{"id":"ce2f132e-9a0e-4940-acd5-66add431b56a","arxiv_id":"2607.04444","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Vlasov-Hartree system is globally well-posed for large low-regularity initial data; repulsive Coulomb interactions make densities decay to zero (with only logarithmic velocity-support growth), and attractive interactions cause only mild growth.","lead":"A rigorous proof that the Vlasov-Hartree system — a coupled model of quantum bosons and classical fermions — has unique global solutions for large low-regularity data. When the interaction is repulsive, the system provably disperses, with fermion velocities growing only logarithmically — behavior not yet proven for the simpler Vlasov-Poisson system alone.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 7(i)'s key log inequality is false for data-dependent constants; the Q(t)≲ln²t proof needs repair, though a direct fix exists.","rationale":"The paper’s central theorems are largely sound, and the reader’s overall CONDITIONAL verdict is appropriate. I focus on the most load-bearing gap in the decay proof: Proposition 7(i) contains a false elementary inequality that invalidates the proof of Q(t)≲ln²t. This is not a matter of disagreement with the physics or consensus; it is an internal inconsistency in a displayed estimate. The reader identified the same defect as a 'concrete defect' but treated another point (virial homogeneity) as the weakest assumption. I agree that the virial identity’s homogeneity is a limitation, but within the Coulomb setting it is an input assumption, not a gap. The Prop 7(i) inequality, by contrast, is a genuine proof error that must be repaired for the stated theorem to be fully established. The repair is straightforward and does not change the qualitative conclusions, so the paper should be accepted conditionally on that repair and on reconciling the other log-exponent bookkeeping errors the reader noted. My concrete test gives a direct verification of both the failure and the repair.","tokens_in":30233,"tokens_out":51768,"duration_ms":468188,"concrete_test":"Check the counterexample numerically: for C=2, t=1 compute ln(e+2(ln²(e+1)−1)) and compare with ln(e+1), confirming the failure. Then verify the repaired homogenization: set w(t)=y(t)−D(ln²(e+t)−1) with D>3C/2 (e.g., D=4C) in the differential inequality y′(t)≤C/(e+t)[ln(e+t)+ln(e+y(t))]; show that w′(t)≤C ln(e+w(t))/(e+t), integrate via F(w)=∫_0^w dr/ln(e+r), and recover w(t)≲ln t ln ln t, hence Q(t)≲ln²t. If both steps succeed, the stated Q bound and Corollary 1 are recovered.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The repulsive-case velocity-support bound Q(t)≲ln²⟨t⟩, which is essential for the full-range decay in Corollary 1, rests on Proposition 7(i). Its proof uses the inequality ln(e + C(ln²(e+t)−1)) ≤ ln(e+t). This is false for data-dependent constants C ≥ 2: at C=2, t=1, the left side is ln(e+2(ln²(e+1)−1)) ≈ 1.43 while the right side is ln(e+1) ≈ 1.31. The stated justification—that the two functions agree at t=0 and 'the former grows slower'—is incorrect; they cross at a C-dependent time. The gap is repairable by choosing a homogenizing subtraction with a sufficiently large coefficient: w(t)=y(t)−D(ln²(e+t)−1) with D>3C/2 gives w′(t)≤C ln(e+w(t))/(e+t), leading again to w(t)≲ln t ln ln t and Q(t)≲ln²t. Thus the theorem’s conclusion may survive, but the proof as written does not establish it. Because Corollary 1’s L^p decay for all p>5/3 relies on subpolynomial Q(t), this gap must be fixed before the central decay claim is fully supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Vlasov-Hartree system with Coulomb interaction V(x)=γ/(4π|x|), γ=±1, in the low-regularity setting where the fermionic density f may be discontinuous. It defines a notion of Lagrangian weak-mild solution, in which φ∈C([0,T];H^s) with s>3/2 so that the Vlasov characteristics are well-defined. The main results are: (i) global existence and uniqueness for large data, with f0∈L∞ with compact support and φ0∈H^s, s∈(3/2,2); (ii) in the repulsive case γ=1, decay estimates including P(t)≲t^{-1}, ∥ρ(t)∥_{L^{5/3}}≲t^{-3/5}, ∥φ(t)∥_{L^6}≲t^{-1/2}, Q(t)≲ln²t, and consequent L^p decay of ρ for all 1<p<∞ and L^q decay of φ for all q>2; (iii) in the attractive case, mild growth bounds Q(t)≲t ln t and ∥φ(t)∥_{H^s}≲t^{5s/3-1} ln^{5s/3-2}t. The proofs are based on energy conservation, a virial-type identity for the coupled system, endpoint Hardy-Littlewood-Sobolev estimates, and log-Lipschitz/Osgood uniqueness arguments.","tokens_in":30366,"tokens_out":7606,"duration_ms":79307,"significance":"If the results hold, this is a significant advance: it appears to be the first global well-posedness and large-data dispersive decay theory for the Vlasov-Hartree system, with no smallness assumption and only low regularity on the fermion density. The paper is largely self-contained and has several genuine strengths: the virial identity in Lemma 10 is verified by direct computation; the endpoint logarithmic estimate in Lemma 4 is concrete and used sharply; the proof has no fitted parameters or data-tuned constants; and the uniqueness argument via Osgood's lemma is appropriate for the low-regularity setting. The main caveat is that one load-bearing step in the proof of the repulsive velocity-support bound, Proposition 7(i), contains a false elementary inequality. This does not appear to be a fatal flaw—the literature-style homogenization repair is straightforward—but it must be fixed before the full decay theorem is established.","major_comments":[{"comment":"The proof asserts ln(e+C(ln²(e+t)−1)) ≤ ln(e+t) and justifies this by saying the two functions agree at t=0 and the former grows slower. This inequality is false for data-dependent constants C that can be large. For example, with C=2 and t=1, the left side is ln(e+2(ln²(e+1)−1)) ≈ 1.43, while the right side is ln(e+1) ≈ 1.31. The constant C in (15) arises from estimates and is not controlled, so the homogenized inequality for w(t) and the resulting bound Q(t)≲ln²⟨t⟩ do not follow as written. This is load-bearing: Corollary 1's full-range L^p decay for p>5/3 uses Q(t)≲ln²t through the interpolation argument at the end of §4.4. The gap appears repairable within the paper's framework: taking w(t)=y(t)−D(ln²(e+t)−1) with D sufficiently large compared to C yields w′(t)≤C ln(e+w(t))/(e+t) and hence w(t)≲ln t ln ln t, which preserves the stated Q(t) bound. I recommend that the author replace th","section":"§4.3, Proposition 7(i)"}],"minor_comments":[{"comment":"Typo: 'exhibts' should be 'exhibits'.","section":"§1.2"},{"comment":"The refined estimate for ∥E∥_{L∞} has a logarithmic denominator ∥φ∥_{L6}². If the L6 norm is zero, the expression is undefined; a trivial continuity/limiting argument would remove this degeneracy.","section":"§2.1, Lemma 4"},{"comment":"In the convergence proof for the error term, the identity Cε = −χ_ε ∗ ∇V is written with a vector-valued mollifier χ_ε(z)=zχ_ε(z); the notation is a little compressed and could be clarified by explicitly indicating the dot product in the convolution.","section":"§4.1, Lemma 10"},{"comment":"The unconventional definition ⟨x⟩=√(2+|x|²) is used repeatedly; this is fine, but it may be worth a short note that all logarithmic factors are evaluated with this convention.","section":"Notation"}],"recommendation":"major_revision","confidential_remarks":"The main technical framework is sound and the central claims are plausible, but the proof of Proposition 7(i) contains a genuine false inequality that undermines the repulsive velocity-support bound and, through it, part of Corollary 1. The issue is localized and a standard repair exists, so I recommend major revision rather than rejection. The author should also audit other places where logarithmic inequalities are justified by 'one function grows slower' with data-dependent constants."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. The paper is the first large-data global theory for the Vlasov-Hartree system, and the Lagrangian weak-mild solution notion is a genuine contribution. The repulsive-case decay — rho in L^{5/3} like t^{-3/5}, phi in L^6 like t^{-1/2}, and especially Q(t) less than or similar to ln^2 t — is stronger than anything known for Vlasov-Poisson without symmetry assumptions. The virial identity (Lemma 10) checks out by direct computation, and the consequences are internally consistent. The citation pattern looks honest: the authors rely on standard tools, cite the recent derivation of the model, and explicitly acknowledge in Section 1.5 that the screened Coulomb potential breaks the virial identity.\n\nThe soft spots are real but local. Proposition 7(i) uses the inequality ln(e + C(ln^2(e+t)-1)) <= ln(e+t), which fails for data-dependent constants C >= 2. The stated Q(t) ~ ln^2 t bound does not follow from the proof as written. However, the gap is repairable: a direct Gronwall comparison with a larger subtraction coefficient still gives w(t) ~ ln t ln ln t and hence Q(t) ~ ln^2 t. The theorem's conclusion likely survives, but the author needs to fix that step. There are also log-exponent inconsistencies in Corollary 1 and its proof; the exponent on the log in the statement is three times what the preceding display gives. Since logs are subpolynomial, this is not load-bearing, but it should be cleaned up.\n\nThe s > 3/2 threshold for uniqueness is standard, and the homogeneity of the Coulomb potential is doing essential work in the virial argument — the authors acknowledge this, so I don't hold it against them. The long estimates in Sections 3.3-3.4 and 4.2-4.3 look coherent to the level I could spot-check, and the circularity burden is zero.\n\nVerdict: this is a solid and important paper that deserves a serious referee. I'd recommend sending it to peer review with a request to fix the log inequality and reconcile the exponents. If the author does that, I'd be comfortable with the main claims.","headline":"Strong first large-data theory for Vlasov-Hartree, with a real but repairable log-inequality gap in the velocity-support proof.","tokens_in":31035,"tokens_out":6479,"would_cite":true,"duration_ms":57315,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q83","35Q55","35B40","35D30","35A01","35A02"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that the Vlasov–Hartree system, a mean-field model of a Bose–Fermi mixture, has a unique global Lagrangian weak-mild solution for all large data, and that in the repulsive Coulomb case every density and field decays to zero","keywords":["Vlasov-Hartree system","Bose-Fermi mixture","global well-posedness","large initial data","dispersive decay","virial identity","low-regularity solutions","Coulomb potential"],"falsifier":"A numerical simulation of the repulsive Vlasov–Hartree system with large, discontinuous f0 and exact Coulomb potential should show ∥ρ(t)∥_{L^{5/3}} decreasing like t^{-3/5} and ∥φ(t)∥_{L^6} like t^{-1/2}; a single large-data run that exhibits slower decay, no decay, or a nontrivial asymptotic state would contradict Theorem 2(i). More directly, recompute Lemma 10ε with V replaced by the Yukawa potential e^{-r}/r: Euler's relation gives x·∇V+V=-r e^{-r}, so the error term Err(t) no longer vanishes, marking exactly where the proof mechanism stops.","tokens_in":29927,"feed_emoji":"⚛️","tokens_out":9796,"duration_ms":89152,"temperature":0.7,"pith_summary":"This paper studies the Vlasov–Hartree system, a mean-field model in which fermions obey a classical Vlasov equation and bosons obey a quantum Hartree equation, coupled through the Coulomb potential. It establishes that for any bounded, compactly supported initial fermion density and any boson wavefunction in H^s with s>3/2—so the fermion density may be discontinuous—there is a unique global solution that conserves mass and energy and has well-defined particle trajectories. In the repulsive case the solution always disperses: the fermion density decays like t^{-3/5} in L^{5/3}, the boson field decays like t^{-1/2} in L^6, the potential energy decays like t^{-1}, and the velocity support grows at most like (log t)^2. Interpolation then gives convergence of the fermion density to zero in every L^p for 1<p<∞ and of the boson field to zero in every L^q for 2<q≤∞. If correct, a repulsive Bose–Fermi mixture with finite mass, energy, and spatial moments has no nontrivial steady states: all energy is eventually converted to kinetic form.","feed_headline":"Decay is guaranteed: repulsive Bose-Fermi mixtures relax to zero","feed_subtitle":"Virial identity forces fermion density, boson field, and potential energy to decay; velocities spread at most (log t)^2.","key_machinery":"The load-bearing object is the combined virial–pseudo-conformal identity (Lemma 10): (1/2)∫|(x+it∇)φ|²dx + (1/2)∫∫|x−vt|²f dxdv + t²P(t) = C + ∫_1^t sP(s)ds, where P(t)=∫(V*ρ)|φ|²dx is the potential energy. Because V=γ/(4π|x|) is homogeneous of degree −1, Euler's relation x·∇V+V=0 cancels the error term, and in the repulsive case P≥0, so Gronwall forces P(t)≲t^{-1} and at most linear growth of the two quadratic quantities; a pseudoconformal change then gives the L^6 decay of φ and the L^{5/3} decay of ρ. Around this identity the paper assembles a global-existence machinery: a mollified regularized system, averaging lemmas for compactness of ρ, strong convergence of the characteristic flow, c","core_discovery":"The central claim is that the Vlasov–Hartree system with Coulomb interaction V=γ/(4π|x|) is globally well-posed for large data in a low-regularity class—f0 bounded and compactly supported, φ0 in H^s for 3/2<s<2—with a unique solution that conserves mass and energy and has Lipschitz trajectories even when the fermion density is discontinuous. For repulsive γ=1, the solution disperses: ∥ρ(t)∥_{L^{5/3}}≲t^{-3/5}, ∥φ(t)∥_{L^6}≲t^{-1/2}, potential energy decays like t^{-1}, velocity support grows at most (log t)^2, and interpolation gives ρ(t)→0 in every L^p (1<p<∞) and φ(t)→0 in every L^q (2<q≤∞). For attractive γ=−1, global existence holds with mild growth, Q(t)≲t log t and algebraic H^s growth","pith_inferences":["Going beyond the paper, the virial mechanism suggests a scaling law for homogeneous potentials |x|^{-α}: the same tracking would give ∥φ(t)∥_{L^6}≲t^{-α/2}, so steeper repulsive potentials would disperse faster, though weaker smoothing would complicate low-regularity well-posedness.","The attractive-case mild-growth bounds leave open whether nontrivial steady states are stable; the low-regularity solution class is exactly the one in which fermionic ground states are indicator functions, so one could probe orbital stability of such mixtures with this framework.","Because the proof's decay rests on exact homogeneity, screened or Yukawa interactions may require a different, possibly faster, decay mechanism; a numerical comparison of P(t) between Coulomb and Yukawa would separate the virial effect from generic dispersion."],"forward_implications":["In the repulsive case, every solution in the class converges to the zero equilibrium; no nontrivial steady states with finite mass, energy, and spatial moments can exist, and all energy converts to kinetic form.","The decay rates are quantitative: potential energy ~ t^{-1}, fermion density in L^{5/3} ~ t^{-3/5}, boson field in L^6 ~ t^{-1/2}, and electric field ~ t^{-1} log t.","Fermion velocities spread at most like (log t)^2 in the repulsive case, so even discontinuous initial densities become asymptotically dispersed in every L^p sense.","For attractive Coulomb interactions, no finite-time blow-up occurs within this class; top-order norms grow at worst algebraically and logarithmically, so the conservation laws nearly control the dynamics.","The low-regularity global theory applies to indicator-like initial data, the physically expected zero-temperature ground states of the fermion density."],"fun_headline_variants":["Repulsive Vlasov-Hartree: decay to zero proven","Virial force drives repulsive Vlasov-Hartree to zero","Global well-posedness and decay for Vlasov-Hartree","Attractive vs repulsive: decay or mild growth in Vlasov-Hartree","Dispersion for repulsive Vlasov-Hartree, even with rough densities"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The decay results rest on the exact algebraic homogeneity of the Coulomb kernel: Euler's relation x·∇V+V=0 is what makes the error term in the virial identity vanish; with a screened or non-inverse-distance potential the identity gains a nonzero error term and the paper's decay framework no longer applies.","fun_headline_variants_meta":{"raw":{"variants":["Repulsive Vlasov-Hartree: decay to zero proven","Virial force drives repulsive Vlasov-Hartree to zero","Global well-posedness and decay for Vlasov-Hartree","Attractive vs repulsive: decay or mild growth in Vlasov-Hartree","Dispersion for repulsive Vlasov-Hartree, even with rough densities"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000317,"raw_usage":{"total_tokens":1637,"prompt_tokens":762,"completion_tokens":875,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":506,"completion_tokens_details":{"reasoning_tokens":774}},"tokens_in":506,"tokens_out":875,"duration_ms":8201,"temperature":1.0,"reasoning_tokens":774,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T08:42:09.574316+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A numerical simulation of the repulsive Vlasov–Hartree system with large, discontinuous f0 and exact Coulomb potential should show ∥ρ(t)∥_{L^{5/3}} decreasing like t^{-3/5} and ∥φ(t)∥_{L^6} like t^{-1/2}; a single large-data run that exhibits slower decay, no decay, or a nontrivial asymptotic state would contradict Theorem 2(i). More directly, recompute Lemma 10ε with V replaced by the Yukawa potential e^{-r}/r: Euler's relation gives x·∇V+V=-r e^{-r}, so the error term Err(t) no longer vanishes, marking exactly where the proof mechanism stops.","supporting_citations":[],"review_version":3}