{"id":"8f04fb0e-cf9d-4dcb-93a8-960b90450a99","arxiv_id":"2506.22834","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An L2 extension theorem is proved for upper semi-continuous L2-optimal functions using Lebesgue differentiation, yielding new characterizations and integrability results.","lead":"This paper proves a new extension result for a class of functions in several complex variables called L2-optimal functions, using a standard real-analysis tool called Lebesgue differentiation. It then applies that result to characterize a special family of functions and to extend two classical theorems to this broader setting.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.6 invokes Theorem 1.4 on mφ without proving mφ is L2-optimal; Definition 4.4(1) gives only a mixed-weight estimate, so the characterization of plurisubharmonic functions is not established as written.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: the proof of Theorem 4.6 silently assumes that the multiple coarse L2-estimate property for φ yields the L2-optimality of mφ, enabling Theorem 1.4. This is indeed a gap. Definition 4.4(1) is strictly weaker than Definition 1.1 in two ways: the exponential weight on the left is e^{-φ-ϕ} rather than e^{-mφ-ϕ}, and the constant C_m appears. The proof of Theorem 1.4 in Section 3 relies on L2-optimality of the weight φτ = mφ + τ(ϕ) at the step solving ∂̄u = v with the sharp estimate. Without the sharp constant, the new proof would yield an extension estimate with an extra factor C_m, which is still subexponential and may suffice for Theorem 4.6, but this is not stated or proven. The paper's Theorem 4.6 is a central application; the abstract advertises the characterization of plurisubharmonic functions, so the missing argument directly affects the paper's main claim. I agree with the CONDITIONAL verdict: the central extension mechanism appears plausible, but Theorem 4.6's proof is incomplete as written. A corrected proof should either show that mφ is L2-optimal (up to subexponential constants) from Definition 4.4(1) or provide a direct derivation of the multiple coarse L2-extension property. No additional concerns were found that would change this assessment.","tokens_in":1198,"tokens_out":969,"duration_ms":470790,"concrete_test":"Test the implication by attempting to re-derive the estimate used in Theorem 4.6 directly from Definition 4.4(1). For a cylinder z+P, fix f(z)=1 and solve ∂̄ with weight e^{-mφ-ϕ} using the multiple coarse L2-estimate property. If the resulting left-hand side carries e^{-φ-ϕ} instead of e^{-mφ-ϕ}, Theorem 1.4 cannot be applied to mφ. Then check whether Theorem 1.4's proof tolerates replacing the L2-optimality of mφ by the coarse version with constant C_m, log C_m/m → 0; if the extension estimate gains a factor C_m, the proof of Theorem 4.6 still succeeds because (log C_m)/m → 0, and the gap is a missing stated lemma. If the proof fails to tolerate C_m, the characterization is unproven.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The proof of Theorem 4.6 (Section 4) applies Theorem 1.4 with weight mφ. Theorem 1.4 requires φ to be L2-optimal (Definition 1.1), i.e., the ∂̄ estimate ∫|u|²e^{-φ-ϕ} ≤ ∫⟨B^{-1}f,f⟩e^{-φ-ϕ} for every smooth strictly plurisubharmonic ϕ. Definition 4.4(1), the multiple coarse L2-estimate property, gives only ∫|u|²e^{-φ-ϕ} ≤ C_m ∫⟨B^{-1}f,f⟩e^{-mφ-ϕ} with log C_m/m → 0. The left exponent is e^{-φ-ϕ}, not e^{-mφ-ϕ}, and the constant is C_m, not 1. Even after adding a constant to make φ ≤ 0, one obtains at best ∫|u|²e^{-mφ-ϕ} ≤ C_m ∫⟨B^{-1}f,f⟩e^{-mφ-ϕ}, a coarse L2-optimality with subexponential constant, not the constant-1 L2-optimality demanded by Theorem 1.4. No lemma in the paper supplies the missing implication, so the step 'by Theorem 1.4' in the proof of Theorem 4.6 is unjustified. Because Theorem 4.6 is the basis for the paper's headline characterization, this gap is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves an Ohsawa-Takegoshi-type L2 extension theorem (Theorem 1.4) for upper semi-continuous L2-optimal functions, replacing the generalized Siu lemma by a new Lebesgue-type differentiation theorem (Proposition 2.9). It then claims three applications: a characterization of plurisubharmonic functions via the multiple coarse L2-estimate property for strongly upper semi-continuous functions (Theorem 4.6), Skoda's integrability theorem for strongly upper semi-continuous L2-optimal functions (Theorem 4.7), and the strong openness property (Theorem 4.8). The proof of Theorem 1.4 is detailed and the differentiation lemma is new, but the proofs of the applications contain load-bearing gaps.","tokens_in":18899,"tokens_out":14585,"duration_ms":146788,"significance":"If the applications were established, the paper would be a substantial contribution: Theorem 1.4 genuinely relaxes the regularity assumptions in the Ohsawa-Takegoshi extension theorem, and the applications would extend Deng-Ning-Wang's characterizations and Skoda-type results to settings beyond plurisubharmonicity. The new Lebesgue-type differentiation tool (Proposition 2.9) and the careful proof of Theorem 1.4 are clear strengths. However, the current proofs of Theorems 4.6 and 4.7 rely on unjustified uses of Theorem 1.4, so the significance is conditional on repairing those steps.","major_comments":[{"comment":"The step 'by Theorem 1.4' applied to the weight mφ is unjustified. Definition 4.4(1) yields only the mixed-weight estimate ∫|u|²e^{-φ-ϕ} ≤ C_m ∫⟨B^{-1}f,f⟩e^{-mφ-ϕ} with log C_m/m→0; it does not imply that mφ is L2-optimal in the sense of Definition 1.1, which requires the constant 1 and matching exponents e^{-mφ-ϕ} on both sides. No lemma in the paper bridges this gap, so the proof does not establish that the multiple coarse L2-estimate property implies plurisubharmonicity.","section":"Section 4.1, Theorem 4.6 proof"},{"comment":"The application of Theorem 1.4 to produce F∈H^0(U,φ) with the unweighted estimate ∫|F|²e^{-φ} ≤ C|F(z)|²e^{-φ(z)} is not valid. In the k=n case Theorem 1.4 gives only ∫_U |F|²e^{-φ}/(|w-z|^{2n}(log|w-z|²)^2) dλ ≤ C|F(z)|²e^{-φ(z)}. Because the weight is singular, finiteness of this weighted integral does not imply finiteness of ∫|F|²e^{-φ}, so the constructed F need not belong to the Bergman space H^0(U,φ). The subsequent Bergman-kernel argument and the conclusion I(φ)_x=O_x are therefore not established as written.","section":"Section 4.2, Theorem 4.7 proof"}],"minor_comments":[{"comment":"The heading contains a typo: 'theroem' should be 'theorem'.","section":"Section 4.2 heading"},{"comment":"The phrase 'Now we can strength the Proposition 2.3' should read 'strengthen'.","section":"Proposition 2.9 introductory sentence"},{"comment":"The statement of Lemma 4.2 does not explicitly state its conclusion ('then φ is plurisubharmonic'), although the proof provides it.","section":"Lemma 4.2"},{"comment":"The notation I(e^{-φ_1}) should be I(φ_1); the multiplier ideal sheaf depends on the function φ_1, not on the weight e^{-φ_1} as a separate argument.","section":"Theorem 4.8 proof"},{"comment":"The reduction to the case Supp O/I(φ_1) ⊂ {z_1⋯z_n=0} by means of Lemma 2.16 and Lemma 2.17 is only sketched; the use of a modification to achieve normal crossings should be stated explicitly.","section":"Theorem 4.8 proof"},{"comment":"In the displayed estimate 'there is a holomorphic function F∈H^0(U,φ) such that ∫_U |f|²e^{-φ} ≤ C|f(z)|²e^{-φ(z)}', the integrand should involve F rather than f.","section":"Theorem 4.7 proof"}],"recommendation":"major_revision","confidential_remarks":"The main extension theorem appears to be a genuine and carefully proved technical contribution. However, the two principal applications are not established by the arguments given: Theorem 4.6 silently assumes L2-optimality of mφ from a strictly weaker mixed-weight condition, and Theorem 4.7 draws an unweighted Bergman-space conclusion from a weaker singular-weight estimate. Both gaps are load-bearing for the paper's advertised results. These may be fixable, but the fixes require nontrivial additional work, so I recommend major revision rather than acceptance at this stage."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi,\n\nThe real product here is Theorem 1.4: an Ohsawa-Takegoshi-type L2 extension for upper semi-continuous L2-optimal weights. The proof adapts Liu-Xiao-Yang-Zhou's earlier result, replacing the generalized Siu lemma with a Lebesgue-type differentiation argument. That trade is legitimate. Proposition 2.9 is new, correctly proved, and the constant is explicit. I read the proof of Theorem 1.4 line by line and it holds together. This is a genuinely useful technical step, and it probably deserves to stand on its own.\n\nThe problem is Theorem 4.6, the characterization of plurisubharmonic functions via the multiple coarse L2-estimate property. The proof invokes Theorem 1.4 with the weight mφ. But Theorem 1.4 requires mφ to be L2-optimal in the sense of Definition 1.1: a constant-1 estimate with the same weight on both sides. Definition 4.4(1) gives something strictly weaker: a mixed-weight estimate with a constant C_m and log C_m/m → 0. The left side has e^{-φ-ϕ}, the right side e^{-mφ-ϕ}. No argument in the paper bridges that gap. The stress-test note is correct: the move 'by Theorem 1.4' is unjustified. Since the characterization is the first advertised application and the paper's title calls it out, this is a load-bearing flaw, not a cosmetic one. The theorem might still be true, but the proof as written doesn't establish it.\n\nThe other two applications, Skoda's integrability and strong openness, use Theorem 1.4 on φ itself, not on mφ, so they are less exposed. I didn't check every line, but they look like plausible extensions of known arguments. The paper also has the usual scattering of typos and leans on the author's own [16,17] for structural facts like coherence of multiplier ideals. That is acceptable in a working paper, but it means a referee should verify those citations.\n\nBottom line: the extension theorem is a real step for the L2-methods community; the characterization needs a repaired proof or a revised statement. I would not desk-reject this. Send it to a serious referee, with instructions to focus on Section 4.1.\n\nBest.","headline":"A useful L2 extension theorem for upper semi-continuous weights is undercut by an unjustified step in the paper's headline characterization of plurisubharmonicity.","tokens_in":19500,"tokens_out":4444,"would_cite":true,"duration_ms":44583,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32U05","32D15","42B25","14F18"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper establishes an Ohsawa-Takegoshi-type L2 extension theorem for upper semi-continuous L2-optimal functions by replacing the generalized Siu lemma with a Lebesgue-type differentiation theorem, and derives characterizations of…","keywords":["L2-optimal","Ohsawa-Takegoshi extension","upper semi-continuous functions","plurisubharmonic characterization","multiplier ideal sheaf","strong openness","Skoda integrability","Lebesgue differentiation"],"falsifier":"Search for a strongly upper semi-continuous, non-plurisubharmonic function satisfying the multiple coarse L2-estimate property on the unit ball: such a function would disprove Theorem 1.6. A direct way to test the gap is to check whether each multiple mφ is L2-optimal in the sense of Definition 1.1, since the proof of Theorem 4.6 assumes exactly that implication without proving it.","tokens_in":18364,"feed_emoji":"📐","tokens_out":18828,"duration_ms":183055,"temperature":0.7,"pith_summary":"An upper semi-continuous function is called L2-optimal when the sharp weighted ∂-equation estimates that plurisubharmonic weights always satisfy hold for that weight as well. This paper proves that L2-optimality alone yields an Ohsawa-Takegoshi-type extension theorem on Stein products: for almost every slice, every holomorphic function with finite weighted norm extends to the whole product with a logarithmic-pole estimate whose constant depends only on the dimension and the size of the slice factor. The proof replaces the generalized Siu lemma used in earlier extension arguments by a Lebesgue-type differentiation theorem for slice averages. The paper then derives three consequences: a characterization of plurisubharmonic functions as exactly the strongly upper semi-continuous functions satisfying the multiple coarse L2-estimate property, Skoda's integrability theorem for strongly upper semi-continuous L2-optimal functions, and the strong openness property for multiplier ideal sheaves of L2-optimal weights.","feed_headline":"Singular L2-optimal weights still extend holomorphic functions","feed_subtitle":"The extension yields plurisubharmonic characterizations, Skoda integrability, and strong openness for singular weights.","key_machinery":"The load-bearing mechanism is a Lebesgue-type differentiation theorem (Proposition 2.9) for partially integrated functions: for $f \\in L^1(\\mathbb{C}^n)$, for almost every $z'' \\in \\mathbb{C}^k$, the ball averages of $\\int_{\\mathbb{C}^{n-k}} |f(w',w'') - f(w',z'')|\\,d\\lambda_{n-k}(w')$ tend to zero. This replaces the generalized Siu lemma of earlier extension proofs, and it is what controls the slice averages of $e^{-\\varphi}$ and $|f|^2 e^{-\\varphi}$ that appear in the extension estimate. The proof then follows the minimal-$L^2$-solution route: it solves $\\bar\\partial$-equations with weights $\\varphi + \\tau(\\log r)$, where $r = |w''-z''|^2$, and uses a convex auxiliary weight $\\chi$ so that the error terms from the cut-off function are absorbed, leaving exactly the logarithmic denominator in the final estimate.","core_discovery":"The central assertion is Theorem 1.4. On a Stein domain $U \\times D \\subset \\mathbb{C}^{n-k} \\times \\mathbb{C}^k$ with $D$ bounded, let $\\varphi$ be upper semi-continuous and L2-optimal, and assume the support of the quotient sheaf $\\mathcal{O}/\\mathcal{I}(\\varphi)$ lies inside $H \\times D$ for some hypersurface $H \\subset U$. Then for almost every $z'' \\in D$, every holomorphic function $f$ on $U \\times \\{z''\\}$ with finite weighted $L^2$ norm admits an extension $F \\in \\mathcal{O}(U \\times D)$ satisfying $$\\int_{U \\times D} \\frac{|F|^2 $e^{{-\\varphi}}$}{|w''-z''|^{2k}(\\log |w''-z''|^2)^2}\\,d\\lambda_n \\le C \\int_{U \\times \\{z''\\}} |f|^2 $e^{{-\\varphi}}$\\,d\\lambda_{n-k},$$ where $C$ depends only on $k$ and $\\sup_{w'' \\in D} |w''|^2$. All the paper's applications — the plurisubharmonicity characterization, Skoda's integrability theorem, and strong openness — are derived from this extension statement.","pith_inferences":["A natural next step, not claimed by the paper, is to use the same slice-differentiation machinery in other $L^2$ estimates with slice integrals, such as $L^2$ division problems, where the delicate part is controlling averages over a family of slices.","The proof's almost-everywhere quantification suggests that the exceptional slices are tied to non-Lebesgue points of $e^{-\\varphi}$; investigating whether that exceptional set can be described explicitly would sharpen the statement.","The paper's Remark 2.20 records that the analogous inequality for vector-bundle Hermitian metrics is still missing, so extending this theorem to the bundle setting remains an open direction rather than a consequence.","If the constant in Theorem 1.4 can be refined to a sharp form, the approach would be a direct route to the open converse conjecture, since the known equivalence between optimal $L^2$-extension properties and plurisubharmonicity would then apply."],"forward_implications":["For almost every slice $z''$, the weighted Bergman space on $U \\times \\{z''\\}$ embeds into holomorphic functions on $U \\times D$ with the stated logarithmic weight; in the single-point case $k=n$ this gives a weighted point-evaluation bound.","A measurable function on a domain is plurisubharmonic exactly when it is strongly upper semi-continuous and satisfies the multiple coarse L2-estimate property; if it is only upper semi-continuous, it agrees almost everywhere with a plurisubharmonic function.","Any strongly upper semi-continuous L2-optimal function with nonzero multiplier ideal sheaf and Lelong number below 2 at a point has trivial multiplier ideal germ at that point, extending Skoda's integrability theorem.","Multiplier ideal sheaves of L2-optimal functions have the strong openness property: for increasing sequences of L2-optimal functions, $\\mathcal{I}(\\varphi) = \\bigcup_j \\mathcal{I}(\\varphi_j)$, and $\\mathcal{I}(\\varphi) = \\bigcup_{\\varepsilon>0} \\mathcal{I}((1+\\varepsilon)\\varphi)$ whenever the plurisubharmonic envelope is not identically $-\\infty$.","The paper notes that the extension constant obtained is uniform but not sharp enough to resolve the open converse problem that L2-optimality implies plurisubharmonicity for upper semi-continuous functions."],"supporting_citations":[{"why":"Supplies the definition of L2-optimality, the multiple coarse L2-estimate property, and the characterization theorems that Theorem 1.6 relaxes.","marker":"[8]"},{"why":"The Ohsawa-Takegoshi theorem whose extension to upper semi-continuous L2-optimal weights is the paper's main result.","marker":"[19]"},{"why":"Provides the proof scheme of slice reduction, minimal L2 solutions, and the Siu-type lemma that Section 3 adapts.","marker":"[16]"},{"why":"Gives the minimal-solution construction used in Step Two to produce the extension with the logarithmic pole.","marker":"[3]"},{"why":"Supplies the Vitali covering lemma used in the Hardy-Littlewood estimate behind the Lebesgue-type differentiation theorem.","marker":"[13]"},{"why":"One source of the coherence of multiplier ideal sheaves for L2-optimal weights, used to reduce to the case I(φ)=O.","marker":"[18]"},{"why":"The other source of coherence of I(φ) for L2-optimal weights cited in Lemma 2.11.","marker":"[17]"},{"why":"Provides the Guan-Zhou strong openness argument that Section 4.3 adapts for L2-optimal functions.","marker":"[12]"},{"why":"Supplies Lempert's characterization of germs in the union of multiplier ideals used in the proof of Theorem 1.9.","marker":"[15]"},{"why":"Skoda's integrability theorem, which Theorem 1.8 extends to strongly upper semi-continuous L2-optimal functions.","marker":"[20]"}],"fun_headline_variants":["L2-optimal singular weights extend holomorphic functions","Even singular L2-optimal weights allow holomorphic extension","Holomorphic extension survives singular L2-optimal weights","Singular L2-optimal weights: new extension, Skoda, openness"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of Theorem 4.6 silently assumes that a function satisfying the multiple coarse L2-estimate property for every multiple mφ is itself L2-optimal for each mφ, because only then can the extension theorem be applied to the weight mφ; the definition of that property gives only a weaker estimate with an extra constant and a mismatched exponential weight, and no step in the paper closes that gap.","fun_headline_variants_meta":{"raw":{"variants":["L2-optimal singular weights extend holomorphic functions","Even singular L2-optimal weights allow holomorphic extension","Holomorphic extension survives singular L2-optimal weights","Singular L2-optimal weights: new extension, Skoda, openness"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000484,"raw_usage":{"total_tokens":2373,"prompt_tokens":913,"completion_tokens":1460,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":529,"completion_tokens_details":{"reasoning_tokens":1391}},"tokens_in":529,"tokens_out":1460,"duration_ms":16040,"temperature":1.0,"reasoning_tokens":1391,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:00:12.356405+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for a strongly upper semi-continuous, non-plurisubharmonic function satisfying the multiple coarse L2-estimate property on the unit ball: such a function would disprove Theorem 1.6. A direct way to test the gap is to check whether each multiple mφ is L2-optimal in the sense of Definition 1.1, since the proof of Theorem 4.6 assumes exactly that implication without proving it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the definition of L2-optimality, the multiple coarse L2-estimate property, and the characterization theorems that Theorem 1.6 relaxes."},{"cited_title":"Ohsawa, K","cited_arxiv_id":null,"evidence_quote":"The Ohsawa-Takegoshi theorem whose extension to upper semi-continuous L2-optimal weights is the paper's main result."},{"cited_title":"A simple proof of the Ohsawa-Takegoshi extension theorem","cited_arxiv_id":"1105.2430","evidence_quote":"Gives the minimal-solution construction used in Step Two to produce the extension with the logarithmic pole."},{"cited_title":"Heinonen, Lectures on analysis on metric spaces, Universitext","cited_arxiv_id":null,"evidence_quote":"Supplies the Vitali covering lemma used in the Hardy-Littlewood estimate behind the Lebesgue-type differentiation theorem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"One source of the coherence of multiplier ideal sheaves for L2-optimal weights, used to reduce to the case I(φ)=O."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The other source of coherence of I(φ) for L2-optimal weights cited in Lemma 2.11."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Guan-Zhou strong openness argument that Section 4.3 adapts for L2-optimal functions."},{"cited_title":"Lempert, Modules of square integrable holomorphic germs , Trends Math","cited_arxiv_id":null,"evidence_quote":"Supplies Lempert's characterization of germs in the union of multiplier ideals used in the proof of Theorem 1.9."},{"cited_title":"Skoda, Sous-ensembles analytiques d’ordre fini ou infini dans Cn, Bull","cited_arxiv_id":null,"evidence_quote":"Skoda's integrability theorem, which Theorem 1.8 extends to strongly upper semi-continuous L2-optimal functions."}],"review_version":1}