{"id":"7c994ba0-5e77-4ba4-8195-553fc312ac70","arxiv_id":"1908.08106","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For powers 2 ≤ p < 7/3, two sufficiently close radial positive energy minimizers of the generalized Choquard functional must be identical, and minimizing the energy is equivalent to saturating the Gagliardo-Nirenberg inequality.","lead":"This paper proves that energy-minimizing solitary waves of the generalized Choquard equation are locally unique, and that these minimizers coincide with the optimizers of a sharp Gagliardo-Nirenberg inequality. The result supplies a missing technical step in the orbital stability program for this family of nonlinear Schrödinger-type equations.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 4.1's kernel decay estimate is under-proved: the c1=0 case assumes two zeros of h and imports Lemma 4.1 from [4], yet Theorem 1's analytic continuation depends on the resulting bound |h|/Q < infinity.","rationale":"The central claim of Theorem 1 is a local uniqueness statement for radial positive minimizers. The proof reduces non-uniqueness to a nonzero h in Ker L+, then studies the analytic function K(z) = E_p(sqrt(sigma)(Q+zh)/sqrt(sigma+z^2)). For K to be analytic in a neighborhood of 0 and on the strip Lambda_delta, one needs Re(1+zh/Q) > 1/2 and appropriate log-branch choices; the bound |h|/Q <= C from Proposition 4.1 is exactly what supplies this. The same bound is used for the dominated-convergence limit along z = R + iR. Hence any unproved step in Proposition 4.1 is directly load-bearing. I agree with the Reader that Proposition 4.1 is the weakest assumption. The manuscript's proof of that proposition is not self-contained: it imports Lemma 4.1 from [4] without restating it, asserts the asymptotic expansions (4.6)-(4.8) rather than proving them, and the maximum-principle argument in the c1 = 0 case assumes two zeros of h, while orthogonality only gives one. This is a genuine gap in the written proof, but I do not see a contradiction that would force rejection; the proposition may well be true and the gap repairable. The separate exponent issue in (1.15) affects Section 3 but not Theorem 1, so it is not the primary concern. The Reader's CONDITIONAL verdict is therefore unchanged: the paper should be accepted only after Proposition 4.1 and the imported lemma are fully justified.","tokens_in":11380,"tokens_out":48161,"duration_ms":483405,"concrete_test":"Independently re-derive Proposition 4.1 from (4.1)-(4.4) without invoking Lemma 4.1 of [4], supplying a rigorous treatment of the c1 = 0 case either by proving that any nontrivial radial H^1 solution has at least two large zeros (for example, via a Sturm comparison argument on the system (5.2)) or by proving directly from the integral inequalities that g and B vanish without the two-zero maximum principle. If the two-zero assumption is essential, numerically search for a nonzero radial solution of the linearized system with c1 = 0, d1 != 0 and exactly one zero, and compute sup_{r>0} |h(r)|/Q(r) to see whether the analytic continuation bound survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Proposition 4.1 is load-bearing: it supplies the bound |h|/Q <= C used in Section 2 to justify analyticity of D((Q+zh)^p) near 0 and on the strip Lambda_delta, and to pass to the limit z = R + iR. The proof of Proposition 4.1 is not complete. In the case c1 = 0, the maximum-principle argument on (4.9) assumes h has two zeros r1 < r2. But h is orthogonal to Q and Q > 0 only guarantee that h has at least one zero; an eventually sign-definite h tending to 0 can have a single zero and no interior negative minimum, so the stated contradiction does not apply. The subsequent conclusion c1 = d1 = 0 also depends on Lemma 4.1 imported from [4] without statement or proof, and on asymptotic expansions (4.6)-(4.8) that are asserted rather than derived. If a nonzero kernel element with c1 = 0, d1 != 0 exists, the ratio |h|/Q may still be bounded, but the proof as written does not establish the dichotomy needed for the analytic continuation. If instead the true decay rate of h were e^{-(p-2)r}, then |h|/Q would grow like e^{(3-p)r} and the argument in Theorem 1 would collapse. A complete proof of Proposition 4.1 is therefore a prerequisite for the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies the generalized Choquard equation i∂tu + Δu + I(|u|p)|u|p−2u = 0 in R3 and, in particular, the local uniqueness of radial positive minimizers of the constrained energy Ep(u)=12‖∇u‖L22−12pD(|u|p,|u|p) under ‖u‖L22=σ. Theorem 1 asserts that, for 2≤p<7/3, any two such minimizers that are sufficiently close in H1rad must coincide. The proof strategy is to reduce local uniqueness to directional uniqueness along elements h of the kernel of the linearized operator L+; if a direction h gives a flat expansion of the normalized energy, the function K(z) obtained by replacing ε by a complex parameter z is shown, via analytic continuation on a domain Ωδ, to be constant, and taking z→∞ along the diagonal yields Ep(√σh)=Ep(Q), making √σh a minimizer. A zero of h coming from orthogonality h⊥Q then contradicts Lemma 5.1. Theorem 2 characterizes minimizers of Eσ as exactly those functions attaining equality in the Gagliardo–Nirenberg inequality with best constant C∗. The paper also contains auxiliary ODE lemmas and Fuchs–Painleve series arguments.","tokens_in":11644,"tokens_out":30166,"duration_ms":357019,"significance":"If the proof can be completed, this is a valuable contribution: it gives a local uniqueness statement for a nonlocal Choquard model in a range of p where classical Sturm comparison arguments are not available, and it does so without requiring non-degeneracy of the linearized operator L+. The analytic-continuation mechanism is quite elegant, and the link established in Theorem 2 between constrained minimizers and equality cases of the Gagliardo–Nirenberg inequality is useful in its own right. However, the central uniqueness theorem is currently not rigorously established, because the key decay estimate for elements of the kernel of L+ (Proposition 4.1) has a substantive gap in its proof. Since that estimate is used to justify the analytic continuation and the limit along z=R+iR, the main claim is not yet proven as written.","major_comments":[{"comment":"The proof of the case c1=0, d1≠0 is incomplete. The argument assumes that h(r)=0 has two roots r2>r1>r0 and then applies the maximum principle to the interval [r1,r2]. However, orthogonality h⊥Q with Q>0 only guarantees that h has at least one zero; h may have exactly one zero and, for example, be negative near the origin and positive on the tail. In that situation no interval with two boundary zeros exists, and the maximum principle applied to equation (4.9) gives no contradiction. This step is what forces c1=d1=0 and hence the decay estimate |h(r)|≲e−r/r in (4.5). That estimate is used in Section 2 to justify the bound |h|/Q≤C, the analytic continuation of K(z), and the limit along z=R+iR. If h instead decays like e−(p−1)r, then |h|/Q grows like e−(p−2)r and the whole argument in Theorem 1 collapses. A complete treatment of the single-zero and eventually-positive case is therefore required.","section":"Section 4, Proposition 4.1"},{"comment":"Even apart from the two-zero issue, the reduction from the system (4.1) to the integral inequalities used with Lemma 4.1 is only sketched. The asymptotic expansions (4.6)–(4.8) are asserted as 'verified in a similar way' without a derivation, and the boundedness of ψ=|g|+|B| as well as the exact form of the inequality (4.11) are not checked. For p>2 the exponential factors e−(p−2)s and e−ps make the reduction plausible, but for p=2, which is included in the statement of Theorem 1, the factor e−(p−2)s is identically 1 and the displayed estimates do not lead to (4.11). The manuscript should either give a complete proof of Proposition 4.1 or explicitly restrict the proof to p>2 and quote the classical p=2 uniqueness result from the introduction.","section":"Section 4, equations (4.6)–(4.11)"}],"minor_comments":[{"comment":"The final compactness step, where a sequence hk in the kernel of L+ is replaced by a limit h∗ with ‖h∗‖L2=1, is not stated explicitly; since the kernel is finite-dimensional (Lemma 5.2), this step is valid but should be spelled out.","section":"Section 2, proof of Theorem 1"},{"comment":"The proof of Lemma 5.1 only discusses the case u′(r0)<0 and omits the symmetric case u′(r0)>0; in the case u′(r0)=0 it invokes uniqueness for the Cauchy problem without stating the required regularity or Lipschitz condition on V(|u|)u. This is likely easy to repair, but the lemma is used for the final contradiction in Theorem 1 and should be proved completely.","section":"Section 5, Lemma 5.1"},{"comment":"The proof of Theorem 2 assumes the existence of a minimizer v for the functional Fσ in (1.13). Since Fσ has infimum 0 and the existence of extremals for the Gagliardo–Nirenberg inequality is nontrivial, the paper should provide a reference or a short argument for this existence.","section":"Section 3, Theorem 2"},{"comment":"The title and abstract promise orbital stability of solitary waves, but the paper actually proves local uniqueness of minimizers and a characterization of best constants; the connection to orbital stability is only via the classification in [3]. The wording should be adjusted to match the content of the results.","section":"Title and abstract"}],"recommendation":"major_revision","confidential_remarks":"The analytic-continuation method is genuinely interesting, and the main theorem would be a solid contribution if the kernel-decay estimate were proved rigorously. At present the proof of Proposition 4.1 has a real gap in a load-bearing place, so I cannot recommend acceptance. I also want to flag that the proof as written does not actually cover p=2, despite the theorem statement; the authors should either invoke [7] explicitly or restrict the statement. Finally, the dependence on Lemma 4.1 from the authors' companion work [4] should be made fully transparent, ideally by including a complete proof of that lemma in this paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the paper takes a genuinely open step—local uniqueness of radial ground states for the generalized Choquard equation with p>2—and the main idea is good. The analytic-continuation trick, extending K(z) to a domain where it must be constant and then pushing z to infinity along R+iR, is a real departure from the usual Sturm/shooting approach, and it does not require the kernel of L+ to be trivial. The equivalence between the two variational problems (Theorem 2) and the formula for the Gagliardo-Nirenberg constant are clean and useful. This deserves a serious referee.\n\nThe soft spots are concentrated in Section 4. Proposition 4.1 is load-bearing: the bound |h|/Q ≤ C is what justifies analyticity of D((Q+zh)^p) and the limit z=R+iR. The proof as written is not complete. In the c1=0 case, the maximum-principle argument assumes h has two zeros r1<r2 so that it has an interior negative minimum. But h⊥Q and Q>0 only guarantee at least one zero; an eventually sign-definite h tending to zero with a single zero is compatible with the assumptions, and the stated contradiction does not apply. The subsequent conclusion c1=d1=0 also imports Lemma 4.1 from [4] without statement or proof, and the asymptotic expansions (4.6)–(4.8) are asserted rather than derived. If c1=0 and d1≠0 is possible, the ratio |h|/Q might still be bounded, but the proof as written does not establish that; if h decayed like e^{-(p-2)r}, the argument in Theorem 1 collapses. This is a gap, not a minor omission.\n\nSmaller issues: equation (1.15) has an exponent typo—it should be β^{1-γ} on the right, not β^{γ-1}; the derivation in the proof uses the corrected version. The boundary with your own paper [4] is not drawn explicitly; the title of [4] suggests a uniqueness result that may overlap with Theorem 1, so the authors should state exactly what is new. Lemma 5.2 on dim(Ker L+) is plausible but only sketched; a few lines showing how Fuchs–Painleve local solvability yields a global dimension bound would help.\n\nNone of this changes my view that the result is probably true and the method is publishable. The right reader is someone working on uniqueness and stability for Choquard-type equations. Send it to a serious referee, but the referee should be told to demand a complete proof of Proposition 4.1 and a fixed (1.15).","headline":"A clever analytic-continuation proof of local uniqueness for generalized Choquard minimizers, but the central decay estimate is under-proved and must be fixed before the theorem is solid.","tokens_in":12188,"tokens_out":6554,"would_cite":false,"duration_ms":156521,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37K40","35Q55","35Q51"],"pacs":[],"model":"deepseek-v4-flash","headline":"Choquard ground states are locally unique for 2 ≤ p < 7/3","keywords":["generalized Choquard equation","local uniqueness","ground states","orbital stability","Riesz potential","Gagliardo-Nirenberg inequality","radial minimizers","analytic continuation"],"falsifier":"Follow the unique radial positive ground state at $p = 2$ numerically and continue it in $p$ up to $7/3$: if a second radial positive normalized minimizer branches off at some $p \\in [2, 7/3)$, Theorem 1 is false. A cheaper test is to solve the linearized equation $L_+ h = 0$ radially and check whether a nontrivial $h$ obeys $|h(r)| \\lesssim e^{-r}/r$ with a nonzero first coefficient; a numerical counterexample to that decay would invalidate Proposition 4.1 and with it the proof.","tokens_in":11152,"feed_emoji":"🌊","tokens_out":5927,"duration_ms":53898,"temperature":0.7,"pith_summary":"This paper tries to prove that, for the generalized Choquard equation in three dimensions, ground states with fixed $L^2$ mass are locally unique for $2 \\le p < 7/3$: any two radial positive minimizers that are close in $H^1_{\\mathrm{rad}}$ must be identical. The point is that local uniqueness is the missing ingredient for orbital stability of the standing waves. The proof avoids the classical Sturm oscillation and shooting arguments, which break down because the nonlinearity is nonlocal, and instead studies an analytic continuation of the energy along directions in the kernel of the linearized operator. If Theorem 1 holds, orbital stability in the range $p \\in (5/3, 7/3)$ follows from existing stability machinery, because the relevant ground-state branch cannot split.","feed_headline":"Choquard ground states are locally unique for 2 ≤ p < 7/3","feed_subtitle":"The proof closes a gap in orbital stability: close minimizers must coincide, so standing waves cannot branch.","key_machinery":"The load-bearing object is the analytic continuation $K(z) = E_p\\!\\left(\\sqrt{\\sigma}\\, (Q + z h)/\\|Q + z h\\|_{L^2}\\right)$, where $h$ is a nontrivial element of the kernel of $L_+ = -\\Delta + \\omega - p I(Q^{p-1}\\,\\cdot)Q^{p-1} - (p-1) I(Q^p)Q^{p-2}$. The analyticity domain is built with the wedge $\\Lambda_\\delta$ near the diagonal $\\operatorname{Re} z = \\operatorname{Im} z$, using Proposition 4.1: every nonzero kernel element $h$ has $|h(r)| \\lesssim e^{-r}/r$, so the quotient $h/Q$ is bounded and the $p$-th power $(1 + z h/Q)^p$ can be complex-analytically continued without crossing the branch cut of the logarithm. Since the no-branching assumption makes all derivatives of $K$ vanish at $z = 0$, $K$ is constant; letting $z = R + iR$ and using dominated convergence identifies $E_p(\\sqrt{\\sigma} h) = E_p(Q)$, which contradicts the radial ODE lemma. The kernel-dimension bound $\\dim(\\ker L_+) \\le 2$, proved by a Fuchs–Painlevé series expansion, turns the pointwise directional argument into a uniform $\\varepsilon_0$.","core_discovery":"On its own terms, the paper's central discovery is Theorem 1: under $2 \\le p < 7/3$, there is $\\varepsilon > 0$ such that any two radial positive minimizers $Q_1, Q_2 \\in H^1_{\\mathrm{rad}}$ of the constrained energy $E_\\sigma$ with $\\|Q_1 - Q_2\\|_{H^1_{\\mathrm{rad}}} \\le \\varepsilon$ are equal. The proof proceeds by contradiction: if two minimizers approached each other, their difference would converge to a nonzero element $h$ of the kernel of the linearized operator $L_+$; the authors then define $K(z)$, an analytic continuation of the energy along the normalized family $(Q + z h)/\\|Q + z h\\|_{L^2}$, show $K(z)$ is constant in a domain of analyticity, and take the limit along the diagonal $z = R + iR$ to reach $E_p(\\sqrt{\\sigma} h) = E_p(Q)$. Orthogonality of $h$ to $Q$ then gives a zero of $h$ at some radius, and a simple ODE lemma forces $h \\equiv 0$, a contradiction. The paper also proves Theorem 2, an equivalence between minimizers of the Choquard energy and of the associated Gagliardo–Nirenberg functional.","pith_inferences":["The same analytic-continuation device might transfer to other nonlocal equations with different kernels, as long as the kernel elements have enough decay to bound $h/Q$.","The theorem concerns radial positive minimizers with $2 \\le p < 7/3$; a quantitative version of Proposition 4.1, with constants depending only on $p$ and $\\sigma$, would turn the qualitative $\\varepsilon$ into an explicit radius of uniqueness, but the paper does not do this.","Theorem 2 together with Corollary 3.1 gives a formula for the sharp Gagliardo–Nirenberg constant in terms of the ground-state energy; if accurate numerical values of $E_\\sigma$ become available, they would provide an independent test of the uniqueness claim."],"forward_implications":["If Theorem 1 is correct, each radial positive ground state of the generalized Choquard equation in the range $2 \\le p < 7/3$ is an isolated point among radial minimizers of fixed $L^2$ mass.","Combined with the linearized stability classification, this supplies the nonlinear orbital stability of the corresponding standing waves for $p \\in (5/3, 7/3)$: no nearby distinct minimizer can act as a competing orbit.","Theorem 2 implies that, for $p \\in (5/3, 7/3)$, the minimizers of the Choquard energy and the sharp Gagliardo–Nirenberg functional coincide, so the profile of the optimizer is characterized independently of which variational problem one solves.","The analytic-continuation argument only needs decay of kernel elements, not spectral nondegeneracy, so it tolerates a kernel of dimension up to 2; this is a new route to local uniqueness for nonlocal equations."],"supporting_citations":[{"why":"Supplies the integral lemma used at the end of Proposition 4.1 to force a nontrivial kernel element to vanish, closing the contradiction.","marker":"[4]"},{"why":"Establishes the classical uniqueness result for the case $p = 2$, which Theorem 1 extends to the range up to $7/3$.","marker":"[7]"},{"why":"Provides the asymptotic expansions for radial ground states and Riesz potentials used to derive the exponential decay of kernel elements.","marker":"[11]"},{"why":"Exemplifies the Sturm/shooting uniqueness method for local nonlinearities that the paper cannot use directly because of the nonlocal Riesz potential.","marker":"[8]"},{"why":"Gives the linearized stability/instability classification that sets the range $p \\in (5/3, 7/3)$ where orbital stability is the question.","marker":"[3]"},{"why":"Shows the alternative Weinstein-functional Taylor-expansion route whose nondegeneracy condition the present proof avoids.","marker":"[1]"},{"why":"Provides the Fuchs–Painlevé series theorem used to bound the kernel dimension of $L_+$ by 2.","marker":"[5]"},{"why":"Refines uniqueness arguments for local nonlinearities, serving as background contrast for why the nonlocal equation needs a different approach.","marker":"[6]"}],"fun_headline_variants":["Choquard ground states unique locally for p < 7/3","Local uniqueness of minimizers in generalized Choquard","Orbital stability via local uniqueness for Choquard","Choquard standing waves cannot branch: local uniqueness","Proof: Choquard minimizers coincide in H1-radial ball"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Proposition 4.1 claims that every nonzero kernel element $h$ decays like $e^{-r}/r$ and is nonzero for all large radii; the analytic continuation of $K(z)$ and the diagonal limit both lean on the bound $h/Q \\le C$, so if that decay estimate fails the contradiction in Theorem 1 collapses.","fun_headline_variants_meta":{"raw":{"variants":["Choquard ground states unique locally for p < 7/3","Local uniqueness of minimizers in generalized Choquard","Orbital stability via local uniqueness for Choquard","Choquard standing waves cannot branch: local uniqueness","Proof: Choquard minimizers coincide in H1-radial ball"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000969,"raw_usage":{"total_tokens":4122,"prompt_tokens":943,"completion_tokens":3179,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":559,"completion_tokens_details":{"reasoning_tokens":3094}},"tokens_in":559,"tokens_out":3179,"duration_ms":22991,"temperature":1.0,"reasoning_tokens":3094,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:52:23.257956+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Follow the unique radial positive ground state at $p = 2$ numerically and continue it in $p$ up to $7/3$: if a second radial positive normalized minimizer branches off at some $p \\in [2, 7/3)$, Theorem 1 is false. A cheaper test is to solve the linearized equation $L_+ h = 0$ radially and check whether a nontrivial $h$ obeys $|h(r)| \\lesssim e^{-r}/r$ with a nonzero first coefficient; a numerical counterexample to that decay would invalidate Proposition 4.1 and with it the proof.","supporting_citations":[{"cited_title":"Georgiev, M","cited_arxiv_id":null,"evidence_quote":"Supplies the integral lemma used at the end of Proposition 4.1 to force a nontrivial kernel element to vanish, closing the contradiction."},{"cited_title":"Lieb, Existence and uniqueness of the minimizing solution of Cho quard’s nonlinear equation, Stud","cited_arxiv_id":null,"evidence_quote":"Establishes the classical uniqueness result for the case $p = 2$, which Theorem 1 extends to the range up to $7/3$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the asymptotic expansions for radial ground states and Riesz potentials used to derive the exponential decay of kernel elements."},{"cited_title":"McLeod and J","cited_arxiv_id":null,"evidence_quote":"Exemplifies the Sturm/shooting uniqueness method for local nonlinearities that the paper cannot use directly because of the nonlocal Riesz potential."},{"cited_title":"Georgiev and A","cited_arxiv_id":null,"evidence_quote":"Gives the linearized stability/instability classification that sets the range $p \\in (5/3, 7/3)$ where orbital stability is the question."},{"cited_title":"Chang, S","cited_arxiv_id":null,"evidence_quote":"Shows the alternative Weinstein-functional Taylor-expansion route whose nondegeneracy condition the present proof avoids."},{"cited_title":"Hille, Ordinary differential equations in the complex domain, repr int of the 1976 original","cited_arxiv_id":null,"evidence_quote":"Provides the Fuchs–Painlevé series theorem used to bound the kernel dimension of $L_+$ by 2."},{"cited_title":"Kwong, Uniqueness of positive solutions of ∆ u − u + up = 0, Arch","cited_arxiv_id":null,"evidence_quote":"Refines uniqueness arguments for local nonlinearities, serving as background contrast for why the nonlocal equation needs a different approach."}],"review_version":1}