{"id":"21fa9a34-ea50-492a-870a-1d20fd727390","arxiv_id":"2601.20931","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Rotating the Aubry-André model's position and momentum operators hides its localization transition in a rotated basis, and at the transition the model maps onto a massless Dirac fermion in a curved metric.","lead":"This paper shows that rotating the position and momentum operators of the Aubry-André model hides its metal-insulator transition in a rotated basis rather than ordinary space. It also claims that at the transition the model maps onto a massless Dirac fermion in a curved spacetime metric.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The curved-spacetime equivalence is only patchwise and the reconstructed metric diverges on an infinite set; no global smooth metric is established.","rationale":"The reader's verdict is CONDITIONAL, and the weakest-assumption analysis correctly identifies the curved-spacetime equivalence as the load-bearing but unproven component. I find no reason to move the verdict: the central localization transition at J=±K is supported by the Aubry-Andre duality in the rotated basis and by the numerical IPR/NPR data, so the paper need not be rejected. However, the 'coincides with massless Dirac fermion in curved spacetime' statement is stronger than what Section IV establishes; the paper itself confines the equivalence to patches and admits the metric diverges. This is not merely a stylistic issue, because the advertised 'unexpected relation between localization transitions and analog gravity' depends on a global metric interpretation. The concrete test—checking the Z_2 gauge flux and the growth of alpha_n on finite rings—would settle whether a global reconstruction is possible. The localized/delocalized transition itself is not threatened, so the conditional acceptance stands.","tokens_in":9864,"tokens_out":20575,"duration_ms":212225,"concrete_test":"For rational approximants omega/2pi = F_n/F_{n+1} and generic phi, compute the hopping t_n = 2Jtilde cos(1/2(omega n + phi)) at J=K on a ring of N sites. Check the flux product P_N = prod_{n=1}^N sign(t_n). If P_N is not +1 for arbitrarily large N, no periodic gauge transformation makes t_n positive and the equivalence to the Dirac Hamiltonian with t_n=|cos theta_n| fails globally. Independently, solve alpha_{n+1} = t_n^2 / alpha_n with alpha_1=1 and record M_N = max_{1<=n<=N} |alpha_n|; if M_N diverges as N grows, no bounded global metric exists and the curved-spacetime claim is only patchwise. Either outcome would confirm the reader's concern.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest advertised claim includes the statement that at the phase transition the Hamiltonian 'coincides with the lattice Hamiltonian of a massless Dirac fermion in a curved spacetime background.' This is the least secure part of the paper. In Section IV, the metric is reconstructed from the hopping through the recurrence sqrt(alpha_n alpha_{n+1}) = |sin(1/2(omega n + phi))| or |cos(1/2(omega n + phi))| (Eqs. 12-13). The paper explicitly concedes: the equivalence to H_RAA holds only 'on patches where sin and cos do not change sign,' and the metrics 'diverge for values of the phase ... on lattice sites where m+n is even,' an infinite set for irrational omega/2pi. A lattice Dirac Hamiltonian derived from a smooth curved-spacetime metric requires a globally defined alpha(x); a patchwise, singular alpha_n cannot support the claim of coincidence. Moreover, at J=K the hopping amplitude in H_RAA is proportional to cos(theta_n) with sign changes, whereas the Dirac lattice Hamiltonian H_II uses |cos(theta_n)|. In 1D a sign-changing hopping can be gauged to positive only if the total Z_2 flux around the ring vanishes; for a quasiperiodic sign sequence this is not guaranteed. Thus the analog-gravity conclusion is an overstatement unless a global smooth metric (or a gauge transformation removing all sign defects) is exhibited.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper considers a family of one-dimensional tight-binding Hamiltonians obtained from the Aubry-André Hamiltonian by a linear canonical (rotation) transformation of the position and momentum operators. The main claims are: (i) the 'rotated' Aubry-André model H_RAA exhibits a localization-delocalization transition at J=±K, but the localized/delocalized phases are defined with respect to rotated bases |e_m^±>, not the conventional position basis; (ii) this hidden transition leaves a remnant signature in the conventional position basis, namely a dip of the average NPR that scales to zero polynomially at the transition in the thermodynamic limit (Fig. 2c); (iii) at the transition the lattice Hamiltonian coincides with the lattice Hamiltonian of a massless Dirac fermion in a curved 1+1D spacetime with metric ds^2 = α(x)^2 dt^2 − dx^2, where α_n is reconstructed from the hopping amplitudes through Eqs. (12)–(13). The paper is largely a corollary of the exactly solvable Aubry-André model via the canonical transformation, with the Dirac/spacetime identification as the novel interpretive step.","tokens_in":10219,"tokens_out":3627,"duration_ms":33131,"significance":"If the central claims hold, the paper has clear value for the quasiperiodic-localization community: it generalizes the Aubry-André duality to a family of off-diagonal models, shows that the position-basis NPR (or a rotated-basis IPR/NPR) detects the hidden transition, and draws a suggestive connection between localization transitions and analog gravity. The algebraic core is exact and can be verified directly from the text: the Hamiltonian identity connecting H_RAA and H_AA is a canonical rotation, so the transition at J=±K is inherited rigorously from Jitomirskaya's theorem. The numerical support (Figs. 1–2) is genuine and reproducible in structure, and the paper is honest in flagging the patchwise/singular limitations of the metric construction. However, the strongest advertised claim—coincidence with a massless Dirac fermion in curved spacetime—is only partially supported: the equivalence is patchwise, the reconstructed metric is singular on an infinite set, and the sign-gauging issue for the hopping is not addressed. The 'polynomial scaling to zero' of the position NPR is asserted only from a guide-to-the-eye fit at N=F_{21}, the largest size shown.","major_comments":[{"comment":"The curved-spacetime equivalence is established only on patches where sin(ωn/2+φ/2) and cos(ωn/2+φ/2) have fixed sign, and the text explicitly concedes that α_n diverges on an infinite set of sites when ω/2π is irrational. A lattice Dirac Hamiltonian regularizing a smooth curved-spacetime Dirac operator requires a globally defined, nonsingular α(x) (or a well-defined gauge with bounded α_n). As it stands, 'coincides with the Hamiltonian of a massless Dirac fermion in curved spacetime' overstates the result: the paper provides a piecewise, singular metric that reproduces the lattice Hamiltonian only locally. The authors should either soften the claim to 'locally equivalent, up to singularities, to a lattice Dirac Hamiltonian with a piecewise-defined metric,' or prove that a global smooth metric can be constructed, e.g., by removing the gauge/recurrence singularities.","section":"Section IV, Eqs. (12)–(13)"},{"comment":"At J=K the hopping amplitude in H_RAA is 4J cos((ωn+φ)/2), while the Dirac Hamiltonian H_II uses t_n ∝ |cos((ωn+φ)/2)|. The equivalence therefore requires removing sign changes of cos((ωn+φ)/2) by a gauge transformation ψ_n → i^n ψ_n or a U(1) phase. In one dimension a sign-changing hopping can be gauged to a positive hopping only if the total Z_2 flux ∏_n sgn(cos((ωn+φ)/2)) equals +1. For a quasiperiodic sign sequence on a finite ring this product is not guaranteed to be +1, and in the thermodynamic limit the sign sequence is not periodic. The paper does not discuss this gauge obstruction. Without it, H_RAA|J=K is not literally equal to H_II, only equal up to a local gauge that may fail globally. This is a load-bearing issue for the Dirac-coincidence claim.","section":"Section IV, H_I vs H_II and Eqs. (18)–(19)"},{"comment":"The claim that the position-basis NPR at eJ=0 'scales to zero polynomially' is supported only by Fig. 2(c), which shows a guide-to-the-eye straight line in log-log for N = F_{n+1} with n up to 20 (N=10946), with no error bars, no phase-averaging procedure described quantitatively, and no fitted exponent. Since this is the only quantitative statement connecting the hidden transition to conventional position-basis observables, the paper should either provide the fitted exponent, a scaling collapse, or a derivation of the polynomial law. Without it, the statement is an extrapolation from a single sequence of system sizes.","section":"Section V, Fig. 2 and the polynomial NPR scaling"},{"comment":"The rotated-basis IPR/NPR are computed in the basis |e_m^±>, but the paper does not specify the normalization, the boundary conditions, or the precise finite-N implementation of these states. Since |e_m^±> = Σ_n e^{iω n^2/4 ∓ imn}|n>, these are discrete quadratic-phase states; on a finite chain with periodic boundary conditions, the quadratic phase factor e^{iω n^2/4} is not periodic unless a twisted boundary condition is imposed. The numerical results in Fig. 1 therefore depend on an unspecified regularization. Please clarify how |e_m^±> are defined and normalized for finite N.","section":"Section V, Eq. (17) and rotated-basis IPR/NPR"}],"minor_comments":[{"comment":"The abstract mentions 'many-body localization' while the paper treats a single-particle model; the final sentence of the Introduction also repeats this. The text should be corrected to 'localization/Anderson localization.'","section":"Abstract and Introduction"},{"comment":"The NPR notation is written inconsistently (NPR_i = 1/(N·IPR_i)) and the text later uses ⟨NPR⟩; also 'thermodinamic' is a typo for 'thermodynamic.'","section":"Section II, Eq. (2)"},{"comment":"The notation with ⊗ and (a+ib)⊗(c+id)=ac+ibd is nonstandard and could confuse; please define it earlier or use an explicit complex hopping amplitude t_n.","section":"Section III, Eq. (3)"},{"comment":"The product index and exponent (−1)^{m+n+1} are confusing: the product runs over m=1 to n−1, so m+n+1 has a fixed parity? Please clarify the derivation of these recurrences, and define C_I/C_II consistently (typo 'defiining').","section":"Section IV, Eqs. (12)–(13)"},{"comment":"The caption states the phase transition occurs at eJ=0, but the original Hamiltonian also has a transition at eK=0 (J=−K). The text says 'similar plots are obtained for eK=0' without showing them; consider showing both or explicitly noting the symmetry.","section":"Section V, Fig. 2 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is a single-author preprint with a very strong central algebraic claim (exact canonical equivalence), but the advertised analog-gravity conclusion is the least rigorous part. I would recommend major revision rather than rejection because the canonical-transformation core is sound and the limitations are at least partly acknowledged; however, the patchwise/singular metric and the sign-gauging issue need to be fixed or the claims need to be downgraded. The numerical evidence for the position-basis NPR scaling is thin and needs an exponent or a collapse. I would not accept without these changes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core result here is sound but smaller than the advertising. The \"hidden\" transition is just the Aubry–André transition moved by a canonical transformation; the paper says so itself, and the transition location J=±K is inherited from the Jitomirskaya theorem. What is actually new is the specific rotated Hamiltonian in Eq. (4) — an off-diagonal, complex-hopping, traceless AA variant — plus the numerical observation that the transition leaves a remnant dip in the position-basis NPR that scales to zero. The algebra connecting H_RAA and H_AA is exact and clean, and the numerical signature, though limited, is plausible and might be useful for cold-atom or photonic simulators that want AA physics without onsite disorder.\n\nThe soft spots are in proportion. The largest is the \"massless Dirac fermion in curved spacetime\" claim. The reconstruction in Section IV is explicitly patchwise, and the paper concedes that the metric factors diverge on an infinite set of sites. Moreover, at the transition the hopping in H_RAA is proportional to cos(θ_n) with sign changes, while the Dirac lattice Hamiltonian H_II uses |cos(θ_n)|; no gauge transformation removing all sign defects is established for an irrational quasiperiodic sequence. So \"coincides\" is too strong — at best you have a patchwise equivalence up to gauge, and the metric is constructed from the hopping rather than independently predicted. Calling the model \"many-body localization\" anywhere is also an overstatement; this is a single-particle tight-binding model. The position-basis NPR scaling is numerical for one frequency and one system size sequence, with no error bars and no analytic argument; that part is suggestive, not proven.\n\nWho gets value? Researchers working on generalized AA models, off-diagonal disorder, and experimental platforms for AA physics. The mapping in Eq. (4) is a genuinely useful addition to the toolbox. The Dirac/gravity connection is a framing device that should be heavily qualified or removed until a global metric is exhibited. I would not block the paper on the core mapping, but I would send it back for major revision on the Dirac claim and the many-body language. It deserves a serious referee, not a desk reject — just a referee who insists the claims match the math.","headline":"The hidden transition is a real but unsurprising corollary of Aubry–André duality; the paper's genuine novelty is the off-diagonal complex-hopping model and the position-basis NPR dip, while the curved-spacetime claim is materially overstated.","tokens_in":10655,"tokens_out":1785,"would_cite":false,"duration_ms":22866,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A canonical rotation of the Aubry-André model hides a localization transition, detectable in ordinary position space, and at the critical point the Hamiltonian coincides with a massless Dirac fermion in curved spacetime.","keywords":["hidden localization transition","Aubry-André model","canonical rotation","inverse participation ratio","normalized participation ratio","massless Dirac fermion","curved spacetime","analog gravity"],"falsifier":"At the transition J=-K, compute the finite-size scaling of the inverse participation ratio in the rotated basis using Fibonacci approximations of the golden-ratio frequency: if IPR does not decay as N^{-D2} with a nontrivial exponent 0<D2<1 while NPR goes to zero, the hidden multifractal transition is absent. Alternatively, check whether the low-energy spectrum of the rotated Hamiltonian at J=-K contains the linear Dirac crossing predicted by the curved-spacetime Hamiltonian; the absence of such a crossing would refute the claimed coincidence.","tokens_in":9773,"feed_emoji":"🌀","tokens_out":4592,"duration_ms":55813,"temperature":0.7,"pith_summary":"This paper generalizes the Aubry-André model by replacing position and momentum with an arbitrary canonically conjugate pair, effectively rotating the operators by half a step in phase space. The author claims that this rotated model still undergoes a localization transition, but the localized and delocalized phases are defined with respect to the rotated basis, not the original lattice. The transition is hidden in the sense that it leaves only a remnant signature in the conventional position basis: the average normalized participation ratio dips to zero in the thermodynamic limit. At the transition point, the Hamiltonian becomes a lattice model of a massless Dirac fermion in a curved spacetime metric, suggesting a direct link between Anderson localization and analog gravity. A sympathetic reader would take this as a proof-of-concept that localization transitions can be engineered in off-diagonal, quasiperiodic hopping models and that criticality in such models carries a geometric, spacetime-like signature.","feed_headline":"Rotate the Aubry-André operators and a hidden transition emerges","feed_subtitle":"At the critical point the lattice maps onto a massless Dirac fermion in curved spacetime, linking localization to analog gravity.","key_machinery":"The central object is the canonically rotated pair X=(1/2)(ωx+φ)-p and Y=(1/2)(ωx+φ)+p, with the canonical commutator [X,Y]=iω. In the position basis, this transformation turns the usual Aubry-André cosine terms into an off-diagonal tight-binding chain with quasiperiodic, site-dependent hopping amplitudes; in momentum space the Hamiltonian has the same functional form with the roles of position and momentum exchanged. The self-duality at J=±K, a direct consequence of the original Aubry-André self-duality, is what forces the transition. At the transition, the hoppings factor as products of a metric-derived factor α_n, which the paper connects to the zweibein of a 1+1D curved spacetime; the tw","core_discovery":"The paper shows that the Hamiltonian obtained by replacing the canonical pair (X,p) with the rotated pair X=(1/2)(ωx+φ)-p and Y=(1/2)(ωx+φ)+p is formally identical in position and momentum space, up to the exchange of the roles. This self-duality, inherited from the Aubry-André model, forces a localization transition at J=±K, but the localized/delocalized eigenstates are relative to the rotated operators. At the transition, the eigenstates are neither localized nor extended in either the rotated basis or the original position basis: both the inverse participation ratio and the normalized participation ratio vanish in the thermodynamic limit, indicating multifractal critical states. In the or","pith_inferences":["A natural testable extension is to check whether the position-basis NPR dip at the hidden transition persists for non-Hermitian or interacting generalizations of the rotated model; if it does, the dip could serve as a robust diagnostic for hidden localization in systems that do not have a canonical-rotation dual.","The metric-factors α_n diverge on an infinite set of sites, so the Dirac-equivalence is strictly a piecewise statement. It would be worth examining whether a global, regularized metric captures the same low-energy physics or whether the singularities are essential to the critical behavior.","The triality suggests that the Anderson transition in the original Aubry-André model, the Hofstadter butterfly in the Harper model, and the hidden transition in the rotated model are the same phenomenon viewed in three different bases; this points to a geometric interpretation of the mobility edge as a metric deformation.","The construction may extend to higher-dimensional quasiperiodic systems: replacing position and momentum by a rotated pair in two or more dimensions could produce hidden localization transitions with multifractal signatures in multiple observables."],"forward_implications":["The rotated Aubry-André model can be realized as arrays of quantum dots or cold atoms with quasiperiodic, off-diagonal hoppings, and the hidden transition should be observable as a dip to zero of the average NPR in the position basis.","The equivalence at the transition point provides a tabletop realization of a massless Dirac fermion in a quasiperiodically deformed spacetime metric, offering a concrete platform for analog-gravity experiments in condensed-matter systems.","The duality between the rotated model, the Aubry-André model, and the Harper-Hofstadter model (a triality) transfers spectral and localization results among all three, so critical exponents and multifractal dimensions computed in one model apply to the others.","The self-duality at J=±K implies that the hidden transition persists for almost all irrational frequencies and phases, with the transition boundary exactly at J=±K, inherited from the Aubry-André result.","The multifractal nature of the critical states, indicated by the simultaneous vanishing of IPR and NPR, should be measurable through finite-size scaling in both the rotated and position bases."],"fun_headline_variants":["Rotating the operators reveals a hidden localization transition","Hidden transition in rotated Aubry-André yields multifractal critical states","At the hidden transition, the model becomes a Dirac fermion in curved space","Operator rotation exposes a transition to multifractal states"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The equivalence to a massless Dirac fermion in curved spacetime is built from metric factors that are defined only on patches where sine and cosine have a fixed sign and that diverge on an infinite set of lattice sites, so the existence of a global smooth metric is not established.","fun_headline_variants_meta":{"raw":{"variants":["Rotating the operators reveals a hidden localization transition","Hidden transition in rotated Aubry-André yields multifractal critical states","At the hidden transition, the model becomes a Dirac fermion in curved space","Operator rotation exposes a transition to multifractal states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00064,"raw_usage":{"total_tokens":2832,"prompt_tokens":842,"completion_tokens":1990,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":586,"completion_tokens_details":{"reasoning_tokens":1918}},"tokens_in":586,"tokens_out":1990,"duration_ms":15608,"temperature":1.0,"reasoning_tokens":1918,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T07:10:02.125159+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"At the transition J=-K, compute the finite-size scaling of the inverse participation ratio in the rotated basis using Fibonacci approximations of the golden-ratio frequency: if IPR does not decay as N^{-D2} with a nontrivial exponent 0<D2<1 while NPR goes to zero, the hidden multifractal transition is absent. Alternatively, check whether the low-energy spectrum of the rotated Hamiltonian at J=-K contains the linear Dirac crossing predicted by the curved-spacetime Hamiltonian; the absence of such a crossing would refute the claimed coincidence.","supporting_citations":[],"review_version":1}