{"id":"5fef8805-7aff-49f0-8ece-1bb2c9cca408","arxiv_id":"2508.19446","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"New partial ranges are proven for maximal estimates of wave operators acting on orthonormal families, via Wolff-type sphere-intersection bounds.","lead":"This paper proves new maximal-in-time estimates for the wave equation applied to infinitely many orthonormal initial data at once, in dimensions two, three, and four. The results are partial progress toward a conjectured sharp range, and the novelty is a geometric analysis of how thickened light cones intersect.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.4 rests on unproved cone-intersection identities (6.10); a wrong constant or missing 2^{-k} term would destroy the Δ bound and the σ=2/7 gain.","rationale":"The reader identified the geometric intersection bounds as the weakest assumption, and I agree the n=2 refinement is the most delicate. In good faith, the n=3,4 parts appear sound: the reduction chain (§3–§4) is internally consistent, the interpolation exponents match, and the derivation of (5.7) from Lemma 5.2 is a straightforward slicing argument. The omitted proof of Lemma B.1 is not a real concern—that lemma is an elementary Taylor-expansion estimate and can be filled in. The genuinely load-bearing gap is the unproved system (6.10), which is used to compute Δ and the threshold m* in the n=2 proof. If (6.10) is not exact, the claimed improvement from σ=1/4 to σ=2/7 collapses. This does not force a REJECT because the identities may well be correct and fillable; it does warrant keeping the CONDITIONAL verdict, with the condition being a rigorous derivation of (6.10). The proposed concrete test—re-deriving (6.10) and re-checking the W1/W2 estimates—is the minimal check that would settle the concern. I therefore recommend no change to the reader's verdict; the reader's 'partial progress' assessment remains appropriate conditional on that geometric identity.","tokens_in":22144,"tokens_out":29801,"duration_ms":264873,"concrete_test":"Independently derive (6.10) from the cone definitions: for w2-w1 ∈ V_{k,l}(0,2^{-k}j), write the conditions ||x1-x2|-r1|≤2^{-k}, ||x1-x2|-r2|≤2^{-k} at slice m, and solve for r1(m), r2(m). Verify that the exact identities hold with no additional O(2^{-k}) corrections. If a correction appears, recompute Δ and re-run the W1/W2 bounds in Section 6.2; if the resulting |W| no longer satisfies (6.5), Theorem 1.4 fails and the n=2 claim reduces to the σ=1/4 bound of Section 5.4.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The n=2 refinement in Section 6 (Proposition 6.1, Theorem 1.4) depends entirely on the tangential/transversal decomposition and the bound (6.5) for |W(w1,w2)|. The derivation of (6.5) in Section 6.2 uses the asserted identities (6.10): |x1-x2| = r1(m)+r2(m)-2^{-k+3/2}m and |x1-x2| = r1(m)-r2(m)+2^{-k+1/2}j. These are stated with no derivation. From them, the paper computes Δ = (2^{-k+3/2}m)(2^{-k+1/2}j) and obtains the threshold m* satisfying m ≥ 4|x1-x2|/(2^{-k}j), which is crucial for the W1/W2 split. If the correct geometric relation has a sign change, a different constant, or an additional O(2^{-k}) error term, then the lower bound Δ ≥ C2^{-2k}2^{θ(k-l)}m fails for small j. In that case the transversal estimate (6.5) would degrade, the bootstrap optimization θ=1/7 would not close, and Theorem 1.4 would reduce to the weaker σ=1/4 bound already noted in Section 5.4. The n=3,4 theorems are less exposed: they rely on Lemma 5.2, whose proof via Lemma B.1 is omitted but the lemma is elementary and the slicing argument in Section 5.3 is internally consistent. Thus the single most load-bearing unresolved point is the correctness of (6.10).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies maximal-in-time estimates for the half-wave propagator applied to orthonormal systems of initial data. For n=3,4 the authors prove (1.3) for β < min{(2n-1)/(2(n-2s)), (2n-3)/(2n-1-4s)} (Theorem 1.3), and for n=2 for β < min{9/(14(1-s)), 5/(12-14s)} (Theorem 1.4). The proof proceeds through the Frank-Sabin duality principle, reducing the β=2 frequency-localized estimate (3.5) to a Schatten-2 bound (3.7), then to the bilinear cone-intersection estimate I_{k,l} (Proposition 4.1). For n=3,4, the intersection estimate is obtained from Wolff-type sphere intersection lemmas (Lemmas 5.1, 5.2) via level-set slicing, yielding (5.7) and σ=1/2. For n=2, a refinement splits the cone difference into transversal and tangential parts controlled by a parameter θ, and a bootstrap gives σ=2/7. Corollary 1.5 derives almost-everywhere convergence of the density of Schatten-Sobolev operators.","tokens_in":22510,"tokens_out":25443,"duration_ms":225102,"significance":"Assuming the geometric estimates are correct, the results are new and nontrivial: the orthonormal maximal estimate has not been previously treated for the wave equation, and the n=2 bootstrap is an interesting technique. The reduction framework is well organized and the n=3,4 case is convincing. The paper is honest about the partial progress relative to Conjecture 1.2. However, the n=2 result depends on unproved identities (6.10), and the appendix leaves parts of Lemma 5.2's proof as a sketch. These need to be fixed before publication.","major_comments":[{"comment":"The two identities |x1-x2| = r1(m)+r2(m)-2^{-k+3/2}m and |x1-x2| = r1(m)-r2(m)+2^{-k+1/2}j are stated without derivation. They are used to compute Δ = (2^{-k+3/2}m)(2^{-k+1/2}j), to define the threshold m* in (6.11), and to obtain the lower bound Δ ≥ 2^{-k}(r1(m)+r2(m)) for m ≥ m*. This is the only geometric input that produces the σ = 2/7 gain in Theorem 1.4. If the identities hold only up to O(2^{-k}) errors (which the thickness of the level sets suggests), the threshold and the lower bound can fail for small j, and (6.5) degrades to the trivial tangential bound, reducing Proposition 6.1 to the σ = 1/4 result of §5.4. Please give a complete derivation of (6.10), including the exact choice of c0 and the definitions of m and j, or replace the identities by rigorous inequalities with error terms and verify that Δ ≥ C2^{-2k}2^{θ(k-l)}m continues to hold.","section":"§6.2, Eqs. (6.10) and (6.11)"},{"comment":"Lemma B.1 is stated without proof (\"We omit the detailed proof\") and Lemma B.2 is justified in a single paragraph that relies on an unstated containment claim. Lemma 5.2, used for all n=3,4 results, depends on Lemma B.2. While the statements are plausible, the proof should be expanded; in particular, derive (B.3) explicitly and justify the interval length for θ. As written, this is a gap in a central estimate.","section":"Appendix B, Lemma B.2 and Lemma B.1"},{"comment":"The passage from Jtang to the displayed bound is too compressed. The Hölder application (apparently with exponents 4/3 and 4), the translation invariance, and the summation bound ∑_{j≤C2^{(k-l)θ}} 1 ≲ 2^{(k-l)θ} should be shown. The resulting exponent 2^{k/2}2^{2l}Jtang ⪅ 2^{(k-l)θ}∥h∥^{5/4}∥h*χ∥^{3/4} is load-bearing for the bootstrap; it should be independently verifiable.","section":"§6.1, derivation of (6.9)"}],"minor_comments":[{"comment":"The signs in the two displays are inconsistent: the text says |K_{k,l}|² ≲ 2^{(n+1)k}2^{-(n-1)l}, while (4.4) has 2^{(n-1)l}; the latter is needed for the cancellation with Proposition 4.1. Please correct.","section":"§4, Eq. (4.4)"},{"comment":"The second exponent should be (n+1)/4 - (n-1)/(2r), not (n+1)/r - ...; likewise in §A.3 it is written with n-2 in the numerator. These typos obscure the endpoint s=1/2 at r=2.","section":"Appendix A, Prop. A.1 and §A.3"},{"comment":"The symbol j is used both for a scalar index in J_j and for a multi-index in the set J; please disambiguate.","section":"§2.1"},{"comment":"The exponent \"2/4+3/4\" should be 5/4.","section":"§6.1, before (6.9)"},{"comment":"The phrase \"orthonormal basis in H^s\" should be \"orthonormal family in L^2 normalized in H^s\".","section":"§3.1"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about (6.10) is well placed; I believe this is the main obstacle. I would not reject outright, because the n=3,4 part appears sound and the n=2 part may be repairable, but the authors should provide a full proof or modify the claim. The paper is otherwise a good fit for the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kinoshita–Ko–Shiraki have a real result here: new β ranges for maximal estimates of wave operators on orthonormal systems in dimensions 2–4, with a genuinely novel geometric argument. It is partial progress on Conjecture 1.2, not the full conjecture, but the ranges are nontrivial and the technique is new.\n\nThe paper does several things well. The reduction from the orthonormal maximal estimate to a Schatten C2 estimate via the Frank–Sabin duality principle, then to the bilinear cone-intersection estimate (1.7), is clean and carefully laid out. The use of Wolff's sphere-intersection lemma to bound the vertical projection of the intersection of two thickened cones is new in this problem. For n=3,4 the argument yields σ=1/2 in (3.5), and the proof looks solid. The n=2 refinement, with the conic decomposition and bootstrap in θ ending at θ=1/7, is the most interesting part; it genuinely improves the naive σ=1/4.\n\nThe soft spots are real but not disqualifying. First, the proof of Lemma B.1—the elementary angular measure bound—is omitted with 'we omit the detailed proof.' That is easy to fill, but in a referee situation you want to see it. Second, the geometric identities (6.10) for the n=2 tangent case are asserted without derivation. These are load-bearing: the Δ bound and the threshold m* come from them, and the θ=1/7 gain depends on that. The identities look like standard tangent-circle approximations, and I have no reason to think they're false, but a wrong constant or a missing error term would degrade Theorem 1.4 to the weaker σ=1/4. That is a specific, checkable point, and the authors should supply the derivation. Third, Section 4.1 leaves some 2^{εk} losses implicit; that's acceptable but worth spelling out.\n\nI saw no evidence of circularity or fitting. The necessity conditions in Section 2 are independent, and the sufficiency proofs do not invoke the conjecture. The self-citations [2,3] are used for the reduction framework, which is legitimate.\n\nBottom line: this is a serious paper for harmonic analysts working on maximal estimates and orthonormal systems. It deserves a proper referee. I would send it to peer review with a request to fill the n=2 gap and prove Lemma B.1; if the identities in (6.10) check out, the n=2 theorem stands.","headline":"New partial progress on maximal estimates for orthonormal wave systems; the n=2 refinement is clever but rests on an unproved geometric identity.","tokens_in":23071,"tokens_out":3345,"would_cite":true,"duration_ms":31010,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35L05","42B25"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves new maximal-in-time estimates for wave equations with orthonormal initial data in dimensions two, three, and four, on a range of exponents better than what elementary interpolation gives.","keywords":["maximal estimates","wave equation","orthonormal systems","Schatten spaces","cone intersections","sphere intersection","pointwise convergence","Sobolev regularity"],"falsifier":"In dimension 2, take two copies of the thin cone V_{k,l}, translate one by (0, 2^{-k}j) with small j, and compute the Lebesgue measure of the projection of their intersection onto R^2. The claimed bound (6.6) predicts the result is at most C 2^{-k}2^{-2l}|x1−x2|^{-1/2}; measuring a larger power of 2^{-k} would falsify the estimate. The identities (6.10) for the sliced radii can also be checked directly by elementary trigonometry for a near-tangent pair, which would confirm or refute the geometric picture.","tokens_in":22010,"feed_emoji":"🌊","tokens_out":6036,"duration_ms":53494,"temperature":0.7,"pith_summary":"This paper establishes maximal-in-time estimates for the half-wave equation when the initial data form an orthonormal family, in dimensions two, three, and four. The central estimate controls the supremum over small times of the weighted sum of squared evolved data in L^β by the ℓ^β norm of the weights, for a range of β depending on the Sobolev regularity s. Previously such orthonormal maximal estimates were only known in endpoint or trivial cases; the paper makes partial progress toward the conjectured optimal range. The proof reduces the problem to a bilinear integral over intersections of thin cones and bounds its decay using precise sphere-intersection measures. In two dimensions, a tangential/transversal decomposition and a self-similar bootstrapping argument yield an additional gain.","feed_headline":"Maximal wave bounds extend to orthonormal data in dimensions 2–4","feed_subtitle":"A cone-intersection geometry proof sharpens the known range of β, with extra gains in the plane.","key_machinery":"The central object is the set W(w1,w2)=Proj_{R^n}(V_{k,l}(w1)∩V_{k,l}(w2)), the vertical projection of the intersection of two 2^{-k}-thickened one-sided cones of height 2^{-l}. The argument estimates the Lebesgue measure |W(w1,w2)| by slicing the cones into level sets of height 2^{-k} and applying sphere-intersection measure lemmas (Lemmas 5.1 and 5.2). In two dimensions, a further decomposition into tangential and transversal cases, described by identities (6.10) for the sliced radii, yields two competing bounds that are balanced by the parameter θ.","core_discovery":"The paper proves (1.3) for n=3,4 whenever β < min{(2n−1)/(2(n−2s)), (2n−3)/(2n−1−4s)}, and for n=2 whenever β < min{9/(14(1−s)), 5/(12−14s)}. These follow from a frequency-localized β=2 estimate with decay exponent σ=1/2 for n=3,4 and σ=2/7 for n=2, interpolated with the trivial β=1 and β=∞ endpoints. The novelty is the treatment of the β=2 case: after dualizing to a Hilbert–Schmidt bound, the problem becomes a bilinear estimate involving the indicator of a thin cone, and the improvement comes from measuring the vertical projection of the intersection of two such cones. In n=3,4 the paper uses a higher-dimensional sphere-intersection lemma to obtain a bound of the form (5.7); in n=2 it refin","pith_inferences":["Editorial inference: if the geometric measure bound (5.7) could be upgraded to the conjectured σ=1, the same machinery would likely give the optimal ω range in Conjecture 1.2, and possibly orthonormal Strichartz estimates for the wave equation.","Editorial inference: the dependence on the identities (6.10) suggests a purely two-dimensional phenomenon tied to the topology of circle intersections in the plane; a direct numerical sampling of near-tangent configurations could test whether the bound in (6.6) is genuinely sharp.","Editorial inference: the optimization θ=1/7 might be an instance of a general cone-separation tradeoff whose optimal value could be derived by a simpler two-point calculation, and analogously computed for n=3 and n=4 to see whether σ=1/2 is improvable.","Editorial inference: the reduction from orthonormal estimates to Hilbert–Schmidt kernel bounds plus cone intersection measures is likely reusable for other families of oscillatory integral operators, not just wave equations."],"forward_implications":["If the estimates are correct, the density of a Schatten-class operator evolved by the wave flow converges pointwise at t=0 for the parameter ranges in Corollary 1.5.","The results give the first orthonormal-system maximal estimates for wave equations beyond the trivial endpoint, a partial confirmation of Conjecture 1.2 in low dimensions.","The cone-intersection measure bounds may transfer to other operators with conical kernels, such as Klein–Gordon equations, whose maximal estimates the paper shows follow from the same geometric analysis.","The two-dimensional tangential/transversal bootstrap suggests a general mechanism for improving σ beyond the straightforward sphere-intersection bound, though the paper notes that for n≥3 the transversal intersections dominate and the same trick does not apply."],"supporting_citations":[{"why":"Supplies the base sphere-intersection measure lemma (Lemma 5.1) that the cone-projection bounds are built from.","marker":"[22]"},{"why":"Extends the sphere-intersection estimate to n≥3 and yields Lemma 5.2, used for the n=3,4 cases.","marker":"[8]"},{"why":"The duality principle (Lemma 3.2) converts the orthonormal L^β estimate into a Hilbert–Schmidt kernel estimate, the core reduction step.","marker":"[16]"},{"why":"Provides the Schatten–Sobolev density framework and the interpolation-and-reduction strategy that this paper adapts to the wave equation.","marker":"[3]"},{"why":"Provides the maximal-in-time orthonormal estimate method and the pointwise convergence consequence stated in Corollary 1.5.","marker":"[4]"},{"why":"The classical single-datum maximal estimate used as the β=1 endpoint.","marker":"[10]"},{"why":"The sharp single-datum maximal estimate, also used as the β=1 benchmark and comparison.","marker":"[21]"}],"fun_headline_variants":["Wave maximal estimates now cover orthonormal data in 2–4D","Cone geometry yields maximal wave bounds for orthonormal data","Maximal wave bounds extend to orthonormal families in 2–4D","New geometric proof improves maximal wave estimates for orthonormal data"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The whole argument rests on the claim that the vertical projection of two intersecting 2^{-k}-thickened cones has the size stated in (5.7) and (6.5); if that size is even slightly larger in some tangent configuration, the central bilinear estimate loses its decay and the main theorems collapse.","fun_headline_variants_meta":{"raw":{"variants":["Wave maximal estimates now cover orthonormal data in 2–4D","Cone geometry yields maximal wave bounds for orthonormal data","Maximal wave bounds extend to orthonormal families in 2–4D","New geometric proof improves maximal wave estimates for orthonormal data"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001216,"raw_usage":{"total_tokens":4798,"prompt_tokens":658,"completion_tokens":4140,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":402,"completion_tokens_details":{"reasoning_tokens":4062}},"tokens_in":402,"tokens_out":4140,"duration_ms":27930,"temperature":1.0,"reasoning_tokens":4062,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T15:51:50.968529+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In dimension 2, take two copies of the thin cone V_{k,l}, translate one by (0, 2^{-k}j) with small j, and compute the Lebesgue measure of the projection of their intersection onto R^2. The claimed bound (6.6) predicts the result is at most C 2^{-k}2^{-2l}|x1−x2|^{-1/2}; measuring a larger power of 2^{-k} would falsify the estimate. The identities (6.10) for the sliced radii can also be checked directly by elementary trigonometry for a near-tangent pair, which would confirm or refute the geometric picture.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the base sphere-intersection measure lemma (Lemma 5.1) that the cone-projection bounds are built from."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The duality principle (Lemma 3.2) converts the orthonormal L^β estimate into a Hilbert–Schmidt kernel estimate, the core reduction step."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Schatten–Sobolev density framework and the interpolation-and-reduction strategy that this paper adapts to the wave equation."}],"review_version":1}