{"id":"84fc8bba-eeb1-4aad-9119-5124774c47b0","arxiv_id":"2607.22329","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Weak solutions exist for a regularised fluctuating homogeneous Landau equation with moderately soft potentials, small conservative noise in Landau-divergence form, and a refined entropy decay when the noise modes are divergence-free.","lead":"This paper adds thermal noise to the Landau equation, the kinetic equation describing how collisions relax the velocity distribution of a gas or plasma, and proves that the resulting stochastic equation has at least one solution when the noise is weak and a smoothing function replaces the singular square-root of the density. It is a rigorous first step for fluctuating kinetic theory at the Landau level.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix D's d=2 divergence-free basis is not tangential, so Theorem 1.2's noise-basis construction is invalid as written; Theorem 1.1 is not affected.","rationale":"I read the paper as a substantial, carefully executed existence proof for the regularised Itô Landau equation. The four-step approximation scheme, the uniform L² and L¹ compactness arguments, and the exponential entropy estimate are presented in enough detail that I do not find a clear flaw in Theorem 1.1 itself. The exclusion of γ=-2 is explicitly justified, and the small-noise threshold, while delicate, appears fixable by the standard device of shrinking ε₀ to account for the finitely many constants arising in Hölder and exponential-martingale estimates. The reader's conditional verdict is therefore appropriate for the main theorem. However, the paper also advertises Theorem 1.2 as a central contribution: the refined entropy inequality for divergence-free noise bases. There, the d=2 construction in Appendix D is objectively incorrect—the proposed vector harmonics are radial, not tangential. This does not change the verdict on Theorem 1.1, but it strengthens the conditionality of the overall paper: either the construction must be corrected (e.g., using the constant rotation field e_θ in d=2) or Theorem 1.2 must be restricted to d≥3. The reader's weakest_assumption focused on Assumption 2.1 and the Fisher-information constant; I agree those are the key structural assumptions for Theorem 1.1, but I disagree that they are the only load-bearing assumptions, since the divergence-free basis construction is the unexamined support for the paper's second headline result. Hence 'partial' agreement, with the verdict unchanged at CONDITIONAL.","tokens_in":88635,"tokens_out":20960,"duration_ms":182023,"concrete_test":"In dimension d=2, take l=1 and Y_{1,1}=cos θ. Compute T_{1,1}=R∇_{S¹}Y/1 with R=[[0,1],[-1,0]]. Since ∇_{S¹}cosθ=-sinθ e_θ and R e_θ=e_r, one obtains T_{1,1}=-sinθ e_r, for which ω·T_{1,1}=-sinθ ≠ 0. This directly contradicts the required tangentiality ω·T=0 for a mode in L^2_tan,div(S¹;TS¹). Replacing the construction by the constant tangent field e_θ (normalized) would restore a valid, odd, divergence-free basis, but that correction is absent from the manuscript.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's second main result (Theorem 1.2, the refined entropy inequality advertised in the abstract) rests on Assumption 7.3, whose active modes must be tangential and divergence-free on the sphere. Appendix D claims to construct an ONB of L^2_anti,tan via T_{l,m}(ω)=R∇_{S¹}Y_{l,m}/l in dimension d=2, with R=[[0,1],[-1,0]]. But on S¹, ∇_{S¹}Y is proportional to e_θ, and R e_θ = e_r, the radial direction. Hence the proposed T_{l,m} are radial, not tangential: ω·T_{l,m} ≠ 0 in general. Consequently G_k = √A Π g_k vanishes identically for these modes, so the construction does not yield nontrivial active modes satisfying (1.20). The d=3 toroidal construction T=ω×∇Y/√(l(l+1)) is correct, but the theorem is stated for all d≥2. A valid odd, divergence-free, tangential basis does exist in d=2 (the constant rotation field e_θ), so the flaw is repairable, but the appendix as written does not supply it. The main existence theorem (Theorem 1.1) does not use the divergence-free condition and is not invalidated by this issue.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a stochastic perturbation of the homogeneous Landau equation in which the conservative noise is written in nonlocal Landau-divergence form, is antisymmetric in the colliding velocities, and is interpreted in the Stratonovich sense before passing to Itô form. The square-root mobility is replaced by a regular coefficient σ that behaves like √r away from vacuum and linearly near zero, and the noise is projected onto a finite antisymmetric basis. For moderately soft potentials γ∈(−2,0), Theorem 1.1 asserts the existence of probabilistic weak solutions to the regularised Itô equation for sufficiently small noise intensity ε, under regularity and support conditions on the noise basis; the solutions conserve mass and momentum, satisfy an energy inequality, and satisfy the entropy-dissipation inequality (1.18) with an additive constant arising from the Stratonovich–Itô correction. The proof combines Galerkin approximation, artificial diffusion, L²-compactness, an exponential entropy-dissipation estimate, and L¹-compactness via a Desvillettes-type weighted Fisher-information bound. Theorem 1.2 claims a refined entropy inequality for a θ-regularised equation when the noise basis is tangential and divergence-free on the relative-velocity sphere, giving monotonicity of the expected θ-entropy without the additive constant.","tokens_in":88884,"tokens_out":14545,"duration_ms":114700,"significance":"The existence theory for a fluctuating Landau equation with nonlocal, conservative noise is new and the paper is unusually careful about the stochastic singularities: Remark 5.4 honestly explains why γ=−2 cannot be handled in the L² framework; the additive constant in (1.18) is explicitly identified as the cost of the Stratonovich–Itô correction; and the exponential entropy-dissipation estimate (Proposition 5.2) is a genuinely stochastic tool. The formal particle derivation (Appendix A) and the Stratonovich-to-Itô conversion (Appendix B) are worked out in detail and give useful heuristics. The main weakness is that the noise-basis construction behind the second main result is incorrect in d=2 as written, and the small-noise threshold ε₀ is non-explicit and depends on entropy-exponential constants; the latter is acceptable for an existence theorem, while the former needs repair.","major_comments":[{"comment":"The d=2 construction of the divergence-free basis is invalid. In (D.3), T_{l,m}=R∇_{S¹}Y_{l,m}/l with R the 90° rotation. On S¹, ∇_{S¹}Y is tangent, so R∇_{S¹}Y is radial; indeed the explicit formula T_{l,1}(ω)=−(1/√π)sin(lθ)ω confirms ω·T_{l,m}≠0. Such radial modes give G_k=√A Π g_k=0, so they do not generate active noise and Assumption 7.3 is not satisfied. Theorem 1.2 is therefore not established for d=2 as stated. Since a valid tangential divergence-free antisymmetric basis exists in d=2 (e.g., the constant rotation field e_θ), the defect is repairable, but the appendix must be rewritten.","section":"Appendix D / Theorem 1.2"},{"comment":"The passage α→0 for the θ-regularised equation needs a uniform L¹_t W^{1,1}_v bound on f^α. The proof after (7.16) asserts E∫₀ᵀ ∫_{B_R} |∇θ(f^α)|² ≤ C(f₀) 'directly from [Des15, Theorem 1] applied to θ²', but the required control of the entropy H(θ²(f^α)) and of the dissipation D(θ²(f^α)) is not spelled out. This is a load-bearing step for Theorem 1.2; please expand it with the same detail as Lemma 6.4.","section":"Section 7, Lemma 7.5"}],"minor_comments":[{"comment":"The cross-reference 'Assumption 2.12' appears twice; the initial-data condition is displayed as (2.12), not as a numbered assumption. Please correct.","section":"Section 1.3"},{"comment":"The support condition '∪ Supp G_k ⊂ {r∈R+ | K^{-1} ≤ r ≤ K} × R^d × S^{d-1}' is notationally ambiguous. Please specify the coordinate decomposition r=|v−v_*|, z=(v+v_*)/2, ω=(v−v_*)/|v−v_*|.","section":"Assumption 2.1, (2.6)"},{"comment":"The sentence 'these modes are normalised in L²(S¹;R²)' and the subsequent claim that T_{l,m} is tangential contradict the explicit formula T_{l,1}(ω)=−(1/√π)sin(lθ)ω, which is radial. The text should be revised consistently.","section":"Appendix D"},{"comment":"The proof uses H both for the positive part of entropy and as a placeholder in inequalities; please define H⁺ and H⁻ explicitly and use them consistently to avoid confusion.","section":"Lemma 6.3"},{"comment":"The phrase 'countable dense set of full-measure times' is used to justify the supremum lower semicontinuity; this is standard but could be stated more explicitly for the exponential estimates.","section":"Section 6.2, Step 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is mathematically substantial and the main existence theorem (Theorem 1.1) appears sound. The d=2 defect in Appendix D is surprising given the overall care; it must be fixed before publication. The second major comment asks for a fuller justification of the uniform gradient bound in Lemma 7.5. No concerns about novelty or attribution: the bibliography is appropriate and the limitations are honestly stated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the headline: this is the first solution theory for a fluctuating homogeneous Landau equation with conservative velocity-pair noise, and the main existence theorem looks right. The advertised divergence-free basis construction for the refined entropy theorem is, however, wrong in d=2 as written, and the paper needs a repair before that second result can be trusted.\n\nWhat is new and good: the noise is written in Landau-divergence form, antisymmetric in the pair of velocities, and is motivated by the fluctuation–dissipation principle and the covariance of a Kac-like particle system. The paper proves existence of probabilistic weak solutions for the regularised Itô equation with γ∈(−2,0), conserving mass and momentum, with energy and entropy-dissipation inequalities (the latter up to an explicit additive constant). The Stratonovich-to-Itô correction is handled by replacing the square-root mobility with a regular coefficient, and Remark 5.4 honestly explains why γ=−2 fails in the SPDE setting while deterministic theory covers it. The exponential entropy estimate (Prop. 5.2) is a genuinely stochastic ingredient, and the four-step compactness scheme is carefully laid out. The claims are scoped honestly: covariance matching with the particle system is formal, and fluctuation coincidence is deferred.\n\nSoft spots: Appendix D's d=2 construction is not tangential. T_{l,m}=R∇_{S^1}Y_{l,m}/l gives radial vectors because R e_θ = e_r, so ω·T≠0 and the projected fields G_k vanish identically. Thus Assumption 7.3 is not satisfied non-trivially by that construction in d=2. A valid basis does exist (the constant rotation field e_θ, used with antisymmetric z- and r-factors), so the flaw is repairable, but Theorem 1.2 as written lacks a working d=2 basis. This does not touch Theorem 1.1. A second soft spot is that ε_0 inherits exponential-in-entropy constants from the Desvillettes weighted Fisher-information lemma, so the small-noise threshold could be extremely small, though still positive for fixed f0. Also, the existence theory is for the regularised equation: smooth σ, finite K modes, and the physical limits (σ→√·, K→∞, recovering particle-system fluctuations) are open. The paper says so; it is a scope limitation, not an overclaim. Finally, the proofs are very long and many estimates are delegated to earlier propositions; I saw no circularity, but independent verification of the L2 and L1 passages is warranted.\n\nWho is this for: researchers in stochastic kinetic theory and singular SPDEs, and anyone working on Dean–Kawasaki-type equations with nonlocal mobility. It deserves a serious referee. Recommendation: send it to peer review, but require the authors to fix or restrict Appendix D (e.g., state Theorem 1.2 for d≥3 or give the correct d=2 basis) and to state the dependence of ε_0 on the entropy constants more explicitly.","headline":"Main existence theorem for regularised fluctuating Landau equation is solid; the d=2 divergence-free basis construction for the refined entropy theorem is wrong as written, though fixable.","tokens_in":89509,"tokens_out":3994,"would_cite":true,"duration_ms":38705,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60H15","35Q20","35Q84"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that a fluctuating Landau equation with regularised thermal noise has probabilistic weak solutions for moderately soft potentials, conserving mass and momentum and dissipating entropy up to an explicit noise-driven constan","keywords":["fluctuating Landau equation","probabilistic weak solution","conservative thermal noise","fluctuation–dissipation principle","moderately soft potentials","Stratonovich-to-Itô correction","entropy dissipation","nonlocal kinetic SPDE"],"falsifier":"A concrete check: simulate the Galerkin scheme of Section 3 with a divergence-free noise basis supported at scale K and measure the entropy gap E[H(f_t)] + E[∫_0^t D(f_s) ds] − H(f_0) as ε is lowered. The theorem predicts this gap stays bounded by C(‖σ′‖_{L∞}) for all ε below a positive threshold; observing it grow without bound as ε → 0, or finding the threshold collapse to zero as the mode-support scale K grows, would refute the claim.","tokens_in":88378,"feed_emoji":"🎲","tokens_out":16265,"duration_ms":120504,"temperature":0.7,"pith_summary":"The paper introduces thermal noise into the homogeneous Landau equation of kinetic gas theory, producing a stochastic PDE whose noise mimics the random fluctuations of a large particle system around its deterministic mean-field limit. Its central claim is an existence theorem: for moderately soft collisions, provided the noise is regularised (a smooth mobility replaces the singular square-root amplitude, only finitely many antisymmetric noise modes act, and the noise intensity is small), at least one probabilistic weak solution exists. The solution keeps the defining structure of the deterministic Landau equation: mass and momentum are conserved almost surely, energy does not increase, and the Boltzmann entropy is dissipated up to an explicit constant depending on the noise coefficient. For a specially chosen class of noise modes satisfying a tangential divergence-free condition, the expected entropy is non-increasing, recovering deterministic dissipativity exactly. This supplies a weak-solution theory for a fluctuating kinetic equation with genuinely nonlocal, pair-coupled noise — groundwork for checking whether such equations correctly reproduce particle-system fluctuations.","feed_headline":"Small thermal noise keeps the Landau equation solvable","feed_subtitle":"New existence proof: mass, momentum and entropy dissipation survive a regularised noise.","key_machinery":"Load-bearing object: the Landau difference gradient ∇̃φ(v,v_*) = Π(v−v_*)⊥(∇_vφ − ∇_v_*φ_*), which turns the Landau operator into a gradient flow of Boltzmann entropy; the thermal noise is placed on the same collision space in divergence form so its covariance matches the particle-system martingale. The singular √(A f f_*) mobility is replaced by a smooth σ(f)σ(f_*) (σ = √r away from vacuum, linear at 0), and the Stratonovich noise is converted to Itô form, producing nonlocal correction terms that couple v, v_*, w. The proof runs on a three-level approximation (Galerkin truncation, coefficient regularisation, artificial diffusion) with L²-then-L¹ compactness, plus an exponential entropy esti","core_discovery":"On the paper's own terms: the regularised Itô fluctuating Landau equation, with conservative noise whose covariance matches the fluctuation martingale of the underlying particle system, has at least one probabilistic weak solution for every small noise intensity ε < ε₀, for kernels |v−v_*|^{γ+2}, γ ∈ (−2,0), d ≥ 2, and nonnegative initial data of finite energy and finite Boltzmann entropy. The solution conserves mass and momentum almost surely, obeys the energy inequality, and satisfies the entropy-dissipation inequality up to an additive constant C(‖σ′‖_{L∞}); for a tangential divergence-free noise basis, the expected entropy is non-increasing. The proof uses Galerkin truncation, coefficien","pith_inferences":["Beyond the paper: the natural next test is the vanishing-noise limit ε → 0 — the theorem does not quantify how fast solutions return to the deterministic Landau flow, so tracing the ε-dependence of all constants would be needed to see the transition.","Beyond the paper: since the regularised mobility σ is linear rather than square-root near vacuum, the genuinely singular noise amplitude is not covered here; extending a renormalised-solution method — which the paper explains fails for the nonlocal Landau operator — is the implied route to removing the regularisation.","Beyond the paper: the divergence-free condition admits a geometric reading — only noise modes acting purely tangentially, without compressing the collision geometry, preserve entropic monotonicity; numerical comparison of entropy production for general versus admissible bases would show how much deterministic dissipativity survives generic thermal noise.","Beyond the paper: the exclusion of the borderline γ = −2, located by the paper in the failure of the uniform L² estimate, suggests an L¹-based compactness argument might extend the existence theorem to that threshold."],"forward_implications":["If correct, the theorem gives a well-defined weak-solution theory for the regularised fluctuating Landau equation with the same conservation laws as the deterministic equation: mass and momentum are conserved almost surely, and the entropy-dissipation inequality holds up to an explicit noise-driven constant.","The paper's own next step comes into reach: the solution theory is the stated foundation for proving that the Gaussian fluctuations and large deviations of the stochastic equation coincide with those of the underlying conservative particle system.","For noise modes satisfying the tangential divergence-free condition, the additive constant disappears: the expected θ-entropy is non-increasing, matching the deterministic Landau dissipation structure exactly.","The exponential entropy-dissipation estimate, derived from the supermartingale structure of the entropy balance, supplies the stochastic substitute for the deterministic entropy bound and is the tool expected to transfer to other singular fluctuating-hydrodynamics SPDEs."],"fun_headline_variants":["Regularised noise preserves Landau solvability","Noisy Landau equation still has weak solutions","Thermal noise regularised, Landau solvable","Proof: noisy Landau equation has weak solutions","Soft-potential Landau solutions exist under regularised noise"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that physically natural noise bases keep the regularity and support constants of the active modes, and the exponential-in-entropy constant entering the small-noise threshold, finite — so that the existence threshold ε₀ stays positive and the Itô-correction terms stay integrable for small noise.","fun_headline_variants_meta":{"raw":{"variants":["Regularised noise preserves Landau solvability","Noisy Landau equation still has weak solutions","Thermal noise regularised, Landau solvable","Proof: noisy Landau equation has weak solutions","Soft-potential Landau solutions exist under regularised noise"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0013,"raw_usage":{"total_tokens":5143,"prompt_tokens":751,"completion_tokens":4392,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":495,"completion_tokens_details":{"reasoning_tokens":4318}},"tokens_in":495,"tokens_out":4392,"duration_ms":25417,"temperature":1.0,"reasoning_tokens":4318,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T05:05:58.530022+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: simulate the Galerkin scheme of Section 3 with a divergence-free noise basis supported at scale K and measure the entropy gap E[H(f_t)] + E[∫_0^t D(f_s) ds] − H(f_0) as ε is lowered. The theorem predicts this gap stays bounded by C(‖σ′‖_{L∞}) for all ε below a positive threshold; observing it grow without bound as ε → 0, or finding the threshold collapse to zero as the mode-support scale K grows, would refute the claim.","supporting_citations":[],"review_version":1}