{"id":"e3a30a40-be41-4879-8090-a9a6baaf9bf8","arxiv_id":"2502.08812","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":5,"one_line_summary":"For energy-supercritical NLS on compact manifolds, the authors claim almost sure global well-posedness on a full measure set, with an invariant flow and polynomial-in-time Sobolev growth.","lead":"The paper constructs invariant probability measures and almost sure global solutions for energy-supercritical nonlinear Schrödinger equations on compact manifolds, including singular Sobolev data. It extends an earlier inviscid-infinite-dimensional limit method to general manifolds, all dimensions, and all sufficiently large nonlinearities.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.9's full-measure conclusion needs a small-ball estimate that the paper never proves; without μ(Σ)=1 the almost-sure statement of Theorem 1.1 does not follow.","rationale":"The reader's weakest assumption identifies precisely the missing small-mass estimate in Lemma 4.9, and my reading of the proof confirms this as the most load-bearing gap. The chain (4.26)-(4.29) in Lemma 4.9 uses (4.16) as if it gave a global bound on the complement of Σ^i_{N,s'}, but it only controls that complement intersected with E_a^N. The full-measure claim μ(Σ_{s'})=1 is what upgrades the positive-measure ensemble to an almost-sure statement in Theorem 1.1, and Proposition 4.11's globalization argument applies only on Σ_{s'}. Without it, the invariant measure and global flow are only defined on a set of positive but not necessarily full measure. The gap is likely repairable: the non-atomicity proved later (Prop 4.13(1)) should give μ(B_a)→0 as a→0, and a standard portmanteau/limiting argument could replace the faulty step. But the repair is not present in the manuscript, and it requires controlling the a-dependence of the constant in (4.15). I considered other concerns, such as the N-independence of Theorem 4.4's energy bound and the integer-power factorization in Lemma 4.7, but these appear more readily repairable and are less directly connected to the headline theorem's almost-sure content. Therefore the verdict remains CONDITIONAL: the central idea is plausible, but as submitted the proof of the full-measure ensemble is incomplete.","tokens_in":32588,"tokens_out":22787,"duration_ms":207075,"concrete_test":"Attempt the natural repair of Lemma 4.9: fix δ>0 and use Proposition 4.13(1) to choose a_m with μ(‖u‖_{L^2} ≤ a_m) < δ. By tightness and portmanteau, liminf_{N→∞} μ_N(‖u‖_{L^2} > a_m) ≥ 1-δ. Combining with (4.16) gives μ_N(Σ^i_{N,s'}) ≥ 1-δ - C(a_m) i^{-2k}, where C(a_m) comes from (4.15) and may blow up as a_m→0. Check whether the double limit lim_{m→∞} lim_{i→∞} (1-δ - C(a_m) i^{-2k}) = 1 can be made rigorous with the stated estimates. If C(a_m) does not grow too fast, the proof is repairable; if the a-dependence cannot be controlled, Lemma 4.9 fails as written and the full-measure assertion of Theorem 1.1 is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Theorem 1.1 is that the invariant measure μ assigns full measure to the global-flow set Σ. The proof of this fact is Lemma 4.9, which derives μ(Σ_{s'})=1 from Proposition 4.8's bound (4.16): μ_N(E_a^N \\ Σ^i_{N,s'}) ≤ C i^{-2k}, where E_a^N = {u ∈ E_N : ‖u‖_{L^2} > a}. That bound only controls the part of Σ^i_{N,s'} lying outside the ball B_a = {‖u‖_{L^2} ≤ a}. From (4.16) one gets μ_N(Σ^i_{N,s'}) ≥ μ_N(E_a^N) - C i^{-2k}, not μ_N(Σ^i_{N,s'}) ≥ 1 - C i^{-2k}. The missing term μ_N(E_a^N) = 1 - μ_N(B_a) requires a small-ball estimate: for every δ>0, one needs limsup_{N→∞} μ_N(B_a) ≤ δ for some a>0. Proposition 4.5 controls ∫ M(u)(1-χ_R(‖u‖_{L^2}^2)) dμ_N ≤ C R^{-1}, which is a large-mass tail estimate and says nothing about mass near zero. The paper later proves in Proposition 4.13(1) that ‖u‖_{L^2} under μ is non-atomic, hence μ(B_a)→0 as a→0, but this is not uniform in N and the proof does not perform the two-parameter limit (a→0, i→∞) needed to repair (4.28). As written, the chain (4.26)-(4.29) is therefore invalid. This is load-bearing: Proposition 4.11 (globalization) only applies to u0 ∈ Σ_{s'}, and if Σ_{s'} has measure strictly less than 1, the almost-sure global well-posedness asserted in Theorem 1.1 is not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims a probabilistic global well-posedness theorem (Theorem 1.1) for energy-supercritical nonlinear Schrödinger equations on compact Riemannian manifolds. The proof follows the IID-limit framework: a Galerkin system is stochastically damped and driven, stationary measures are constructed, the inviscid limit yields invariant measures μ_N for the Galerkin flow, and a statistical ensemble Σ is extracted on which uniform H^{s-} growth bounds permit a deterministic globalization argument. The main theorem asserts an invariant measure μ with μ(Σ)=1, flow invariance, almost-sure global existence for H^s data, polynomial-in-time growth, and absolute continuity of the L^2-mass distribution.","tokens_in":33093,"tokens_out":11184,"duration_ms":113223,"significance":"Should the proof be correct, this would be a substantial advance: it would extend almost-sure global well-posedness and invariant-measure constructions for energy-supercritical NLS from the torus and ball settings to general compact manifolds, cover all dimensions d≥3 and Sobolev orders up to H^{d/2}, and handle general nonlinearity powers. The explicit dissipation operator and the claimed simplification of the infinite-dimensional limit (Remark 1.3) are attractive features of the approach. However, the stated theorem is currently contingent on two load-bearing technical steps that are not justified as written: the full-measure property of the ensemble and the treatment of non-integer powers in the nonlinearity.","major_comments":[{"comment":"The conclusion μ(Σ_{s'})=1 is not established. Proposition 4.8 provides only μ_N(E_a^N \\setminus Σ_{i,N,s'}) ≤ C i^{-2k}, which controls failure outside B_a={‖u‖_{L^2}≤a}. In the chain (4.26)–(4.28), this is used as μ_N(Σ_{i,N,s'}) ≥ 1 - C i^{-2k}; that step requires μ_N(B_a) to be negligible uniformly in N. Proposition 4.5, Eq. (4.12), is a large-mass tail estimate and gives no control near the origin, while the later non-atomicity statement in Proposition 4.13(1) is not uniform in N, and no two-parameter limit (a→0, i→∞) is performed. Since Theorem 1.1(3) and the globalization Proposition 4.11 both depend on μ(Σ)=1, this gap is load-bearing.","section":"Lemma 4.9, Eq. (4.16)–(4.28)"},{"comment":"The factorization |u|^{2q}u - |v+z|^{2q}(v+z) = w f_{2q}(u,v) - g_{2q}(v,z)z, with f_{2q} and g_{2q} described as polynomials of degree 2q, is an algebraic identity valid only when 2q is an integer. The theorem allows q≥q_{M^d} with non-integer values (e.g., on T^3, Corollary 2.4 gives q≥5/3), and Proposition 3.1 explicitly treats real q>1. This identity is used to prove the inviscid-limit convergence in Proposition 4.6, so the convergence (III) in that proof is not justified for non-integer q without a substitute estimate. The manuscript should either impose integrality of 2q (and of 3k̃, as needed for F∈C^∞ in Proposition 4.2) or prove the required Lipschitz/factorization bounds for real powers.","section":"Lemma 4.7, Eq. (4.13)"},{"comment":"The local theory for non-integer q needs clarification. Proposition 2.1 uses the estimate ‖|P_N u|^{2q}P_N u‖_{H^s} ≲ ‖P_N u‖_{L^∞}^{2q}‖P_N u‖_{H^s} for s up to d/2. For real, non-integer q, the map u↦|u|^{2q}u is not C^{⌈s⌉}-smooth at the origin when s>1, so the standard composition/product estimates used in the contraction argument require additional hypotheses or a separate fractional-calculus argument. Since the theorem claims all q≥q_{M^d}, the non-integer cases must be either excluded explicitly or handled by suitable fractional composition estimates; as written, the local well-posedness input is not fully justified in the stated generality.","section":"Proposition 2.1 and Corollaries 2.3–2.5"}],"minor_comments":[{"comment":"The displayed inequality in (4.19) is tautological as written: the summand should involve μ_N(E_a^N ∩ φ_N^{-lT_0}(B_{i,j}^c)) (or an equivalent set), and the invariance/Chebyshev steps should be displayed explicitly.","section":"Section 4.4, Eq. (4.19)"},{"comment":"The proof cites “the inequality (4.30) below” before Lemma 4.10 is stated; the order should be changed or the citation adjusted.","section":"Lemma 4.9"},{"comment":"In the proof, several occurrences of W^{σ,q} should be W^{σ,p}; the Gronwall display and the surrounding estimates should be checked for consistency.","section":"Lemma 2.7"},{"comment":"Reference [58] is a duplicate of [57] with the same title, journal, volume, and pages; the bibliography should be cleaned up.","section":"References"},{"comment":"The notation s- is defined as s-ε in Section 1.7, but the theorem should state explicitly how ε in that notation is related to the growth exponent in (1.14).","section":"Theorem 1.1"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my read. The paper does something genuinely new: it pushes the IID-limit programme of Sy and Sy-Yu from tori and unit balls to general compact Riemannian manifolds, all dimensions, and singular Sobolev regularities s <= d/2 with all large nonlinearity powers. The local wellposedness argument based on Strichartz estimates is sensible, and the dissipation design using the Cordoba-Cordoba inequality is clever. If Theorem 1.1 is true, it is a significant advance in a genuinely hard supercritical regime. The soft spot the reader flagged is real and load-bearing. Lemma 4.9 tries to prove mu(Sigma_{s'})=1 from (4.16), which only controls mu_N(E_a^N \\ Sigma^i_{N,s'}). To conclude mu_N(Sigma^i_{N,s'}) >= 1 - C i^{-2k} you need mu_N(E_a^N) close to 1, i.e. a small-ball estimate for mu_N(B_a). The paper has a large-mass tail (Prop 4.5) but no small-mass control, and the later non-atomicity of ||u||_{L2} under mu (Prop 4.13) is not uniform in N and comes after the fact. As written, the chain (4.26)-(4.29) is invalid. Since Sigma_{s'} is the set on which the globalization lemma applies, the almost-sure statement of Theorem 1.1 does not follow. This looks repairable, because the dissipative structure should give enough control near zero, but it has to be supplied. There is a second, smaller gap: Lemma 4.7 factorizes |u|^{2q}u - |v+z|^{2q}(v+z) using polynomials of degree 2q. That is fine for integer q, but the theorem allows real q, so for non-integer powers one needs a different estimate. This is likely fixable, but as written it is an unproven step. The text also has rough spots that make verification harder: typos, garbled norm definitions, and some skipped justifications. None of that changes the core, but it adds friction. Bottom line: the paper is worth serious refereeing and the program is credible. I would send it out with a request for major revision, focusing on the small-ball estimate and the power issue. As it stands, I would not cite the theorem without the repair.","headline":"Serious IID-limit extension to compact manifolds with a real, repairable gap in the full-measure ensemble lemma.","tokens_in":33630,"tokens_out":4063,"would_cite":false,"duration_ms":36046,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35A01","35Q55","35R11","60H15","37K06","37L50"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every compact Riemannian manifold of dimension at least 3, the energy-supercritical nonlinear Schrödinger equation has almost-sure global solutions with invariant measures and slow Sobolev-norm growth, even for singular data of…","keywords":["nonlinear Schrödinger equation","energy supercritical","global well-posedness","invariant measure","compact Riemannian manifold","singular Sobolev spaces","inviscid-infinite-dimensional limit","statistical ensemble"],"falsifier":"For a concrete admissible case, such as quintic NLS on the three-dimensional torus, compute (analytically or numerically) the limit measure $\\mu$ obtained by the inviscid-infinite-dimensional procedure and check whether $\\mu(\\{u:\\|u\\|_{L^2}\\le a\\})\\to 0$ as $a\\to 0$; if this small-ball mass fails to vanish, or if an atom at any fixed $L^2$ value appears, then the full-measure ensemble of Theorem 1.1 cannot exist.","tokens_in":32390,"feed_emoji":"🌀","tokens_out":11813,"duration_ms":98327,"temperature":0.7,"pith_summary":"The paper aims to prove that the energy-supercritical nonlinear Schrödinger equation on a general compact Riemannian manifold of dimension at least 3 admits a probabilistic global theory: for almost every initial datum drawn from a specially constructed invariant measure, the solution exists for all time even though the datum may be too rough for classical energy-based well-posedness. This would bypass the standard obstruction that supercritical nonlinearities cannot be controlled by the linear evolution, replacing a conservation law with an invariant measure as the globalizing mechanism. The main theorem asserts the existence, for every admissible singular Sobolev order $s\\in(s_{M^d},d/2]$ and every sufficiently large power $q$, of a full-measure set $\\Sigma$ on which the NLS flow is global, maps $\\Sigma$ to itself, grows at most like $(1+|t|)^\\varepsilon$ in the $H^{s-}$ norm, and leaves the measure invariant, with the $L^2$-norm distribution absolutely continuous. The proof follows an inviscid-infinite-dimensional-limit strategy: build invariant measures for damped stochastic Galerkin approximations, pass to the inviscid limit, and select a statistically typical ensemble of initial data with uniform a priori bounds.","feed_headline":"Supercritical NLS tamed by invariant measures on compact manifolds","feed_subtitle":"For singular data below the H^{d/2} threshold the flow is global, invariant, and grows at most like (1+|t|)^ε.","key_machinery":"Three devices carry the argument. First, a local well-posedness theory for the Galerkin-truncated equation, based on Strichartz estimates that lose a fraction of a derivative on general compact manifolds but not on tori or Zoll manifolds; this sets the regularity thresholds $s_{M^d}$. Second, a dissipation operator $L_s(u)=(-\\Delta)^{s-1}u+C_{d,s}\\|u\\|_{H^{s-}}^{3\\tilde{k}-1}u$ inserted into a damped-driven stochastic Galerkin equation; a pointwise fractional-derivative inequality makes the dissipation rate of the energy coercive, producing stationary-measure bounds that are uniform in the damping and the Galerkin dimension. Third, an inviscid-infinite-dimensional-limit (IID-limit) procedure: stationary measures of the stochastic system converge, as the damping goes to zero, to invariant measures for the deterministic Galerkin flow; then a probabilistic representation principle plus Chebyshev-type tail estimates on the invariant measure selects a full-measure set of initial data whose Galerkin trajectories obey a uniform $(1+|t|)^\\varepsilon$ growth bound. A globalization lemma converts those uniform bounds into global existence in $H^s$ for the limiting flow.","core_discovery":"The central claim is that energy-supercritical NLS on compact manifolds is almost-surely globally well-posed for singular data. For every compact Riemannian manifold $(M^d,g)$ of dimension $d\\ge 3$, every Sobolev order $s\\in(s_{M^d},d/2]$, and every power nonlinearity $q\\ge q_{M^d}$, the paper constructs a set $\\Sigma=\\Sigma_{q,s,\\varepsilon}\\subset H^s$ and a probability measure $\\mu$ with the following properties: the solution map $\\varphi^t$ is a global flow on $\\Sigma$ with $\\varphi^t\\Sigma=\\Sigma$; every trajectory satisfies $\\|\\varphi^t u_0\\|_{H^{s-}}\\le C(\\|u_0\\|_{H^s})(1+|t|)^\\varepsilon$ for all $t$; $\\mu(\\Sigma)=1$; $\\mu$ is invariant under $\\varphi^t$; and the law of the functional $u\\mapsto\\|u\\|_{L^2}$ under $\\mu$ is absolutely continuous with respect to Lebesgue measure. On the torus and on Zoll manifolds the admissible regularity range extends to all $s>s_{q,d}=d/2-1/q$, the critical scaling exponent, while on general compact manifolds it is restricted to $s>s_{M^d}=d/2-1/(2q)$ by the available Strichartz estimates.","pith_inferences":["The same ensemble construction could plausibly transfer to other energy-supercritical dispersive equations on compact domains (for instance nonlinear wave or Hartree equations) whenever a coercive dissipation operator and Strichartz estimates are available; the paper does not state this.","The dependence on small-ball negligibility of the $L^2$-norm distribution suggests a direct stress test: compute the stationary measures' small-ball mass for the cubic NLS on the 3-torus; failure of $\\mu(\\{u:\\|u\\|_{L^2}\\le a\\})\\to0$ would break the full-measure property while leaving the trajectory bounds intact.","Because the growth bound holds for every $\\varepsilon>0$, one might conjecture sharper sub-polynomial (e.g. logarithmic) growth for typical data; the present argument is only constructed to give the $\\varepsilon$-polynomial bound.","The restriction $q\\ge q_{M^d}$ stems from relying on linear Strichartz estimates; as the paper's remark indicates, multilinear refinement should lower the nonlinearity threshold, bringing more physical nonlinearities into the theorem's admissible range."],"forward_implications":["For every compact Riemannian manifold of dimension $d\\ge 3$ and every admissible singular regularity up to $d/2$, energy-supercritical NLS admits a global flow defined almost surely with respect to an invariant probability measure, so the supercritical obstruction is circumvented in a measure-theoretic sense.","The bound $\\|\\varphi^t u_0\\|_{H^{s-}}\\le C(1+|t|)^\\varepsilon$ for every $\\varepsilon>0$ gives quantitative long-time control: typical Sobolev norms grow at most like a very small power of time on compact manifolds, where scattering is absent.","The invariance of $\\mu$ and the flow-invariance of the full-measure set $\\Sigma$ imply recurrence-type behavior: typical trajectories return infinitely often to every set of positive measure, providing a statistical substitute for scattering on bounded domains.","On tori and Zoll manifolds the result covers all regularities above the critical exponent $s_{q,d}=d/2-1/q$, while on general compact manifolds it covers $s>s_{M^d}=s_{q,d}+1/(2q)$, widening the previously known range for singular data from balls to arbitrary compact geometries.","The absolute continuity of the $L^2$-norm distribution rules out atoms, in particular at $u=0$, so the invariant measure is not trapped on a single orbit or near the zero state."],"supporting_citations":[{"why":"Supplies the Strichartz estimates with a fractional derivative loss that anchor the local well-posedness theory on general compact manifolds.","marker":"[15]"},{"why":"Provides the refined periodic Strichartz estimates that extend the admissible regularity range on tori down to the critical exponent.","marker":"[5]"},{"why":"Provides the Strichartz estimates for Zoll manifolds used to reach the critical exponent in that geometry.","marker":"[14]"},{"why":"Introduces the inviscid-infinite-dimensional-limit strategy and the absolute-continuity result (Theorem 9.1) that the present proof relies on for the full-measure ensemble.","marker":"[55]"},{"why":"Gives the local-time argument cited to prove that the L2-norm distribution under the limit measure is absolutely continuous and atom-free.","marker":"[49]"},{"why":"The pointwise fractional-derivative inequality that makes the dissipation rate of the energy coercive in the a priori estimates.","marker":"[23]"},{"why":"Extends the inviscid-infinite-dimensional-limit method to singular data on the unit ball, the singular-regularity starting point generalized here.","marker":"[56]"}],"fun_headline_variants":["Invariant measures yield global flow for supercritical NLS on compact manifolds","Almost sure global well-posedness for supercritical NLS on manifolds","Supercritical NLS has global flow almost surely on compact manifolds","Invariant measures construct global flows for supercritical NLS on manifolds","Energy-supercritical NLS: almost sure global well-posedness on manifolds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the measure assigns negligible total weight to initial data whose overall size ($L^2$ norm) is small; the proof imports this small-ball negligibility from a cited local-time result, and without it the constructed set $\\Sigma$ need not have full measure even if all trajectory bounds hold.","fun_headline_variants_meta":{"raw":{"variants":["Invariant measures yield global flow for supercritical NLS on compact manifolds","Almost sure global well-posedness for supercritical NLS on manifolds","Supercritical NLS has global flow almost surely on compact manifolds","Invariant measures construct global flows for supercritical NLS on manifolds","Energy-supercritical NLS: almost sure global well-posedness on manifolds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000968,"raw_usage":{"total_tokens":4115,"prompt_tokens":938,"completion_tokens":3177,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":554,"completion_tokens_details":{"reasoning_tokens":3077}},"tokens_in":554,"tokens_out":3177,"duration_ms":16666,"temperature":1.0,"reasoning_tokens":3077,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T23:37:55.813722+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a concrete admissible case, such as quintic NLS on the three-dimensional torus, compute (analytically or numerically) the limit measure $\\mu$ obtained by the inviscid-infinite-dimensional procedure and check whether $\\mu(\\{u:\\|u\\|_{L^2}\\le a\\})\\to 0$ as $a\\to 0$; if this small-ball mass fails to vanish, or if an atom at any fixed $L^2$ value appears, then the full-measure ensemble of Theorem 1.1 cannot exist.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Strichartz estimates with a fractional derivative loss that anchor the local well-posedness theory on general compact manifolds."},{"cited_title":"Bourgain","cited_arxiv_id":null,"evidence_quote":"Provides the refined periodic Strichartz estimates that extend the admissible regularity range on tori down to the critical exponent."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Strichartz estimates for Zoll manifolds used to reach the critical exponent in that geometry."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the inviscid-infinite-dimensional-limit strategy and the absolute-continuity result (Theorem 9.1) that the present proof relies on for the full-measure ensemble."},{"cited_title":"Shirikyan","cited_arxiv_id":null,"evidence_quote":"Gives the local-time argument cited to prove that the L2-norm distribution under the limit measure is absolutely continuous and atom-free."},{"cited_title":"C´ ordoba and D","cited_arxiv_id":null,"evidence_quote":"The pointwise fractional-derivative inequality that makes the dissipation rate of the energy coercive in the a priori estimates."},{"cited_title":"Almost sure global well-posedness for the energy supercritical NLS on the unit ball of $\\mathbb{R}^3$","cited_arxiv_id":"2007.00766","evidence_quote":"Extends the inviscid-infinite-dimensional-limit method to singular data on the unit ball, the singular-regularity starting point generalized here."}],"review_version":1}