{"id":"a0d5780a-a115-4d83-8319-dff773766983","arxiv_id":"2411.16424","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A complete orthonormal eigenbasis for the radially symmetric Lévy Fokker-Planck operator is constructed, with explicit fractional Hermite-type eigenfunctions and integer eigenvalues.","lead":"This paper constructs a complete set of exact eigenfunctions for the fractional Fokker-Planck equation, giving a basis for radially symmetric solutions of the fractional heat equation. A generalist should read it because it upgrades a leading-order convergence result into a full asymptotic expansion, a step that had been missing for fractional diffusion and Lévy processes.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Uniqueness proof in Prop. 5.22 never eliminates the sin(2πw) coefficient, leaving a nonconstant 2s-periodic multiplier; Theorem 1.1's simplicity is not established.","rationale":"The reader's weakest_assumption identifies Proposition 5.22 as load-bearing, and my review agrees. The proof of that proposition is the only place where the paper excludes nonconstant 2s-periodic multipliers for the Mellin solution. I traced the final Liouville step: after the boundedness/cancellation arguments, the coefficient b is forced to zero but the coefficient a of sin(2πw) remains. The resulting multiplier v(z)=a sin²(2πw) is nonconstant, entire, and compatible with the growth estimates used in the proof, so the printed argument does not establish uniqueness. Because Theorem 1.1 explicitly claims uniqueness of each eigenfunction up to scalar, this is a genuine gap in the central claim. I do not assert the theorem is false: the Fourier-conjugation route (Lemma 4.1 plus the classical spectrum of L1 on L2(exp)) may provide a repair, but that route is not the one presented and requires an explicit surjectivity/regularity argument for the transformation u↦u†. The paper's own caveats (unproved bound in Section 4.4, n=1 pathology) further support a conditional verdict. The concrete check I propose is aimed at settling whether the missing a-term can actually occur in L2†, which would resolve whether the gap is cosmetic or fatal.","tokens_in":46476,"tokens_out":24250,"duration_ms":215130,"concrete_test":"For s=1/2, n=2, ν=0, construct the candidate u_a by setting the periodic multiplier in (5.12) to v(z)=sin²(2πz) (i.e. φ(w)=(1/(2i))(e^{2πiw}−e^{-2πiw})), and compute u_a by Mellin inversion. Then evaluate the L2† norm (4.12) by closing the integration contour and summing residues, or numerically via the series (4.14). If the norm is finite for a≠0, then eigenvalue 0 has an eigenfunction not proportional to e_0^{(1/2)}, disproving the uniqueness part of Theorem 1.1. If the norm diverges, repeat for a generic irrational s (e.g. s=1/√2) to check whether the missing estimate can be repaired; an analytical check that the printed bound (5.15) imposes no condition on a at w=0 confirms the gap is not closed by the current argument.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Proposition 5.22, the proof reduces the 2s-periodic multiplier v(z) to v(z)=sin(2πw)φ(w) with w=(n−z)/(2s). A generalized Liouville argument gives φ(w)=a sin(2πw)+b cos(2πw)+c. The paper then shows b=c=0, but a is never forced to vanish. Hence v(z)=a sin²(π(n−z)/s), a genuinely nonconstant 2s-periodic function, satisfies all the regularity and growth bounds used in the proof (it is entire and grows like e^{2π|Im w|}, within the allowed e^{5π|Im w|/2}). Substituting this v into (5.12) yields a second family of solutions to the Mellin eigenvalue equation (5.2) that are not scalar multiples of (5.3). If any such u belongs to L2† for some k∈N, then Theorem 1.1's claim that eigenfunctions are uniquely given by (1.10) fails, and the basis may include extra elements. The proof's opening step (u∈L2† implies u∈H^k† for all k) is also stated as a consequence of Lemma 4.1, which only gives the conjugation Ls u=f ⇔ L1 u†=f†; for the H^k claim one needs elliptic regularity for the local equation, which is not shown.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the spectral problem for the radial Lévy Fokker–Planck operator L_s = (−Δ)^s − (1/(2s)) x·∇ − n/(2s) on a new weighted space L2†(s). The main theorem (Theorem 1.1) claims that L2† has a complete orthonormal basis of eigenfunctions e_k with eigenvalues k, given explicitly in Euclidean variables by (1.10) and in Fourier variables by (1.11). The proof uses an intertwining transform Φ_s that relates L_s to the local operator L_1, together with Mellin transform techniques. The paper also treats the adjoint fractional Ornstein–Uhlenbeck operator and derives a corollary giving all-order asymptotics for radial solutions of the fractional heat equation.","tokens_in":46729,"tokens_out":10366,"duration_ms":87047,"significance":"The explicit, parameter-free formula for the fractional analogue of Hermite/Laguerre eigenfunctions and the identification of the natural weighted space are potentially valuable contributions. The intertwining relation (1.8), the Mellin symbol computation in Proposition 3.7, and the explicit eigenfunction formulae in Sections 5.2–5.4 are elegant and essentially self-contained. If the completeness and uniqueness claims survive scrutiny, the paper would provide a clean spectral description with immediate applications to asymptotic expansions for the fractional heat equation. However, the uniqueness proof in Proposition 5.22 leaves open a concrete possibility of additional eigenfunctions, so the central claim is not yet established.","major_comments":[{"comment":"The proof never eliminates the coefficient a in the final expression φ(w)=a sin(2πw)+b cos(2πw)+c. After showing b=c=0, the multiplier v(z)=sin(2πw)φ(w) reduces to a sin²(π(n−z)/s), which is a nonconstant 2s-periodic function satisfying the growth bound (5.15). Substituting this v into (5.12) yields additional solutions of the Mellin eigenvalue equation (5.2) that are not scalar multiples of (5.3). If any such solution belongs to L2† for some k∈N, then Theorem 1.1's claim that eigenfunctions are uniquely given by (1.10) fails. The paper must either force a=0 or prove that these additional solutions lie outside L2†; neither is currently done.","section":"§5.5, Prop. 5.22, eqs. (5.12)–(5.15)"},{"comment":"The assertion 'By Lemma 4.1, u∈L2† implies that u∈Hk† for all k' is not a consequence of Lemma 4.1, which only states the equivalence L_s u=f ⇔ L_1 u†=f†. To justify (5.13) for every k one needs a weighted elliptic regularity statement for the local Ornstein–Uhlenbeck operator L_1 on L²(exp); no such regularity result is proved or cited in the paper. This step is load-bearing because the bound (5.15) on φ is derived from the asserted H^k† membership.","section":"§5.5, first sentence of the proof"},{"comment":"The claimed equivalence (4.15) with the explicit δ_m from (4.17) is not proved. Lemma 4.9 derives the formula for δ_m, but the text immediately concedes that its upper bound is 'a problem in the field of number theory which we do not attempt to consider here.' Since (4.15) is used to describe L2† as a fractional Gevrey–Sobolev space, either the upper bound must be supplied or that description should be labeled as conditional or formal.","section":"§4.3–4.4, eq. (4.15) and Lemma 4.9"}],"minor_comments":[{"comment":"The condition n ≥ 2s appears in the introduction, while Theorem 1.1 assumes n ≥ 2; the role of the former condition and its relation to the latter should be clarified.","section":"§1, first paragraph and Remark 5.2"},{"comment":"The phrase 'without loss of generality' is not accurate for the full spectrum, since non-radial eigenfunctions are not characterized; the paper only treats the radial part of the spectrum, and this should be stated more carefully.","section":"§1, restriction to radial functions"},{"comment":"The notation U_k is used both for the Mellin-space function defined in (5.5) and for its inverse Mellin transform U_k(r); this overlap should be resolved to avoid confusion.","section":"§1, eq. (1.12), and §5.3"},{"comment":"The sentence 'The results for L2‡ are parallel to those for L2‡' should presumably read 'those for L2†'.","section":"§6, first paragraph"},{"comment":"There are typographical errors: 'Insituto' in the affiliation, 'Fokker–Plank' in the introduction, 'eingenfunctions' in Section 6, and 'greaterorsimilar' in Lemma 5.12 where an inequality symbol is intended.","section":"throughout"}],"recommendation":"major_revision","confidential_remarks":"The main concern is exactly the one identified in the stress-test note: the uniqueness proof in Proposition 5.22 does not exclude the coefficient a, leaving open a nonconstant 2s-periodic multiplier sin²(π(n−z)/s) that would add extra eigenfunctions. This is a genuine, load-bearing gap rather than a presentation issue. The paper is otherwise interesting and the gap may be repairable by showing that the additional solutions fail to belong to L2†, or by an alternative argument. The unproved upper bound for δ_m in Lemma 4.9 is also openly acknowledged by the authors; if §4.3 is only heuristic, that should be stated explicitly. The paper fits the journal scope, but in its current form the central theorem is not fully supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe paper is worth a serious look. It constructs a genuinely new Hilbert space L2† for the fractional Fokker–Planck operator, gives explicit eigenfunctions in Euclidean and Fourier variables, and proves completeness of the basis for radial functions (modulo one gap I'll flag). The core intertwining observation—change Fourier variable ϑ=ζ^s to reduce the fractional operator to the classical one—is clean, and the Mellin machinery is well executed. The Fourier-side formulas ζ^{2sk}e^{-ζ^{2s}} were already in the physics literature [44,45,27,28], and the authors say so; the real contribution is the functional-analytic frame and the Mellin proof, which is not circular.\n\nWhere it's soft: the uniqueness proof in Prop. 5.22 does not actually finish. The generalized Liouville step leaves a free coefficient a in φ(w)=a sin(2πw)+b cos(2πw)+c. The proof forces b=0 (using poles of Γ(1−w) at w=1 for n≥2), but a is never eliminated. That leaves a nonconstant 2s-periodic multiplier v(z)=a sin²(π(n−z)/s) that satisfies the bounds used, so the claimed uniqueness of the eigenfunctions is not established. Whether those extra solutions land in L2† is not shown. This is load-bearing for Theorem 1.1's \"uniquely\" statement; it may be fixable by a finer L2† argument, but right now the basis claim rests on an unproven step. The opening step \"u∈L2† implies u∈H^k†\" also needs a local regularity argument that is only invoked, not proved.\n\nMinor: the upper bound for δ_m in Lemma 4.9/eq. (4.15) is explicitly unproved (number-theoretic difficulty), so the Gevrey–Sobolev interpretation is heuristic at that point. The \"without loss of generality\" radial reduction is not carried out; the paper proves the theorem for radial functions, which is fine, but the WLOG language overstates.\n\nNet: the framework and explicit formulas are likely correct and important; the uniqueness gap is specific and could be repaired. I'd send it to a good referee but expect the referee to demand the fix. I'd cite it for the L2† space and the Mellin multipliers, with caution on the uniqueness claim.","headline":"A valuable but not-yet-finished spectral framework for the fractional Fokker–Planck operator; the uniqueness proof in Prop. 5.22 has a real gap that a referee should push the authors to close.","tokens_in":47292,"tokens_out":3450,"would_cite":true,"duration_ms":33386,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R11","35P10","33C45","35K05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that the Lévy Fokker–Planck operator has a complete explicit eigenbasis with eigenvalues 0,1,2,…, giving all-order asymptotics for fractional heat solutions.","keywords":["Lévy Fokker–Planck equation","fractional heat equation","Mellin transform","fractional Hermite polynomials","Laguerre polynomials","fractional Ornstein–Uhlenbeck","spectral basis","eigenfunction expansion"],"falsifier":"Solve the Mellin functional equation (5.2) for $n=1$, $\\nu=1/2$ and check whether the resulting function lies in $L^2_\\dagger$; a nonzero admissible solution would contradict Proposition 5.22 and break the claimed completeness of the basis.","tokens_in":46255,"feed_emoji":"🧮","tokens_out":7648,"duration_ms":69144,"temperature":0.7,"pith_summary":"This paper aims to settle the fine asymptotic description of the fractional heat equation by giving the full spectrum of the Lévy Fokker–Planck operator that appears after passing to self-similar variables. It constructs a weighted Hilbert space in which the operator is self-adjoint, proves that its eigenvalues are exactly the nonnegative integers, and writes the eigenfunctions explicitly as fractional analogues of Hermite polynomials—Laguerre polynomials composed with $r^{2s}$ and transported by a Mellin multiplier (equivalently $\\zeta^{2sk}e^{-\\zeta^{2s}}$ in Fourier variables). The dual fractional Ornstein–Uhlenbeck problem is treated by the same method. If correct, every radially symmetric solution of the fractional heat equation in this space has a convergent eigenfunction expansion, yielding asymptotic formulas to all orders rather than only convergence to the fundamental solution.","feed_headline":"Fractional heat equation gets an explicit eigenfunction basis","feed_subtitle":"Mellin variables make the fractional operator classical, giving expansions to every order.","key_machinery":"The machinery is the Mellin transform on radial functions together with the Mellin multiplier $\\Lambda_s(z)$ defining a quasi-isometry $\\Phi_s$; $\\Phi_s$ encodes the change of radial Fourier variable $\\vartheta=|\\xi|^s$ that converts the nonlocal Lévy Fokker–Planck operator into the classical local one ($L_s u=f \\iff L_1 u^\\dagger=f^\\dagger$). Under Mellin transform, $(-\\Delta)^s$ becomes multiplication by a symbol $\\Theta_s(z)$ of order $2s$, and the eigenvalue equation reduces to a first-order difference equation whose solutions are quotients of Gamma functions. The integer eigenvalues are selected by the requirement of exponential decay in $r^{2s}$, which occurs exactly when the Kummer function reduces to a Laguerre polynomial; the resulting Mellin–Barnes integrals give explicit Fox–Wright and Legendre-function formulas. This machinery replaces the ODE uniqueness argument of the local case with a Mellin-domain uniqueness proof.","core_discovery":"The central discovery is that for every $s\\in(0,1)$ and $n\\ge 2$, the operator $L_s u=(-\\Delta)^s u-\\frac{1}{2s}x\\cdot\\nabla u-\\frac{n}{2s}u$ is self-adjoint on a weighted Hilbert space $L^2_\\dagger(s)$ and has pure point spectrum $\\{0,1,2,\\dots\\}$: there is a complete orthonormal basis $\\{e_k\\}$ with $L_s e_k=k e_k$. In Fourier variables the eigenfunctions are $\\hat e_k(\\zeta)=e^{-\\zeta^{2s}}\\zeta^{2sk}$, $k\\in\\mathbb{N}$, and in Euclidean variables they are given by a Mellin multiplier applied to Laguerre polynomials, $e_k=\\Phi_s\\{(\\frac{r^{2s}}{4})^{-n(1-s)/(2s)}e^{-r^{2s}/4}L_k^{(n-2)/2}(\\frac{r^{2s}}{4})\\}$. The same construction gives a dual basis for the fractional Ornstein–Uhlenbeck operator $L^*_s$, and separation of variables then yields a unique solution of the fractional heat equation with a full asymptotic expansion to all orders.","pith_inferences":["Editorial inference: the change of variable $\\vartheta=|\\xi|^s$ plus Mellin multipliers may be a general template—any radially symmetric diffusion whose Fourier symbol is a pure power of $|\\xi|$ and whose drift is linear should have the same integer spectrum after the same reparametrization.","Editorial inference: the paper's own Remark 5.23 hints that in one dimension, non-integer eigenvalues may correspond to extra solutions outside $L^2_\\dagger$ but inside a larger distribution space; these could describe long-time tails of 1D Lévy processes that the Hilbert-space expansion misses.","Editorial inference: the Gevrey–Sobolev interpretation of $L^2_\\dagger$ suggests the basis is best adapted to initial data with super-polynomial Fourier decay; testing the expansion on data with only algebraic decay should reveal slower-than-exponential convergence of the coefficients.","Editorial inference: adding a Hardy-type potential $c u/r^{2s}$ breaks the explicit formulas but preserves the Mellin structure, so one could test numerically whether the spectrum shifts continuously away from integers as $c$ varies."],"forward_implications":["Every radially symmetric initial datum in $L^2_\\dagger$ gives a unique solution $\\phi(t,r)=\\sum_{k=0}^\\infty a_k e^{-kt}e_k(r)$ to the fractional heat equation.","As $t\\to\\infty$, the solution behaves like the self-similar fundamental solution to leading order, with all higher-order corrections indexed by $k$ and decaying like $t^{-k}$ in self-similar variables.","The dual basis $\\{\\omega_k\\}$ gives an equally explicit eigenfunction expansion for the fractional Ornstein–Uhlenbeck equation, with the same integer eigenvalues.","For $s\\to 1$ the construction recovers the classical Hermite and Laguerre spectral theory of the heat and Ornstein–Uhlenbeck equations.","Non-integer solutions of the eigenvalue equation fail to lie in $L^2_\\dagger$, so the spectrum of the operator in this space is exactly $\\mathbb{N}$."],"supporting_citations":[{"why":"Supplies the compactness and spectral theorem for the local weighted Fokker–Planck problem, which is transported to the fractional setting.","marker":"[29]"},{"why":"Provides the local self-similar heat-equation framework whose eigenbasis is the starting point for the fractional analogue.","marker":"[14]"},{"why":"Gives the Mellin transform, inversion theory, and Mellin–Barnes integral tools used throughout.","marker":"[39]"},{"why":"Supplies the Mellin symbol of the fractional Laplacian via a conformal/extension method.","marker":"[8]"},{"why":"Gives a more recent Mellin definition of the fractional Laplacian used to verify the symbol.","marker":"[38]"},{"why":"Provides the formal Fourier-side eigenfunctions $\\zeta^{2sk}e^{-\\zeta^{2s}}$ from the physics literature, which the paper makes rigorous.","marker":"[44, 45]"},{"why":"Gives the asymptotic expansion of the Fox–Wright function used to establish the algebraic decay of non-integer eigenfunctions.","marker":"[52]"},{"why":"Supplies the Bessel and Legendre integral identities used to evaluate the critical case $s=1/2$.","marker":"[22]"},{"why":"Provides the ODE uniqueness characterization in the local case that motivates the Mellin-based uniqueness proof.","marker":"[36]"}],"fun_headline_variants":["Mellin transform gives explicit eigenfunctions for fractional heat","Fractional Ornstein-Uhlenbeck has pure point spectrum","Explicit basis for fractional Fokker-Planck via Mellin multipliers","Laguerre polynomials expand fractional heat solutions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The completeness claim rests on the uniqueness step Proposition 5.22, which assumes membership in $L^2_\\dagger$ forces membership in every higher energy space $H^k_\\dagger$ and then rules out nonconstant periodic multipliers; if that step fails, extra eigenfunctions may exist.","fun_headline_variants_meta":{"raw":{"variants":["Mellin transform gives explicit eigenfunctions for fractional heat","Fractional Ornstein-Uhlenbeck has pure point spectrum","Explicit basis for fractional Fokker-Planck via Mellin multipliers","Laguerre polynomials expand fractional heat solutions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000871,"raw_usage":{"total_tokens":3798,"prompt_tokens":1001,"completion_tokens":2797,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":617,"completion_tokens_details":{"reasoning_tokens":2728}},"tokens_in":617,"tokens_out":2797,"duration_ms":20019,"temperature":1.0,"reasoning_tokens":2728,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:10:18.133981+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Solve the Mellin functional equation (5.2) for $n=1$, $\\nu=1/2$ and check whether the resulting function lies in $L^2_\\dagger$; a nonzero admissible solution would contradict Proposition 5.22 and break the claimed completeness of the basis.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the compactness and spectral theorem for the local weighted Fokker–Planck problem, which is transported to the fractional setting."},{"cited_title":"Escobedo, O","cited_arxiv_id":null,"evidence_quote":"Provides the local self-similar heat-equation framework whose eigenbasis is the starting point for the fractional analogue."},{"cited_title":"Paris, D","cited_arxiv_id":null,"evidence_quote":"Gives the Mellin transform, inversion theory, and Mellin–Barnes integral tools used throughout."},{"cited_title":"DelaTorre, M","cited_arxiv_id":null,"evidence_quote":"Supplies the Mellin symbol of the fractional Laplacian via a conformal/extension method."},{"cited_title":"Pagnini, C","cited_arxiv_id":null,"evidence_quote":"Gives a more recent Mellin definition of the fractional Laplacian used to verify the symbol."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the asymptotic expansion of the Fox–Wright function used to establish the algebraic decay of non-integer eigenfunctions."},{"cited_title":"S.; Ryzhik, I","cited_arxiv_id":null,"evidence_quote":"Supplies the Bessel and Legendre integral identities used to evaluate the critical case $s=1/2$."},{"cited_title":"Mizoguchi","cited_arxiv_id":null,"evidence_quote":"Provides the ODE uniqueness characterization in the local case that motivates the Mellin-based uniqueness proof."}],"review_version":1}