{"id":"5a7e42db-8458-46ea-9c5c-44311addd164","arxiv_id":"2509.01484","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For bounded potentials whose spatial derivative grows at most like |x|^δ with δ<1, the 1D quantum harmonic oscillator is reducible in L^2 for a Cantor set of frequencies of asymptotically full measure.","lead":"This mathematics paper proves that a quantum harmonic oscillator with a bounded, time-quasi-periodic potential can be reduced to a constant-coefficient system when the potential's spatial derivative grows slower than |x|. The result weakens earlier decay conditions and moves toward an open problem posed by Eliasson in 2011.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proof of Lemma 2.2(c), used to place the KAM transformations on ℓ²_p, contains an invalid inequality; without a fix the L²/H^p statement is unsupported.","rationale":"The reader's stated weakest assumption is the diagonal-decay estimate Lemma 2.6, but that estimate appears carefully derived and plausible. My reading of the manuscript identifies a different, more immediate gap: the proof of Lemma 2.2(c), a structural lemma that is reused at every KAM step and in the final limiting argument to place the transformations on ℓ²_p and to justify the conjugation identity in B(ℓ²,ℓ²_{−2}). The displayed inequality in Appendix A is demonstrably false for a simple M_+-matrix, so the written proof does not establish the needed weighted boundedness. The main KAM iteration and measure estimates are not the issue; the issue is the bridge from the formal infinite-matrix calculation to the L²/H^p statement of the main theorem. Since this can likely be repaired without changing the main ideas, the appropriate disposition is CONDITIONAL acceptance pending a corrected proof of Lemma 2.2(c).","tokens_in":25569,"tokens_out":48650,"duration_ms":505712,"concrete_test":"Independently settle Lemma 2.2(c): prove or disprove that A,[N,A]∈B(ℓ²) implies N^{s/2} A N^{-s/2} is bounded for s=2 and s=−2. As a minimal check, re-run the Appendix A Cauchy–Schwarz step with the rank-one matrix A_1^2=1, u_2=1, s=2; the displayed inequality is false, so the written proof must be modified. If the lemma is true, supply a corrected proof (for example via a weighted Schur test using Lemmas B.1 and B.3, or via interpolation); if it is false, the KAM control of U in B(ℓ²_p) and the B(ℓ², ℓ²_{−2}) conjugation identity fail.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Appendix A's proof of Lemma 2.2(c) is not valid as written. In the s>0 part, the first displayed inequality\n  Σ_i i^s |Σ_j A_i^j u_j|²  ≤  Σ_i ( Σ_j (A_i^j/(1+|i−j|)) (i/j)^{s/2} j^{s/2}|u_j| )²\nfails already for the rank-one matrix A_1^2=1, u_2=1, s=2: the LHS is 1 and the RHS is 1/4. Inserting 1/(1+|i−j|) on A_i^j shrinks the summand, and the following Cauchy–Schwarz line places the denominator on the wrong factor. The same problem affects the s<0 estimate. This is not cosmetic: Lemma 2.2(c) is invoked in Lemma 3.2 to conclude that exp(S_m) acts boundedly on ℓ²_p, and at the end of §3.3 to assert that the conjugation identity holds in B(ℓ², ℓ²_{−2}). Those are exactly the steps that turn the formal matrix KAM iteration into the L²-reducibility and H^p-regularity claims of Theorems 2.4 and 1.1. The statement of Lemma 2.2(c) may be true and repairable, but the argument supplied does not prove it.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a reducibility theorem for the one-dimensional quantum harmonic oscillator with a time quasi-periodic, bounded potential whose spatial derivative grows at most like |x|^δ, δ<1. Working in the Hermite basis, the authors represent the perturbation as an infinite matrix and show that its difference matrix decays along the diagonal; a KAM iteration then conjugates the system to an autonomous diagonal equation on a Cantor set of asymptotically full measure in the frequency parameter. The main results include L^2 reducibility, H^p regularity of the transformation for p∈[0,2], almost-periodic solutions, and pure point spectrum of the Floquet operator. The proof is structured as an abstract matrix reducibility theorem (Theorem 2.4) with a verification of the matrix-space hypotheses for the Schrödinger potential (Lemma 2.6).","tokens_in":25975,"tokens_out":27467,"duration_ms":298766,"significance":"If fully established, the result gives a substantial step toward Eliasson's question on bounded perturbations of the harmonic oscillator: it removes the decay assumption on the potential itself and replaces it by a mild growth condition on its derivative, with an explicit measure estimate. The key technical novelty, the difference-matrix decay estimate (Lemma 2.6), is well targeted and goes beyond earlier work in [24,25]. The KAM framework is standard, but the paper provides explicit constants and uses a useful separation of the operator-norm control and the element-decay control. The auxiliary lemmas are largely self-contained and the main theorem is stated with quantitative bounds. However, two proof gaps in the current version—the invalid proof of Lemma 2.2(c) and the parameter handling in Proposition 3.1—must be repaired before the result is fully supported.","major_comments":[{"comment":"The first displayed inequality in the s>0 part of the proof is false. It asserts that Σ_i i^s |Σ_j A_i^j u_j|² is bounded by a sum in which A_i^j is replaced by A_i^j/(1+|i−j|); this factor shrinks the summand and cannot be introduced by the triangle inequality. For the rank-one matrix A_1^2=1 with u_2=1, s=2, the left side is 1 and the right side is 1/4. The same defect occurs in the s<0 part. This is load-bearing because Lemma 2.2(c) is used in Lemma 3.2 to show exp(S_m)∈B(ℓ²_p) and in §3.3 to justify the conjugation identity in B(ℓ²,ℓ²_{−2}). The statement is salvageable: from N A = A N + [N,A] one gets ||N A u|| ≤ ||A|| ||u||_1 + ||[N,A]|| ||u|| ≤ (||A||+||[N,A]||)||u||_2, and duality/interpolation covers s∈[−2,2]. The proof must be rewritten.","section":"Appendix A, proof of Lemma 2.2(c)"},{"comment":"The proof of the measure estimate (3.3) is not logically consistent as written. The symbol D is reused for the input domain and for the Diophantine set D(γ,K) from Lemma B.7, and then the text says \"Setting κ = γ^{1+2/α}\", although κ is an input parameter and γ is a separate parameter. The claimed bound Meas(D\\D') ≤ Cκ^{ν1}K^{ν2} does not follow without choosing the relation between γ and κ. The repair is straightforward: introduce D_γ from Lemma B.7, set D' = D ∩ D_γ \\ F, choose γ = κ^{α/(α+2)} for κ small, and use Meas(D\\D') ≤ Meas(Π\\D_γ)+Meas(F). Since Lemma 3.2 relies on (3.3) for the measure decay (3.32), this point must be corrected; it is local but load-bearing.","section":"Section 3.1, proof of Proposition 3.1"}],"minor_comments":[{"comment":"In the definition of F, the union is over i,j∈Z but should be over i,j∈N (diagonal indices in the matrix spaces).","section":"Section 3.1, definition of F"},{"comment":"The formula for σ_m−σ_{m+1} is garbled in the displayed line; it should be σ_0(m+1)^{-2}/(2 Σ_{i≥1} i^{-2}) so that the total loss is σ_0/2.","section":"Section 3.2, parameter choices"},{"comment":"The growth condition (1.6) writes ln^{δ'}(2+|x|); this should be (ln(2+|x|))^{δ'} to be meaningful.","section":"Remark 1.4"},{"comment":"The estimate |⟨V' h_{i+1}, h_j⟩| ≤ C(i∧j)^{δ/2} is used without comment. It follows by taking the minimum of the two Cauchy–Schwarz bounds ||(1+|x|)^δ h_{i+1}|| ||h_j|| and ||h_{i+1}|| ||(1+|x|)^δ h_j||, together with Lemma B.4. State this explicitly.","section":"Lemma 2.6"},{"comment":"The statement has a typo: H should be −d²/dx² + x² (the second derivative is missing a square).","section":"Appendix B, Lemma B.4"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a natural and timely problem, and the overall strategy appears sound. The main obstacles are two technical but fixable gaps: the invalid proof of Lemma 2.2(c) and the γ/κ parameter confusion in Proposition 3.1. I see no reason to doubt the authors' claims, but the current text does not yet constitute a complete proof of the advertised L^2/H^p reducibility. I recommend major revision and encourage the authors to repair these points carefully."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does what it claims: it replaces the |x|^{-1} decay condition on ∂_xV from [24] with |x|^δ growth for any δ<1, and it does so by a mechanism that is actually new. The split between operator-norm estimates for the homological equation and element-decay estimates for the measure estimates is a sensible, non-obvious combination. The main theorem is a genuine step toward Eliasson's open problem, not a solution; the significance estimate 6 is fair for a one-dimensional result with bounded potentials only on the derivative.\n\nLemma 2.6 is the technical heart, and it checks out. The estimate (2.6) deserves the emphasis: it converts derivative growth into the diagonal decay α=(1−δ)/2 needed for the Melnikov conditions. The KAM iteration itself is standard in outline, with explicit exponents; the measure estimates are handled carefully. No fitted parameters, no circularity, and the citations to [24,25,29] are legitimate framework citations rather than a hidden reduction.\n\nThe soft spot is the proof of Lemma 2.2(c) in Appendix A. The stress-test note is correct: the first displayed inequality in the s>0 part is false, and the rank-one example with A_1^2=1, u_2=1, s=2 gives LHS 1 and RHS 1/4. The same problem appears in the s<0 estimate. I want to be clear about the consequence. The lemma statement itself is true, and the gap is repairable in a few lines: use \\bar A_i^j=(1+|i−j|)|A_i^j|, apply Cauchy-Schwarz as |A_i^j| = \\bar A_i^j/(1+|i−j|), bound sup_i Σ_j \\bar A_i^j² by Lemma B.1, and then use Lemma B.3. That removes the invalid inequality. So this is a local proof error, not a load-bearing flaw in the theorem. But as printed, Appendix A does not prove the lemma, and since Lemma 2.2(c) is used to place the conjugation in B(ℓ2,ℓ2_{-2}), the authors should fix it before publication.\n\nOne minor notational stumble: in the proof of Proposition 3.1, D is used both for the current domain and for the good set from Lemma B.7; the intended construction D'=D∩D_γ\\F is clear enough.\n\nI did not machine-check every inequality; the reader's MODERATE confidence seems right. If I were the editor, I would send this to a serious referee. The right outcome after the Appendix repair is acceptance; this is a meaningful, honest improvement in a hard problem.","headline":"A genuine step toward Eliasson's open problem, with one local proof gap in Appendix A that is easily repaired; worth refereeing.","tokens_in":26411,"tokens_out":18560,"would_cite":true,"duration_ms":204871,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35P05","37K55","81Q15"],"pacs":[],"model":"deepseek-v4-flash","headline":"For all δ<1 and almost every frequency, a bounded quasi-periodic potential with derivative growth |x|^δ can be removed from the 1D quantum harmonic oscillator by a unitary change of variables.","keywords":["reducibility","quantum harmonic oscillator","quasi-periodic potential","KAM theory","Hermite basis","difference matrix","Melnikov conditions","pure point spectrum"],"falsifier":"For any potential satisfying (1.2), compute the Hermite difference matrix ΔP_i^j = ∫ V(h_{i+1}h_{j+1}-h_i h_j) dx. The proof's key bound (2.6) claims this is O((i∧j)^{(δ-1)/2}); for V(x,θ)=cos(x−θ_1) it predicts O((i∧j)^{-1/2}). A numerical or analytic check that this quantity decays slower than any positive power for such a potential would falsify Lemma 2.6 and thereby the KAM mechanism.","tokens_in":25509,"feed_emoji":"⚛️","tokens_out":10467,"duration_ms":112544,"temperature":0.7,"pith_summary":"The paper proves that a one-dimensional quantum harmonic oscillator remains reducible when perturbed by a time-quasi-periodic potential that is only bounded, provided the potential's spatial derivative grows no faster than |x|^δ with δ<1. Reducibility means that for almost every frequency vector (in a Cantor set of asymptotically full measure), a unitary, real-analytic change of variables removes the time dependence completely, leaving a diagonal constant-coefficient equation. This matters because it relaxes a long-standing boundary: earlier results required spatial decay of the perturbation or of its derivative, while here the derivative is allowed to grow almost linearly. The proof works by moving from decay in the eigenvalues and tails of the spectrum to decay in the difference matrix of the perturbation in the Hermite basis. If the theorem is correct, the dynamics is almost-periodic, Sobolev norms stay nearly constant, and the Floquet operator has pure point spectrum for those frequencies.","feed_headline":"Quantum oscillator reducible under bounded quasi-periodic potentials","feed_subtitle":"Covers potentials whose spatial derivative may grow like |x|^δ for any δ<1; solutions become almost-periodic.","key_machinery":"The carrying object is the Hermite-basis matrix P(θ) of the potential and its difference matrix ΔP, entries P^{j+1}_{i+1}-P^j_i. The relevant class M_b^α consists of bounded matrices whose difference matrix decays like (i∧j)^{-α}. Lemma 2.6 shows P∈M_b^α with α=(1-δ)/2. The KAM step solves the homological equation [A,S]-i∂_tS=Ã-P+R with S in a companion class M_b^{α+}; Lemma 2.2 supplies the algebra of these classes, controlling operator norms and the diagonal decay that enters the Melnikov measure estimates.","core_discovery":"The paper's central claim (Theorem 1.1): if V is bounded and |∂_xV(x,θ)|≤C(1+|x|)^δ with δ<1, then for all small ε there is a Cantor set Π_ε⊂[0,2π)^n of asymptotically full measure such that for ω∈Π_ε the equation i∂tψ=(-∂xx+x^2+εV(x,ωt))ψ is reducible in L². A unitary, real-analytic conjugacy sends it to an autonomous diagonal equation i∂tφ=H^∞φ, H^∞=diag{λ_i^∞}, with |λ_i^∞-(2i-1)|≤Cε. The conjugacy is C¹ in ω and ε^{2/3}-close to the identity. Consequences: solutions are almost periodic, H^p norms stay within 1+Cε of their initial values, and the Floquet operator has pure point spectrum.","pith_inferences":["The threshold δ<1 enters only through the diagonal decay exponent α=(1-δ)/2, so the method's real requirement is polynomial diagonal decay of the difference matrix; replacing it by logarithmic decay is a natural next step, which the authors flag in Remark 1.4.","If the standing question about purely bounded potentials with no derivative control has a positive answer, this result suggests the mechanism will come from the difference matrix of the perturbation rather than from any decay of the potential itself.","One could directly test the key estimate numerically for oscillating potentials such as cos(x−ωt): Lemma 2.6 predicts |ΔP_i^j|≤C(i∧j)^{-1/2}, and checking whether this decay actually holds would probe the foundation of the KAM argument independently of the iteration.","The abstract reducibility theorem is stated for any diagonal operator with linearly spaced eigenvalues and controlled spacing; transferring the proof to other one-dimensional Schrödinger operators with such spectra is plausible but not carried out here."],"forward_implications":["For every δ<1 and small ε, most frequency vectors in the parameter space make the perturbed oscillator unitarily conjugate to a diagonal autonomous equation.","Solutions of the Cauchy problem are almost periodic and their H^p norms for 0≤p≤2 remain within a factor 1+Cε of the initial norm for all time.","The associated Floquet operator has pure point spectrum on the good frequency set.","The eigenvalues of the reduced equation are ε-close to the unperturbed eigenvalues 2i−1, and the conjugating transformation is within Cε^{2/3} of the identity.","The paper notes that a logarithmic growth assumption on ∂_xV would likely suffice as well, since logarithmic decay of the difference matrix would still control the measure estimates."],"supporting_citations":[{"why":"Supplies the weighted L² bound for Hermite eigenfunctions that turns the growth condition on ∂_xV into diagonal decay of the perturbation matrix.","marker":"[19]"},{"why":"Provides the difference-matrix framework and the previous KAM scheme for perturbations with decaying derivative that this paper extends.","marker":"[24]"},{"why":"Established an earlier reducibility result for bounded perturbations on R^d; the corollaries on almost-periodicity and Floquet spectrum follow its arguments.","marker":"[17]"},{"why":"Introduced the KAM iteration for the quantum harmonic oscillator with polynomial decay, whose spectral and measure estimates are adapted here.","marker":"[18]"},{"why":"Showed the required tail decay can be logarithmic; the measure-estimate strategy for the Melnikov conditions is inherited from it.","marker":"[29]"},{"why":"Gives the creation and annihilation identities for Hermite functions used to express the difference matrix ΔP in terms of integrals involving V'.","marker":"[9]"},{"why":"The classical KAM theorem whose spectral assumption motivates why tail decay was previously needed and where this paper's diagonal-decay substitute enters.","marker":"[28]"}],"fun_headline_variants":["Quantum oscillator reducible under bounded quasi-periodic potentials","Bounded quasi-periodic potentials tame quantum oscillator via KAM","Quantum harmonic oscillator reducible to diagonal autonomous system","KAM reducibility for 1D oscillator with quasi-periodic bounded potential","Quasi-periodic bounded perturbations yield almost-periodic quantum solutions"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"Everything rests on the estimate |ΔP_i^j(θ)| ≤ C(i∧j)^{(δ−1)/2} in Lemma 2.6: if the perturbation's difference matrix does not decay polynomially toward the diagonal, the small-divisor frequency conditions cannot be imposed and the KAM iteration does not close.","fun_headline_variants_meta":{"raw":{"variants":["Quantum oscillator reducible under bounded quasi-periodic potentials","Bounded quasi-periodic potentials tame quantum oscillator via KAM","Quantum harmonic oscillator reducible to diagonal autonomous system","KAM reducibility for 1D oscillator with quasi-periodic bounded potential","Quasi-periodic bounded perturbations yield almost-periodic quantum solutions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000649,"raw_usage":{"total_tokens":2792,"prompt_tokens":695,"completion_tokens":2097,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":439,"completion_tokens_details":{"reasoning_tokens":2010}},"tokens_in":439,"tokens_out":2097,"duration_ms":16151,"temperature":1.0,"reasoning_tokens":2010,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:30:10.012242+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For any potential satisfying (1.2), compute the Hermite difference matrix ΔP_i^j = ∫ V(h_{i+1}h_{j+1}-h_i h_j) dx. The proof's key bound (2.6) claims this is O((i∧j)^{(δ-1)/2}); for V(x,θ)=cos(x−θ_1) it predicts O((i∧j)^{-1/2}). A numerical or analytic check that this quantity decays slower than any positive power for such a potential would falsify Lemma 2.6 and thereby the KAM mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the weighted L² bound for Hermite eigenfunctions that turns the growth condition on ∂_xV into diagonal decay of the perturbation matrix."},{"cited_title":"Liang, Z.Q","cited_arxiv_id":null,"evidence_quote":"Provides the difference-matrix framework and the previous KAM scheme for perturbations with decaying derivative that this paper extends."},{"cited_title":"Grébert, E","cited_arxiv_id":null,"evidence_quote":"Established an earlier reducibility result for bounded perturbations on R^d; the corollaries on almost-periodicity and Floquet spectrum follow its arguments."},{"cited_title":"Grébert, L","cited_arxiv_id":null,"evidence_quote":"Introduced the KAM iteration for the quantum harmonic oscillator with polynomial decay, whose spectral and measure estimates are adapted here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Showed the required tail decay can be logarithmic; the measure-estimate strategy for the Melnikov conditions is inherited from it."},{"cited_title":"Chodosh, Infinite matrix representations of isotropic pseudodifferential operators,Methods Appl","cited_arxiv_id":null,"evidence_quote":"Gives the creation and annihilation identities for Hermite functions used to express the difference matrix ΔP in terms of integrals involving V'."},{"cited_title":"Pöschel, A KAM-theorem for some nonlinear partial differential equations,Ann","cited_arxiv_id":null,"evidence_quote":"The classical KAM theorem whose spectral assumption motivates why tail decay was previously needed and where this paper's diagonal-decay substitute enters."}],"review_version":1}