696-17b1-400f-afc0-0ee8b03191d7","arxiv_id":"1908.07879","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For multiple operator integrals, boundedness on the Haagerup tensor product of compact operators implies complete boundedness, and both are equivalent to a pointwise factorization of the symbol.","lead":"This paper proves that certain multilinear maps built from functions of normal operators are automatically 'completely bounded' whenever they are bounded, and it characterizes exactly which functions produce such maps. This gives a clean criterion for multiple operator integrals used in perturbation theory to be well behaved.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 7 (Section 3.2), the unproved bridge between multiple operator integrals and Schur multipliers via equation (8), is load-bearing for Theorem 8's equivalence (ii)⇔(iii); it is likely true but must be proved.","rationale":"The reader's weakest assumption matches my main concern. The paper's central theorem is coherent and the proof strategy is sound: (i)⇒(ii) by the JTT module-map lemma, (iii)⇒(iv) by duality and factorization, (iv)⇒(ii) by finite-dimensional approximation. The only genuinely load-bearing unproved step is Proposition 7. It is a structural identity, not a restatement, and the manuscript itself flags it as omitted. It should be proved or cited with a precise reference. The unresolved '(??)' in Section 4 is a smaller issue that can be fixed by citing the complete metric surjection from Pisier or Effros-Ruan. These are rigor and exposition gaps, not evidence of mathematical error; the conditional verdict is appropriate. If Proposition 7 were independently verified, the paper could be accepted as is.","tokens_in":13517,"tokens_out":24925,"duration_ms":247165,"concrete_test":"Verify Proposition 7 directly: (a) prove (8) for φ=f1⊗...⊗fn and rank-one K_i by computing both sides on the kernel and using ρ_i^{-1} f(A_i)ρ_i = M_{f_i}; (b) show the left and right sides are w*-continuous in φ and norm-continuous in each K_i∈S2, so the identity extends to all φ∈L∞ and all S2 kernels. As a spot check, write out the n=3 case explicitly and compare the integral kernels. If this succeeds, the bridge (ii)⇔(iii) is sound.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central risk is Proposition 7 in Section 3.2. It asserts the identification Γ_{A1,...,An}(φ)(ρ1K1ρ2^{-1},...,ρ_{n-1}K_{n-1}ρn^{-1}) = ρ1 Λ(φ)(K1,...,K_{n-1}) ρn^{-1}, i.e. equation (8). The proof is omitted: 'The proof is similar and we leave it to the reader.' The equivalence (ii)⇔(iii) in Theorem 8 reduces multiple operator integrals to Schur multipliers exactly through this identity; without it, the restriction argument in (ii)⇒(iii) and the equality of norms collapse. The identity is almost certainly correct: it holds for elementary φ=f1⊗...⊗fn because ρ_i intertwines f(A_i) with multiplication by f on L2(λAi), and both sides extend by w*-continuity in φ for fixed S2 kernels. But the paper does not supply this argument, and a chain of equivalences whose pivotal bridge is unstated is not fully rigorous as written. A secondary missing detail is the unresolved reference '(??)' used in Section 4 to justify the complete metric surjection q:S1(L2(λAi))→L1(λAi).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies multiple operator integrals Γ_{A_1,...,A_n}(φ) acting on Haagerup tensor products of compact operator spaces over a separable Hilbert space. Its main result, Theorem 8, gives four equivalent conditions for such an integral to be bounded: (i) boundedness of the extension to S∞(H) h⊗ ... h⊗ S∞(H); (ii) complete boundedness of this extension; (iii) complete boundedness of the associated continuous multilinear Schur multiplier Λ(φ); and (iv) a pointwise factorization φ(t_1,...,t_n)=⟨a_1(t_1),[a_2(t_2)...a_{n−1}(t_{n−1})](a_n(t_n))⟩ with a_1∈L∞(λ_{A_1};H_1), a_n∈L∞(λ_{A_n};H_{n−1}), and a_i∈L∞_σ(λ_{A_i};B(H_i,H_{i−1})) for separable Hilbert spaces H_i. The proof proceeds through a module-map argument using Juschenko–Todorov–Turowska's Lemma 3.3, a restriction to spectral subspaces that identifies Γ with the Schur multiplier Λ, a duality/factorization argument via L∞_σ spaces, and a finite-rank approximation in the converse direction. The paper also states equality of the bounded, completely bounded, and factorization norms.","tokens_in":13725,"tokens_out":15882,"duration_ms":137889,"significance":"The result is a natural extension of the Juschenko–Todorov–Turowska characterization of continuous multilinear Schur multipliers to general normal operators, and the automatic passage from boundedness to complete boundedness is a useful theorem. The norm formula relating ‖Γ‖, ‖Γ‖_cb, ‖Λ‖_cb, and the infimum over factorizations is a strong and clean statement. The overall strategy is coherent and uses standard operator-space tools; in particular, the finite-rank approximation in (iv)⇒(ii) is explicit and gives a constructive cb bound. The main reservation is that one structural bridge, Proposition 7, is not proved in the manuscript and is essential for (ii)⇔(iii); once that proof is supplied, the result appears credible.","major_comments":[{"comment":"Proposition 7 is the only bridge between the multiple operator integral Γ and the multilinear Schur multiplier Λ, and it is used in the proof of (ii)⇒(iii) to restrict Γ to S∞(H_{i+1},H_i) and to identify the restriction with Λ(φ), including the cb-norm inequality ‖Λ(φ)‖_cb ≤ ‖Γ_{A_1,...,A_n}(φ)‖_cb. The proof is omitted with the remark 'The proof is similar and we leave it to the reader.' This is a load-bearing step: without it the equivalence (ii)⇔(iii) in Theorem 8 is not established as written. The author should supply a complete proof, for example by checking (8) for elementary tensors φ=f_1⊗...⊗f_n, where it follows from ρ_i f(A_i)ρ_i^{-1}=M_{f_i}, and then extending by w*-continuity in φ for fixed S2 kernels. If a proof already appears in [4, Proposition 9] for the analogous two-operator case, a precise reference to the exact statement would also be acceptable, but it should be explicit.","section":"Section 3.2, Proposition 7 (equation (8))"},{"comment":"The construction of the quotient map q:S1(L2(Ω))→L1(Ω) is invoked with an unresolved cross-reference 'by (??)'. This map is then used to form Q=q_1⊗...⊗q_n, whose kernel N is identified via Proposition 3(iv), and the vanishing of u on N is what produces the factor v. Although the statement that q is a complete metric surjection is plausible and likely standard, the manuscript does not provide the supporting statement. Please replace the placeholder by the correct equation and either prove the complete metric surjectivity or cite a theorem where it appears.","section":"Section 4, proof of (iii)⇒(iv), page 9"}],"minor_comments":[{"comment":"The displayed domain of Λ(φ) on page 1 is written as S2(L2(Ω_{n−1}),L2(Ω_n)) × ... × S2(L2(Ω_1),L2(Ω_2)), which is inconsistent with the convention used after (7) and in Proposition 7, where the first argument lies in S2(E2,E1). Please correct this typo.","section":"Introduction and Section 3.2, definition of Λ(φ)"},{"comment":"In the definition of σ_i^N, the displayed formula writes σ_i^N(X_1) but should read σ_i^N(X_i); also σ_1^N should be A_1^N π_N(X_1). This is a typographical error, and the intended argument is clear from the preceding formula.","section":"Section 4, proof of (iv)⇒(ii)"},{"comment":"In the sentence 'the element [a^i_{kl}]_{1≤k,l≤N} ∈ M_N(L∞(λB)) has a norm less than ‖a_i‖∞', the symbol λB should be λA_i.","section":"Section 4, proof of (iv)⇒(ii)"},{"comment":"After obtaining u = v∘Q, the complete boundedness of v follows from the complete metric surjectivity of Q; this step is implicit and should be stated explicitly for the reader.","section":"Section 4, proof of (iii)⇒(iv)"}],"recommendation":"major_revision","confidential_remarks":"The missing proof of Proposition 7 appears to be straightforward and the result likely correct, but the manuscript as submitted leaves a central bridge to the reader. I recommend major revision rather than rejection, and I would ask the editor to require a full proof or precise citation for Proposition 7 and a correction of the unresolved cross-reference in Section 4. The paper fits the scope of math.FA well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper proves the expected characterization: a multiple operator integral on the Haagerup tensor product of compact operators is bounded if and only if it is completely bounded, and if and only if its symbol admits the pointwise factorization in (9). That is a real result, and it is new for general normal operators. The previous work [12] only covered continuous multilinear Schur multipliers on L2-spaces, so this genuinely extends the scope. The paper also supplies a new proof of [12, Theorem 3.4] along the way, which is a nice bonus.\n\nThe overall strategy is coherent. The proof chain (i)⇒(ii)⇒(iii)⇒(iv)⇒(ii) is assembled from standard operator-space ingredients: the module-map lemma of Juschenko–Todorov–Turowska, the duality of Proposition 4, the Haagerup tensor product calculus, and finite-rank approximation. The finite-rank step in (iv)⇒(ii) is particularly clean—it builds the approximating symbols φN and uses Wittstock's theorem to control the cb norm. The reliance on the author's own construction in [4] is legitimate; the target statement is not being assumed. Circularity is not a real concern here.\n\nThe soft spots are real but fixable. The main one is Proposition 7 in Section 3.2. This is the bridge that identifies Γ with Λ via the unitaries ρi, and it is exactly what makes (ii)⇔(iii) work and the norm equality hold. The paper dismisses it with \"the proof is similar and we leave it to the reader.\" That is not good enough for a load-bearing step. It is almost certainly true—for elementary tensors it follows from the intertwining of ρi with functional calculus, and the general case should follow by w*-continuity—but the argument needs to appear in the paper. A referee should insist on this.\n\nThere is also an unresolved cross-reference \"(??)\" in Section 4, used to justify that the map q: S1(L2(Ω)) → L1(Ω) is a complete metric surjection. That is a minor issue but should be cleaned up. There are a few typos, like \"Kn1\" in (7), which are harmless.\n\nOverall, the central argument holds up. This paper deserves a serious referee and likely publication after revision. The right recommendation is to send it to peer review, with the clear request that Proposition 7 be proved in full and the cross-reference fixed.","headline":"Theorem 8 is the right result and the proof is mostly sound, but Proposition 7—the unproved bridge to Schur multipliers—is load-bearing and needs to be written down before the paper is fully rigorous.","tokens_in":14293,"tokens_out":1497,"would_cite":true,"duration_ms":16536,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L07","47B49"],"pacs":[],"model":"deepseek-v4-flash","headline":"For any finite family of normal operators, a multiple operator integral is bounded on the Haagerup tensor product of compact-operator spaces if and only if its symbol admits a measurable vector-valued factorization, and in that case the…","keywords":["multiple operator integrals","complete boundedness","Haagerup tensor product","multilinear Schur multipliers","factorization of symbols","operator spaces","normal operators","perturbation theory"],"falsifier":"The unproved bridge is Proposition 7, so the shortest check is computational: for $n=2$, take $H=\\mathbb{C}^2$, $A_1=\\mathrm{diag}(0,1)$, $A_2=\\mathrm{diag}(1,0)$, and a symbol such as $\\varphi(s,t)=st$; compute $\\Lambda(\\varphi)$ from the kernel formula in Section 3.2 and $\\Gamma_{A_1,A_2}(\\varphi)(X)$ from the definition for an explicit $X$. The claim asserts the two sides of equation (8) are equal for every $X$ and every symbol, so a single mismatch disproves the bridge; if the example agrees, repeat the check with $\\varphi=\\mathbf{1}_{E\\times F}$ for a Borel rectangle.","tokens_in":13243,"feed_emoji":"","tokens_out":14113,"duration_ms":120348,"temperature":0.7,"pith_summary":"The paper sets out to characterize exactly when a multiple operator integral, the basic multilinear device used in perturbation theory to express differences and derivatives of functions of normal operators, extends from Hilbert-Schmidt inputs to a bounded map on the Haagerup tensor product of compact-operator spaces. It establishes that boundedness is the same as complete boundedness, and that both are equivalent to a concrete factorization of the symbol: the symbol must equal a pointwise inner product of a vector-valued function with a chain of operator-valued functions. The theorem also gives that the norm of the integral, its completely bounded norm, and the norm of the associated continuous multilinear Schur multiplier all coincide with the infimum product of the factor norms. This is the operator-integral analogue of the known factorization characterization for continuous multilinear Schur multipliers.","feed_headline":"Operator integrals are bounded exactly when their symbols factor","feed_subtitle":"An inner-product factorization of the symbol is exactly what makes the integral map completely bounded on compact operators.","key_machinery":"The load-bearing object is the Haagerup tensor product $S_\\infty(H)\\otimes^h\\cdots\\otimes^h S_\\infty(H)$ (with $n-1$ factors), the operator-space tensor product that detects complete boundedness of multilinear maps. The proof works in three steps: first it shows $\\Gamma_{A_1,\\dots,A_n}(\\varphi)$ is a multilinear module map over the commutants of the von Neumann algebras generated by $A_1$ and $A_n$, which forces boundedness to imply complete boundedness; second, using the unitaries $\\rho_i$ that identify $L^2(\\lambda_{A_i})$ with spectral subspaces of $H$, it identifies the restriction of $\\Gamma$ with the continuous multilinear Schur multiplier $\\Lambda(\\varphi)$; third, complete boundedness of $\\Lambda$ is converted, through the complete isometry $CB(E,B(H,K)) = ((K_c)^*\\otimes^h E\\otimes^h H_c)^*$, into a functional on a Haagerup tensor product of trace-class spaces, whose factorization yields the pointwise inner-product representation of the symbol.","core_discovery":"The central result, Theorem 8, states that for any $n\\ge 2$, any normal operators $A_1,\\dots,A_n$ on a separable Hilbert space $H$, and any bounded Borel symbol $\\varphi$ on the product of their spectra, the multiple operator integral $\\Gamma_{A_1,\\dots,A_n}(\\varphi)$ extends to a bounded map on the Haagerup tensor product of $n-1$ copies of $S_\\infty(H)$ if and only if it extends completely boundedly, if and only if the associated continuous multilinear Schur multiplier $\\Lambda(\\varphi)$ is completely bounded, if and only if $\\varphi$ factors almost everywhere as $\\varphi(t_1,\\dots,t_n)=\\langle a_1(t_1),[a_2(t_2)\\cdots a_{n-1}(t_{n-1})](a_n(t_n))\\rangle$, where $a_1$ and $a_n$ take values in separable Hilbert spaces and the intermediate $a_i$ are weak-* measurable operator-valued functions on the spectral measures of the corresponding $A_i$. In that case the norm of $\\Gamma$, its completely bounded norm, and the norm of $\\Lambda$ all equal the infimum of $\\|a_1\\|_\\infty\\cdots\\|a_n\\|_\\infty$ over all such factorizations.","pith_inferences":["Because the proof uses the Haagerup tensor product and $S_\\infty$, it does not settle the analogous question for Schatten classes $S_p$ with $1\\le p<\\infty$; a natural extension would test whether some analogous factorization characterizes boundedness there.","The separability assumptions on $H$ and on the intermediate Hilbert spaces are used to apply structural theorems, but the mechanism itself looks measure-theoretic, so the same statement may hold for non-separable spaces or for arbitrary families of commuting normal operators; this is an extrapolation, not a claim of the paper.","The factorization identity isolates a purely spectral condition on $\\varphi$, independent of $H$; this suggests that one could check boundedness numerically by searching for vector-valued functions $a_i$ on the spectra, which may be useful for concrete operator-function estimates."],"forward_implications":["Boundedness of a multiple operator integral on the Haagerup tensor product automatically implies complete boundedness, with equal norms.","The norm of the integral is the infimum over factorizations of the product of the factor norms, so norm estimates become an optimization over vector-valued functions.","The same factorization condition characterizes complete boundedness of the associated continuous multilinear Schur multiplier, so the two subjects are identified by the unitary bridge.","For perturbation theory, the theorem supplies a concrete criterion for when the multilinear operator integrals representing higher-order differences of operator functions extend to compact operators."],"supporting_citations":[{"why":"Defines the multiple operator integrals used here and supplies their w*-continuity and isometric extension theorem, which anchors the construction.","marker":"[4]"},{"why":"Provides the scalar-valued spectral measures and cyclic vectors behind the unitaries that identify spectral subspaces, along with the von Neumann algebra facts used in the module-map step.","marker":"[7]"},{"why":"Gives the identification between weak-* measurable vector-valued L-infinity spaces and bounded operators on L1, used to turn abstract factorizations into measurable coefficient functions.","marker":"[9]"},{"why":"Supplies the continuous multilinear Schur multiplier characterization that is generalized here, together with the module-map lemma used for automatic complete boundedness.","marker":"[12]"},{"why":"Provides the Haagerup tensor product factorization theorem and the operator-space tools used throughout the proof.","marker":"[16]"},{"why":"Provides the duality complete isometry for completely bounded maps and the injectivity and projectivity properties of the Haagerup tensor product.","marker":"[19]"}],"fun_headline_variants":["Symbol factorization is the key to bounded operator integrals","Operator integrals: bounded iff symbols factor","Factorable symbols yield exactly the bounded operator integrals","Boundedness of operator integrals characterized by symbol factorization","Integral maps on compact operators: bounded iff symbol factors"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole chain of equivalences depends on an identification between the operator integral and a certain kernel-smoothing multiplier map, an identification that is asserted without proof in Proposition 7; if that identification fails for some tuple of operators and symbols, the chain breaks.","fun_headline_variants_meta":{"raw":{"variants":["Symbol factorization is the key to bounded operator integrals","Operator integrals: bounded iff symbols factor","Factorable symbols yield exactly the bounded operator integrals","Boundedness of operator integrals characterized by symbol factorization","Integral maps on compact operators: bounded iff symbol factors"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000976,"raw_usage":{"total_tokens":4106,"prompt_tokens":864,"completion_tokens":3242,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":480,"completion_tokens_details":{"reasoning_tokens":3170}},"tokens_in":480,"tokens_out":3242,"duration_ms":485411,"temperature":1.0,"reasoning_tokens":3170,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:54:44.135266+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The unproved bridge is Proposition 7, so the shortest check is computational: for $n=2$, take $H=\\mathbb{C}^2$, $A_1=\\mathrm{diag}(0,1)$, $A_2=\\mathrm{diag}(1,0)$, and a symbol such as $\\varphi(s,t)=st$; compute $\\Lambda(\\varphi)$ from the kernel formula in Section 3.2 and $\\Gamma_{A_1,A_2}(\\varphi)(X)$ from the definition for an explicit $X$. The claim asserts the two sides of equation (8) are equal for every $X$ and every symbol, so a single mismatch disproves the bridge; if the example agrees, repeat the check with $\\varphi=\\mathbf{1}_{E\\times F}$ for a Borel rectangle.","supporting_citations":[{"cited_title":"When do triple operator integrals take value in the trace class?","cited_arxiv_id":"1706.01662","evidence_quote":"Defines the multiple operator integrals used here and supplies their w*-continuity and isometric extension theorem, which anchors the construction."},{"cited_title":"Conway, A Course in Operator Theory , Graduate Studies in Mathematics, Vol","cited_arxiv_id":null,"evidence_quote":"Provides the scalar-valued spectral measures and cyclic vectors behind the unitaries that identify spectral subspaces, along with the von Neumann algebra facts used in the module-map step."},{"cited_title":"Dunford, B","cited_arxiv_id":null,"evidence_quote":"Gives the identification between weak-* measurable vector-valued L-infinity spaces and bounded operators on L1, used to turn abstract factorizations into measurable coefficient functions."},{"cited_title":"Juschenko, I","cited_arxiv_id":null,"evidence_quote":"Supplies the continuous multilinear Schur multiplier characterization that is generalized here, together with the module-map lemma used for automatic complete boundedness."},{"cited_title":"Pisier, Introduction to operator space theory , London Mathematical Society, Lecture note Series 294, 2003","cited_arxiv_id":null,"evidence_quote":"Provides the Haagerup tensor product factorization theorem and the operator-space tools used throughout the proof."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the duality complete isometry for completely bounded maps and the injectivity and projectivity properties of the Haagerup tensor product."}],"review_version":1}