{"id":"1a1f95d4-6178-4819-896a-ea4443aeb421","arxiv_id":"2501.13771","paper_version":2,"verdict":"REJECT","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"A Littlewood-Paley proof is given for a pointwise upper bound on the fundamental solution of the 1D fractional Fokker-Planck equation, with polynomial decay exponent 2+2s and an arbitrarily small epsilon.","lead":"This paper proves an upper bound on how fast the fundamental solution of the fractional Fokker-Planck equation decays in position and velocity. Decay bounds of this kind are building blocks for well-posedness and regularity proofs in kinetic PDE theory.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.2 is false: at s=1/2, m1=2, m2=0, η=0, ξ=8, the left side is ≈0.00458 while the right side is ≈0.00195.","rationale":"The reader's weakest_assumption correctly flags Lemma 3.2 as suspect, and the verdict REJECT is consistent with my reading. However, the reader's specific scaling objection (using ∂_ξ M for m1=1) does not by itself produce a falsehood, because the exponential factor e^{-M} can make the bound true there; the real counterexample occurs at m1=2, where the RHS decays too fast. Thus the same lemma is the problem, but the concrete mechanism differs. I also agree that the omitted case |x|≥|v| is a serious gap; the proof literally states 'Further details are omitted here' and then gives the desired bound. Nevertheless, the false Lemma 3.2 is the most load-bearing issue because it breaks the dyadic machinery used to prove all cases. Since a false lemma is a decisive correctness failure, the reader's REJECT verdict should stand. My recommendation is UNCHANGED: the preprint does not meet the standard of a supported theorem, even though the underlying result may well be true (as suggested by [9]).","tokens_in":15134,"tokens_out":18747,"duration_ms":144600,"concrete_test":"Evaluate Lemma 3.2 at s=1/2, m1=2, m2=0, η=0, ξ=8 with M(ξ,η)=∫_0^1 |(ξ-η)τ+η| dτ. Compute the left side |∂_ξ^2 e^{-M(8,0)}| = (1/4)e^{-4} ≈ 0.00458 and the right side 1/8^3 = 1/512 ≈ 0.00195. If the left side exceeds the right side, Lemma 3.2 is false. Then recompute the dyadic block estimate (3.4)-(3.6) for this case using the correct derivative bound; if the resulting ℓ1-sum over m1,m2 does not converge to the claimed power law, Theorem 1.1 is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central dyadic estimates in Lemma 3.3 all invoke Lemma 3.2, so the displayed bound (3.2) must be valid for every derivative order and every dyadic block. It is not. Take s=1/2, so M(ξ,0)=|ξ|/(2s+1)=|ξ|/2. For m1=2, m2=0, η=0, the left side of (3.2) is |∂_ξ^2 e^{-|ξ|/2}| = (1/4)e^{-|ξ|/2}. At ξ=8 this is (1/4)e^{-4} ≈ 0.00458. The right side of (3.2) is (|ξ|^{1+2s-m1}+|η|^{1+2s-m2})/(|ξ|^{m1+1}+|η|^{m2+1}) = (|ξ|^0+0)/(|ξ|^3+0) = 1/512 ≈ 0.00195. Since 0.00458 > 0.00195, the inequality fails. The source of the failure is visible already in Lemma 3.1: its final displayed bound drops the |ξ|^{2s-m1-m2} term that is present in the subcase |ξ|≥4|η|, giving an underestimate when η is small relative to ξ. Because Lemma 3.2 inherits this underestimate and Lemma 3.3 applies Lemma 3.2 in every dyadic block, the ℓ1-summation in Section 4 is not justified. The claimed pointwise bound of Theorem 1.1 is therefore unsupported by the proof as written. This is a concrete counterexample, not merely a missing detail. Separately, the proof of case 3 (|x|≥|v|) is only asserted without derivation, and case 2.2 (1/2<s<1) is summarized with 'we omit the details'; these gaps compound the failure, but the false Lemma 3.2 is the more fundamental obstruction.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the fundamental solution of the one-dimensional fractional Fokker-Planck equation ∂tf + v∂xf + |Dv|^{2s}f = 0. The fundamental solution is written as an explicit oscillatory integral with phase M(ξ,η)=∫_0^1 |(ξ−η)τ+η|^{2s}dτ. The authors prove derivative estimates for M and e^{-M} (Lemmas 3.1 and 3.2), then via Littlewood-Paley dyadic decomposition and several integrations by parts (Lemma 3.3) they bound each dyadic piece and sum the pieces (Section 4) to obtain the pointwise upper bound in Theorem 1.1. The result is claimed for all s∈(0,1) and all derivative orders.","tokens_in":15520,"tokens_out":5523,"duration_ms":41554,"significance":"If the pointwise bound is correct, it gives a sharp (up to ε) polynomial decay in the kinetic variables for the fundamental solution and its derivatives, with a transparent Fourier/Littlewood-Paley proof. The paper uses no fitting parameters, and the statement is consistent with the known two-sided heat-kernel estimates of [9]. However, the proof as written contains a false key lemma and omits essential cases, so the result is not established by this manuscript.","major_comments":[{"comment":"The estimate (3.2) is false. Take s=1/2, m1=2, m2=0, η=0, ξ=8. Then M(ξ,0)=|ξ|/2 and ∂_ξ^2 e^{-M} = (1/4)e^{-|ξ|/2}, so the left side is (1/4)e^{-4} ≈ 0.00458. The right side of (3.2) is |ξ|^{1+2s-m1}/(|ξ|^{m1+1}+1) = 1/8^3 = 1/512 ≈ 0.00195, so the inequality fails. Since Lemma 3.3 invokes (3.2) in every dyadic block, the dyadic summation in Section 4 is unjustified. The source of the failure is visible in Lemma 3.1, whose final '∼' display drops the |ξ|^{2s-m1-m2} term present in the subcase |ξ|≥4|η|.","section":"Lemma 3.2, Eq. (3.2)"},{"comment":"The entire case |x|≥|v| is dismissed with the sentence 'Similar to case 2' and the asserted bound ∑_{m1,m2} ~K_{m1,m2} ≲ |x|^{-(2+2s+b1-ε)} |v|^{-(b2+ε)} is not derived. Because the theorem's final bound requires this regime, the proof is incomplete.","section":"Section 4, case 3 (|x|≥|v|)"},{"comment":"The case 1/2<s<1 is only summarized: the text says 'Given its similarity to case 2.1, we just summarize key findings here' and 'we omit the details here'. The three displayed estimates for H_{m1,m2} are stated without derivation, and the summation to the claimed bounds is not shown. This is a separate case in the main theorem, so the omission is load-bearing.","section":"Section 4, case 2.2 (1/2<s<1)"}],"minor_comments":[{"comment":"The integers n1,...,n10 appear in the estimates before being defined; they should be quantified in the statement of the lemma, since the proof later chooses them.","section":"Lemma 3.3 statement"},{"comment":"The phrase 'there exists ε > 0 ... where ε arbitrarily small' is ambiguous; the theorem should state that for every sufficiently small ε>0 the bound holds with a constant depending on ε.","section":"Theorem 1.1 statement"},{"comment":"The notation ⟨x, x+v⟩ is not defined, and the equality |∂_x^{b1} ∂_v^{b2} K(1,x,v)| = |∂_x^{b1} ∂_v^{b2} K(1,x,−x−v)| in the final display is not justified from the Fourier integral representation.","section":"End of proof, Section 4"},{"comment":"There are several typos, e.g., 'stvarepsilons' in the proof of Lemma 3.3 and 'FUNDAMENT AL' in the title header of the arXiv source; a careful copyedit is needed.","section":"Throughout"}],"recommendation":"reject","confidential_remarks":"The paper's central proof is invalid because Lemma 3.2 is false. I recommend rejection. If the authors can repair Lemma 3.2, supply the omitted cases, and address the summation issues, the result may still be true, but the current manuscript is not publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this is a short note giving a Littlewood-Paley proof of a 1D pointwise upper bound for the fundamental solution of the fractional Fokker-Planck equation. The result itself is not new: the authors admit in Remark 1.3 that Hou and Zhang [9] already proved two-sided heat kernel bounds in arbitrary dimensions, and two-sided bounds imply their upper bound. The only real novelty would be the proof technique, not the theorem.\n\nThe paper does some things well. The Fourier representation of the kernel is derived cleanly, the scaling corollary is natural, and there are no fitted parameters or circular arguments. The citation pattern looks honest: [9] is acknowledged even though it undercuts the novelty.\n\nThe soft spots are serious. Lemma 3.2 is false as stated. Take s = 1/2, η = 0, m1 = 2, m2 = 0. Then M(ξ,0) = |ξ|/2, so the left side of (3.2) is (1/4)e^{-|ξ|/2}. The right side is |ξ|^0 / |ξ|^3 = 1/|ξ|^3. At ξ = 8, that is 0.00458 versus 0.00195, a concrete failure. The origin of the problem is visible in Lemma 3.1: the final displayed equivalence drops a |ξ|^{2s-m1-m2} term that is present in the |ξ| ≥ 4|η| subcase, and at η = 0 it underestimates by a power of |ξ|. Since Lemma 3.3 applies Lemma 3.2 in every dyadic block and Section 4 sums those blocks, the main estimate is not supported.\n\nThere is also an independent gap: case 3, |x| ≥ |v|, is disposed of with “similar to case 2” and no computation, and case 2.2 (1/2 < s < 1) is only summarized. Even if Lemma 3.2 were repaired, the proof is incomplete.\n\nWho is this for? Someone who specifically wants an analytic Littlewood-Paley proof rather than the stochastic-process proof in [9] might care, but they should wait for a corrected version. As written, the central lemma is false and a main case is omitted.\n\nRecommendation: desk reject. If the authors can fix Lemmas 3.1 and 3.2 and actually supply the missing case, the paper might become a modest technical note, but it does not meet the bar for refereeing now.","headline":"The main lemma is false as stated, a key case is not proved, and the theorem is already subsumed by Hou-Zhang's two-sided bounds; this is not ready for review.","tokens_in":16094,"tokens_out":3843,"would_cite":false,"duration_ms":35525,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35A08","35Q84"],"pacs":[],"model":"deepseek-v4-flash","headline":"The fractional Fokker-Planck kernel satisfies a pointwise polynomial decay bound for all derivatives.","keywords":["fractional Fokker-Planck equation","fractional Kolmogorov equation","fundamental solution","pointwise upper bound","Littlewood-Paley decomposition","fractional Laplacian","kinetic equation","derivative estimates"],"falsifier":"Evaluate Lemma 3.2 at $\\eta=0$ for $(m_1,m_2)=(1,0)$: compute $\\partial_\\xi e^{-|\\xi|^{2s}/(2s+1)}$ and compare both sides over $\\xi>0$; if no uniform constant works, the dyadic sums in Lemma 3.3 cannot be uniform. Separately, computing $|K(1,x,0)|$ for large $x$ at several values of $s$ would test the claimed $\\langle x\\rangle^{-(2+2s)}$ decay without relying on the omitted $|x|\\ge |v|$ calculation.","tokens_in":14884,"feed_emoji":"📉","tokens_out":18105,"duration_ms":146587,"temperature":0.7,"pith_summary":"The paper sets out to prove a pointwise upper bound for the fundamental solution --- the kernel that propagates a point source --- of the one-dimensional fractional kinetic Fokker-Planck equation $$\\partial_t f + v\\partial_x f + |D_v|^{2s} f = 0, \\qquad 0<s<1.$$ The claimed bound controls every derivative $\\partial_x^{b_1}\\partial_v^{b_2}K(1,x,v)$ by a product of powers of $\\langle x\\rangle$, $\\langle x+v\\rangle$, and $\\langle x,x+v\\rangle$, up to an arbitrarily small loss $\\varepsilon$. If the bound is correct, the kernel and all of its derivatives have polynomial tails simultaneously, and the scaling law for $K(t,x,v)$ turns the time-one statement into a $t$-dependent bound for every time. The authors offer this as the upper estimate needed in well-posedness and regularity arguments for fractional kinetic equations. The proof works on the Fourier side, using a frequency-block decomposition and derivative estimates for the phase $M$; the symmetric regime $|x|\\ge |v|$ and the range $1/2<s<1$ are summarized rather than written out in full.","feed_headline":"All derivatives of fractional Fokker-Planck kernel decay polynomially","feed_subtitle":"A frequency-block proof gives uniform decay in the natural kinetic variables x and x+v at every time.","key_machinery":"The load-bearing object is the phase $$M(\\xi,\\eta)=\\frac{1}{2s+1}\\frac{|\\xi|^{2s}\\xi-|\\eta|^{2s}\\eta}{\\xi-\\eta}.$$ Lemma 3.1 bounds its derivatives; Lemma 3.2 transfers those bounds to $e^{-M}$ through the higher-chain-rule formula for derivatives of composed functions; Lemma 3.3 localizes the Fourier integral with dyadic cutoffs $\\chi(|\\xi|/2^{m_1})\\chi(|\\eta|/2^{m_2})$ and integrates by parts in $\\eta$ --- twice for $0<s\\le 1/2$ and three times for $1/2<s<1$ --- to convert powers of $1/v$ into decay. The three dyadic regimes $|m_1-m_2|\\le 2$, $m_2\\ge m_1+3$, and $m_2\\le m_1-3$ are then summed separately. The Fourier representation evaluates the inverse transform at $(x,-x-v)$, which is why the final weight is expressed through $x$ and $x+v$.","core_discovery":"On its own terms, the central claim is Theorem 1.1: for $K$ the fundamental solution of (1.1), for every $b_1,b_2\\in\\mathbb{N}$ and every $0<s<1$ there is an arbitrarily small $\\varepsilon>0$ with $$|\\$partial_x^{{b_1}}$\\$partial_v^{{b_2}}$K(1,x,v)| \\lesssim \\frac{1}{\\langle x,x+v\\$rangle^{{2+2s-2\\varepsilon}}$\\langle x\\$rangle^{{\\varepsilon+b_1}}$\\langle x+v\\$rangle^{{\\varepsilon+b_2}}$},$$ where $\\langle u,w\\rangle=(1+|u|^2+|w|^2)^{1/2}$. The proof derives $K$ by space-velocity Fourier transform and characteristics, obtaining the phase $M(\\xi,\\eta)=\\int_0^1 |(\\xi-\\eta)\\tau+\\eta|^{2s}\\,d\\tau$, then proves derivative bounds for $M$ and for $e^{-M}$ and sums Littlewood-Paley blocks after integration by parts in $v$. The fully written computation covers $0<s\\le 1/2$ in the regime $|v|\\ge |x|$; the regime $|x|\\ge |v|$ is stated to be similar, and the case $1/2<s<1$ is summarized as analogous.","pith_inferences":["If Lemma 3.2 survives a direct homogeneity check, the dyadic summation is the only place it enters; a short calculation at $\\eta=0$ would settle whether the displayed exponents are harmless or need repair.","The omitted $|x|\\ge |v|$ case is likely reachable by the same scheme with integration by parts in $\\xi$ instead of $\\eta$, since the final representation shows the decay must be split between $x$ and $x+v$.","A two-sided version would need lower bounds for $e^{-M}$ on the same dyadic blocks, so the phase formula turns sharpness into a convexity question; the paper proves only the upper side.","The same frequency-block framework should adapt to higher-dimensional kinetic equations by replacing $M$ with the vector analogue of the same time integral and using anisotropic dyadic blocks."],"forward_implications":["Because the combined weight $\\langle x,x+v\\rangle^{-(2+2s)}$ is integrable over $\\mathbb{R}^2$, the bound puts $K$ and all its derivatives into natural weighted $L^1$ spaces.","The scaling identity $K(t,x,v)=t^{-1-1/s}K(1,x/t^{1+1/(2s)},v/t^{1/(2s)})$ upgrades the time-one estimate to an explicit $t$-dependent decay bound at every time.","Since $b_1$ and $b_2$ are arbitrary, the kernel is smooth with controlled tails in both $x$ and $v$, not merely integrable.","The bound has the expected heat-kernel shape for fractional Kolmogorov equations, which is the upper estimate the authors point to for future well-posedness and regularity arguments."],"supporting_citations":[{"why":"The explicit Gaussian fundamental solution for the $s=1$ kinetic equation, the classical case the fractional theorem is designed to extend.","marker":"[11]"},{"why":"The fractional heat-kernel asymptotic $K(1,x)\\lesssim (1+|x|)^{-(1+2s)}$, the one-dimensional model whose decay exponent is generalized.","marker":"[21]"},{"why":"A polynomial upper bound for kinetic integro-differential equations, the neighbouring result this paper complements with a Littlewood-Paley proof.","marker":"[15]"},{"why":"Two-sided heat-kernel estimates for nonlocal kinetic operators posted while this work was completed; the paper states its method and results are independent of them.","marker":"[9]"}],"fun_headline_variants":["Every derivative of fractional Fokker-Planck kernel decays polynomially","Fractional Fokker-Planck kernel: pointwise decay for all derivatives","Littlewood-Paley proof of polynomial decay for Fokker-Planck kernel","Uniform decay bounds for fractional Fokker-Planck fundamental solution"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole dyadic summation rests on Lemma 3.2's pointwise control of derivatives of $e^{-M}$, and that displayed bound is not homogeneous for pure $\\xi$-derivatives: at $\\eta=0$ and order $(1,0)$ the left side scales like $|\\xi|^{2s-1}$ while the stated right-hand side scales like $|\\xi|^{2s-2}$, so the block estimates in Lemma 3.3 inherit whatever defect that bound has.","fun_headline_variants_meta":{"raw":{"variants":["Every derivative of fractional Fokker-Planck kernel decays polynomially","Fractional Fokker-Planck kernel: pointwise decay for all derivatives","Littlewood-Paley proof of polynomial decay for Fokker-Planck kernel","Uniform decay bounds for fractional Fokker-Planck fundamental solution"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000297,"raw_usage":{"total_tokens":1671,"prompt_tokens":844,"completion_tokens":827,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":460,"completion_tokens_details":{"reasoning_tokens":744}},"tokens_in":460,"tokens_out":827,"duration_ms":6952,"temperature":1.0,"reasoning_tokens":744,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T15:36:55.788467+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate Lemma 3.2 at $\\eta=0$ for $(m_1,m_2)=(1,0)$: compute $\\partial_\\xi e^{-|\\xi|^{2s}/(2s+1)}$ and compare both sides over $\\xi>0$; if no uniform constant works, the dyadic sums in Lemma 3.3 cannot be uniform. Separately, computing $|K(1,x,0)|$ for large $x$ at several values of $s$ would test the claimed $\\langle x\\rangle^{-(2+2s)}$ decay without relying on the omitted $|x|\\ge |v|$ calculation.","supporting_citations":[{"cited_title":"P´ olya, On the zeros of an integral function represen ted by Fourier’s integral, Messenger of Math , 52 (1923), 185–188","cited_arxiv_id":null,"evidence_quote":"The fractional heat-kernel asymptotic $K(1,x)\\lesssim (1+|x|)^{-(1+2s)}$, the one-dimensional model whose decay exponent is generalized."}],"review_version":1}