{"id":"750d7a44-cb42-49c0-8037-339c10298494","arxiv_id":"2607.24400","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Vlasov–Poisson is locally well-posed in d≥2 for finite-mass data with weighted L^{p>d} velocity envelopes and arbitrarily small uniform velocity Hölder regularity.","lead":"The paper proves local existence and uniqueness for the Vlasov–Poisson system when the initial density is only slightly better than L^d in space and has an arbitrarily small Hölder modulus in velocity alone. That anisotropic low-regularity class sits below the usual full phase-space Sobolev theories and still yields a time-integrable Hölder electric field.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"No significant objection identified; the bootstrap the reader flagged is closed explicitly in the text.","rationale":"The reader's verdict and weakest-assumption analysis match my own read. The paper's architecture — twist lemma under a small L¹_t C^{1,α} field condition, dispersive representation (37), Hölder conversion via the G_θ modulus, Schauder fixed point with Aubin–Lions compactness plus tightness upgrade, and pointwise-in-time Loeper stability for uniqueness — is standard and each interface between lemmas checks out on careful reading. The one genuinely delicate point (smallness closure) is handled with the correct scaling: the exponent 1−q > 0 is exactly what makes T-smallness absorb the data size, and the restriction α < 1 − d/p is both necessary for time integrability of (11) and sufficient for the fixed point; the paper is transparent about this and about the endpoints. What remains is residual constant-tracking risk spread across Lemmas 3, 4, and 7 — several absorptions and weight comparisons whose errors would not be visible structurally but would show up numerically. Hence no verdict change: ACCEPT stands, MODERATE confidence is appropriate, and the proposed test would convert the soft reliance on tracked constants into direct evidence for the two central estimates (37) and (41).","tokens_in":18510,"tokens_out":6995,"duration_ms":365603,"concrete_test":"In d=2, prescribe a smooth divergence-generated C^{1,α} field E with ∫₀^T‖E‖_{C^{1,α}}ds just below the δ0 traced through Lemma 3, and take f0(y,w)=χ(y)⟨w⟩^{-m}(1+ε g(w)) with g a non-smooth θ-Hölder function (m>3). Compute ρ_E(t,x) two ways: (i) brute-force backward characteristic integration of (36), and (ii) the reparametrized formula (37) using the numerically inverted map v↦X_{0,t}(x,v). Compare the two densities pointwise and compare [ρ_E(t)]_{C^α} against the right side of (41). Agreement validates Lemmas 3–4 and the δ0 calibration; a discrepancy localizes exactly which absorption or weight-comparability step fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I read the same spot the reader flagged — Lemma 3's smallness hypothesis (19) and its closure for the self-consistent field — and checked whether it actually closes or merely appears to. It closes. In Lemma 7, ‖E_ϱ‖_{L¹_t C^{1,α}} is bounded by T^{1−1/r} times the Schauder bound C((M+R_p)T^{1/r}+R_α); substituting R_α = 2C A₀ T^{1/r−q} with q = d/p+α < 1 gives CA₀(T + T^{1−q}), which is below any fixed δ₀ for T small depending only on the stated parameters. δ₀ itself is reduced finitely many times in Lemma 3 (M ≤ 2, the N_t absorption, the H absorption), each time via T∫‖∇E‖ ≤ 1/2 with T ≤ 1, so δ₀ = δ₀(d) is honest. I also audited the secondary load-bearing steps: (i) the I1 estimate in Lemma 4 uses the |h| ≤ 1 cutoff in G_θ legitimately because (26) gives |ΔW_t| ≤ C|ℓ|/t ≤ 1 for |ℓ| ≤ ct, and (|ℓ|/t)^θ ≤ (|ℓ|/t)^α in that regime; (ii) the weight comparability ⟨W_t⟩ ≈ ⟨(x−y)/t⟩ follows from (24) at O(1) distance; (iii) the uniqueness Gronwall uses R(t) ~ t^{−d/p}, integrable since p > d, and the moment-free truncation in Lemma 8 is correct because the truncated couplings have equal mass, making the field difference dipole-decaying and weakly lower-semicontinuous in L²; (iv) tightness (51) needs only f₀ ∈ L¹. The endpoints p = d and θ = 0 are openly flagged as where the mechanism breaks. I find no internal inconsistency; correctness risk is the ordinary one of a long constant-tracking AP argument without machine verification.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript proves local well-posedness for the Vlasov–Poisson system on R^d_x × R^d_v, d≥2, in an anisotropic low-regularity class: the initial datum f0≥0 has finite mass, its weighted velocity envelope G0(x)=ess sup_v⟨v⟩^m|f0(x,v)| belongs to L^p_x for some p>d (with m>d+1), and a uniform weighted C^θ modulus in velocity, G_θ, also belongs to L^p_x. No spatial derivative of f0 is assumed. The central estimate is a nonlinear mixing bound [ρ(t)]_{C^α_x} ≲ t^{−d/p−α}(‖G0‖_p+‖Gθ‖_p) for any α<min{θ,1−d/p}, obtained from a characteristic reparametrization (the terminal-velocity-to-initial-position map is a global C^1 diffeomorphism under a small L^1_t C^{1,α} field hypothesis, Lemma 3) and an exact density representation (37). Time integrability of the singularity (since d/p+α<1) yields E∈L^1_t C^{1,α}, which closes a Schauder fixed point on the density via Aubin–Lions–Simon compactness and tightness. Uniqueness follows from Loeper's H^{−1} field estimate applied pointwise in time with an integrable coefficient R(t)~t^{−d/p}, plus a truncation argument (Lemma 8) removing moment assumptions. Continuous dependence on the datum in L^1∩L^p is also shown. The endpoints p=d and θ=0 are identified as critical for the method, and a nonuniqueness open problem is formulated.","tokens_in":18929,"tokens_out":7016,"duration_ms":244041,"significance":"If correct — and the argument appears sound — this is a solid contribution to the low-regularity theory of Vlasov–Poisson. It is, to my knowledge, the first well-posedness result in an anisotropic class with no spatial derivative of f0, going below the phase-space Sobolev theories of Jeong–Tae and Tae and adding uniqueness and C^{1,α} field regularity where the finite-energy Lagrangian theory of Ambrosio–Colombo–Figalli gives only existence. Particular strengths worth recording: the argument is complete and self-contained with all constants tracked; the mechanism (ballistic mixing converting fractional velocity regularity into spatial Hölder regularity) is identified transparently and shown to be sharp for the method, with the endpoints p=d and θ=0 openly flagged rather than hidden; the uniqueness proof is genuinely pointwise-in-time, using only an integrable L∞ singularity of the density; and §7 formulates precise, well-motivated open problems (the θ=0 endpoint, nonuniqueness below the density-stability threshold) that are likely to stimulate further work. The result is incremental in technique over the author's characteristic-reparametrization program [20, 21] but the data class,","major_comments":[{"comment":"Lemma 8 (p. 15) is the analytical input on which the uniqueness theorem rests, since (64)–(65) are applied with R=R(t)~t^{-d/p} unbounded as t↓0, so any hidden mass- or support-dependence of the constant would be load-bearing. The truncation/normalization step is currently compressed into two sentences ('applied after normalization by their common mass, gives... The normalization introduces no additional mass factor'). Please expand this: (i) state the scaling argument showing the constant in (64) depends only on d and R=max\\|ρ_i\\|_∞ and not on the common mass (the estimate for probability densities rescales cleanly, but this should be shown); (ii) justify ρ_{i,n}→ρ_i in L¹ (one line: ρ_{i,n}≤ρ_i with masses converging to M); (iii) justify the passage from distributional convergence of the truncated fields plus the uniform L² bound to (64) via weak lower semicontinuity, including why the","section":"§5, Lemma 8"}],"minor_comments":[{"comment":"Abstract vs. Theorem 1: the abstract states the mixing bound as [ρ(t)]_{C^α} ≲ t^{-d/p-ε} 'ε>0 small', whereas (11) gives t^{-d/p-α} with the explicit admissible range α<min{θ,1−d/p} (which need not be small when θ and 1−d/p are large). Please harmonize the notation so the abstract reflects the actual statement.","section":"Abstract / Eq. (11)"},{"comment":"The restriction m>d+1 is slightly stronger than what the proof needs: the current kernel K̃_t(z)=t^{-d}⟨z/t⟩^{-m+1} belongs to L^{p'} precisely when (m−1)p'>d, i.e. m>1+d−d/p, which is weaker than m>d+1 for finite p. Consider either noting the sharp condition or stating that m>d+1 is chosen for uniformity in p.","section":"Remark 2 / Lemma 4, Eq. (39)"},{"comment":"Reference [6] is listed as 'unpublished survey notes'. Please provide a stable citation (preprint number, journal status, or URL) if available, since §7 leans on it for the 2D Euler context.","section":"References"},{"comment":"The passage from an a.e.-defined section v↦f0(x,v) to a C^θ_loc representative is asserted; one sentence on measurability of the resulting selection (e.g., via the essential-supremum definition of G_θ) would make expressions like f0(y,W_t(x,y)) fully rigorous.","section":"§1.3, Eq. (3)"},{"comment":"In the uniqueness proof it would help the reader to note explicitly that R(t)=max_i‖ρ_i(t)‖_∞ is finite for every t>0 by (10) even though ρ_0 need only lie in L^p, so that Lemma 8 applies at each positive time and only the time-integrability of R(t) is needed for Gronwall.","section":"§5, proof of uniqueness"},{"comment":"The derivation ∥E_ϱ∥_{L^1_tC^{1,α}} ≤ CA_0T + CA_0T^{1−q} absorbs (M+R_p) into A_0; since R_p is chosen 'comparable to' the data norms rather than equal to them, a brief remark that the final T depends only on the stated parameters would close the bookkeeping.","section":"§4.2, Lemma 7"}],"recommendation":"minor_revision","confidential_remarks":"The characteristic reparametrization technique is closely adapted from the author's prior joint work [20, 21] and the Schauder framework from [28]; the self-citation pattern is transparent about this (§1.4) and the application here — a local theory in a genuinely new anisotropic data class — is a legitimate new result rather than a repackaging. Several key comparators ([24], [34], [8]) are very recent preprints; the author may wish to update their status at revision. The manuscript fits the scope of a specialist PDE/kinetic-theory journal well. No concerns about overlap or novelty disclosure beyond the above."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The new thing here is a local Lagrangian well-posedness theorem for Vlasov–Poisson that only needs a weighted uniform C^θ_v modulus of f0 controlled in L^p_x (p>d, θ>0 arbitrarily small), with no positive-order spatial derivatives of the datum. That sits below the phase-space Sobolev/Besov theories (Jeong–Tae, Tae, weighted Sobolev works) and still produces enough mixing that ρ lands in L^1_t C^α and E in L^1_t C^{1,α}, so Loeper uniqueness applies with an integrable L^∞ density singularity.\n\nWhat the paper does well is the physical-space route: short-time characteristic reparametrization (the twist lemma with Jacobian Hölder control), exact density/current formulas, Young-convolution dispersive bounds that convert velocity Hölder into spatial Hölder, then a clean Schauder set with Aubin–Lions–Simon compactness and a continuous density map. Uniqueness is the standard Loeper estimate plus Gronwall with the t^{-d/p} bound; the moment-free truncation argument is careful. The discussion is honest about the endpoints p=d and θ=0.\n\nThe soft spot the reader flagged—the smallness ∫∥E∥_{C^{1,α}}≤δ0 needed for the twist—is closed explicitly in Lemma 7 by the choice of R_α and T small. Secondary steps (weight comparability, the |h|≤1 cutoff in G_θ, tightness from L^1 only) check out on a second pass. Correctness risk is the ordinary one for a long constant-tracking AP existence proof without machine verification; nothing load-bearing looks broken. Self-citations to the author’s earlier characteristic work are technique, not circular.\n\nThis is for people who care about low-regularity kinetic Cauchy theory. It will not move global existence or applied modeling, but the anisotropic class is real and the write-up is usable. I would send it to referees; it deserves a serious read, not a desk reject.","headline":"Solid local well-posedness for VP in a genuinely anisotropic low-regularity class; the bootstrap closes and the argument is complete for what it claims.","tokens_in":19895,"tokens_out":515,"would_cite":true,"duration_ms":9011,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q49","35Q83","35Q85","35A01"],"pacs":[],"model":"grok-4.5","headline":"Vlasov–Poisson is locally well-posed when initial data has only L^{d+} spatial density control and an arbitrarily small positive Hölder modulus in velocity.","keywords":["Vlasov–Poisson equation","local well-posedness","fractional velocity regularity","kinetic mixing","Schauder fixed point","Lagrangian solutions","characteristic reparametrization"],"falsifier":"Exhibit two distinct Lagrangian solutions on a common short interval for a single nonnegative finite-mass datum whose weighted velocity envelope and weighted velocity Hölder envelope both belong to some L^p with p greater than dimension, or show that the density Hölder mixing bound fails for such data even under a field with arbitrarily small L^1_t C^{1,α} norm.","tokens_in":19625,"feed_emoji":"⚡","tokens_out":1092,"duration_ms":32710,"temperature":0.7,"pith_summary":"The paper proves that the Vlasov–Poisson system on phase space admits a unique short-time Lagrangian solution under anisotropic assumptions far weaker than classical phase-space Sobolev regularity. The initial distribution needs finite mass, a weighted velocity envelope in L^p in space for some p greater than the dimension, and only a tiny positive Hölder regularity in the velocity variable, controlled in the same L^p space. Free-streaming already turns that velocity modulus into spatial Hölder regularity of the macroscopic density; the paper shows the same ballistic mixing survives a self-consistent electric field that is small in a time-integrated sense. The resulting density is integrable in time with values in a positive Hölder class, so the field is regular enough to close a characteristic construction. Existence comes from a Schauder fixed point on the density; uniqueness follows from a Loeper-type stability estimate that tolerates an integrable blow-up of the density as time approaches zero.","feed_headline":"Tiny velocity smoothness unlocks Vlasov–Poisson well-posedness","feed_subtitle":"Local unique solutions exist with only L^{d+} density control and no full phase-space derivatives","key_machinery":"Nonlinear mixing via characteristic reparametrization: for a field with small L^1_t C^{1,α} norm the map from terminal velocity to initial position is a global diffeomorphism, yielding an exact free-transport-like formula for the density in which a velocity Hölder increment becomes a spatial Hölder increment of size |ℓ|/t, producing the time-integrable bound on [ρ(t)]_{C^α}.","core_discovery":"For dimension at least two, any p greater than d, and any positive velocity Hölder exponent however small, finite-mass nonnegative data whose weighted velocity supremum and weighted velocity Hölder envelope both lie in L^p in space generate a unique local Lagrangian solution of Vlasov–Poisson. The density obeys the nonlinear mixing bound that its spatial Hölder seminorm decays like a power of time strictly better than t^{-1}, hence is time-integrable in a positive Hölder class, and the electric field lies in L^1_t of C^{1,α}.","pith_inferences":["The argument marks p = d and zero velocity Hölder as genuine endpoints of the Hölder-flow method, suggesting any extension there must replace pointwise Jacobian control by log-Lipschitz or regular-Lagrangian-flow tools.","The open nonuniqueness question posed at the end is the natural kinetic counterpart of known flexibility results for 2D Euler below Yudovich: either a rigidity principle protects Vlasov–Poisson far below bounded density, or a singular nonnegative finite-mass counter-example exists.","Because the assumptions are anisotropic, physically natural data that are rough in space but mildly regular in velocity fall inside the theorem while remaining outside isotropic Sobolev local theories."],"forward_implications":["Macroscopic density need not be uniformly bounded up to time zero; an integrable singularity of order t^{-d/p} is enough for uniqueness.","No positive-order spatial derivative of the initial distribution is required—only a weighted uniform velocity Hölder modulus controlled in L^p.","The electric field belongs to L^1 in time with values in C^{1,α}, closing the characteristic flow.","On sets of data with uniformly bounded mass and envelopes, the solution map is continuous into continuous-in-time L^1 ∩ L^p phase-space densities.","The same mechanism recovers the classical free-transport conversion of velocity regularity into spatial regularity of the density as a special case."],"fun_headline_variants":["Tiny velocity Hölder unlocks local VP well-posedness","L^{d+} density plus scrap of velocity smoothness suffices","Fractional v-regularity yields unique local Vlasov–Poisson","Nonlinear mixing bound gives time-integrable Hölder density","Any positive velocity Hölder + L^{d+} data: local unique VP"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The self-consistent electric field must remain small enough, when integrated in time in the C^{1,α} norm, that particle trajectories stay a global diffeomorphism close to free streaming; that forces the existence time to be short.","fun_headline_variants_meta":{"raw":{"variants":["Tiny velocity Hölder unlocks local VP well-posedness","L^{d+} density plus scrap of velocity smoothness suffices","Fractional v-regularity yields unique local Vlasov–Poisson","Nonlinear mixing bound gives time-integrable Hölder density","Any positive velocity Hölder + L^{d+} data: local unique VP"]},"model":"grok-4.5","effort":"low","cost_usd":0.003845,"raw_usage":{"total_tokens":1250,"prompt_tokens":805,"num_sources_used":0,"completion_tokens":71,"cost_in_usd_ticks":38448000,"prompt_tokens_details":{"text_tokens":805,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":374,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":805,"tokens_out":71,"duration_ms":7770,"temperature":1.0,"reasoning_tokens":374,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T15:52:11.818282+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Exhibit two distinct Lagrangian solutions on a common short interval for a single nonnegative finite-mass datum whose weighted velocity envelope and weighted velocity Hölder envelope both belong to some L^p with p greater than dimension, or show that the density Hölder mixing bound fails for such data even under a field with arbitrarily small L^1_t C^{1,α} norm.","supporting_citations":[],"review_version":1}