{"id":"e3f8ac0f-e023-474b-aaf8-b0e45e965cbb","arxiv_id":"2508.12714","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Anderson localization is proven at the spectral edge for alloy-type Anderson-Bernoulli models with exponential long-range hopping on Z^d and R^d.","lead":"This paper proves that random disorder in certain materials can keep quantum states localized even when particles can hop across long distances, specifically near the edge of their energy range. It extends a known mathematical proof to a broader family of alloy-type Anderson models on lattices and in continuous space.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Initial-scale Green's function estimate for alloy-type Bernoulli potential is the least secure step; needs explicit verification.","rationale":"The reader's weakest assumption is precisely the initial-scale Green's function estimate for the alloy-type potential with long-range hopping. I agree that this is the least secure part of the argument. The abstract gives no details of how Klopp's Floquet-Bloch method and quantitative uncertainty principle are adapted, and the alloy-type structure introduces correlations that may invalidate the standard estimates. However, this is a concern about unverified steps, not a demonstrated inconsistency. Without the full text, we cannot determine whether the proof overcomes this obstacle. Therefore the appropriate verdict remains UNVERDICTED, consistent with the reader's low-confidence assessment. The concrete test would resolve the concern by checking a specific non-delta kernel and long-range hopping case, exposing any hidden assumptions about the kernel support or the decay rate.","tokens_in":649,"tokens_out":6564,"duration_ms":87825,"concrete_test":"Independently re-derive the initial-scale estimate for a minimal alloy kernel with support radius R=1, e.g., φ(n)=δ_{n,0}+δ_{n,e_1}, and exponential hopping t(m,n)=e^{-μ|m-n|}, following the paper's adaptation of Klopp's lemma. Verify the claimed Green's function bound for a sparse Bernoulli configuration where variables are 1 on a periodic sublattice and 0 elsewhere. Check that the bound is uniform in μ (for μ above some fixed threshold, not scaling with box size) and in the kernel support radius. If the proof requires φ to be a delta (V(n)=ω_n) or requires the alloy kernel to be of rank one, the alloy-type claim is not substantiated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—Anderson localization near the spectral edge for an alloy-type Anderson-Bernoulli model with exponential long-range hopping—rests on an initial-scale Green's function estimate. The abstract states that Bourgain's method is followed, but that Klopp's Floquet-Bloch/uncertainty-principle technique is adapted for this estimate. This is the natural pressure point: in Bourgain's original Bernoulli model, the on-site random variables directly give a potential that is either 0 or 1 at each lattice site. In an alloy-type model, the potential at a site is a convolution of Bernoulli variables with a single-site kernel, so the effective random potential is correlated, multi-valued, and its 'zero set' has a more complex geometry. Klopp's quantitative uncertainty principle typically requires a potential with a positive lower bound on a set of positive density, which a sparse Bernoulli configuration may not provide. The proof must establish that the initial-scale resolvent bound holds uniformly over all configurations in the support of the large-deviation estimates, with constants depending only on the model parameters (coupling constant, decay rate, kernel support). The abstract does not indicate how this uniformity is achieved or what conditions on the alloy kernel are needed. If the uncertainty-principle step fails for kernels of support radius greater than zero, or if the decay rate of the hopping couples badly with the configuration combinatorics, the theorem as stated would not follow. This is a correctness risk specific to the extension beyond Bourgain's model.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims to prove Anderson localization near the spectral edge for an alloy-type Anderson-Bernoulli model on Z^d with exponential long-range hopping, and an analogous model on R^d. The proof is said to follow Bourgain's multi-scale method, with initial-scale Green's function estimates obtained by adapting Klopp's Floquet-Bloch/quantitative uncertainty principle approach. Only the abstract is available for review; no proofs, hypotheses, or technical statements are provided.","tokens_in":911,"tokens_out":1395,"duration_ms":18114,"significance":"If the theorem is correct, it constitutes a meaningful extension of Bourgain's 2004 localization result from the standard Bernoulli-on-site model to alloy-type potentials with nontrivial single-site kernels and exponential long-range hopping, and to the continuum. The combination of Bourgain's multi-scale analysis with Klopp's uncertainty-principle technique is a plausible and potentially valuable strategy. However, since the full text is unavailable, the significance can only be assessed provisionally; no machine-checked proofs, numerical verification, or derivations can be credentialed from the abstract alone.","major_comments":[{"comment":"The abstract only states that the proof 'adapts' Klopp's method and 'is mainly based on Bourgain's method.' No details of the multi-scale analysis are given. In particular, it is not clear how the long-range hopping terms interact with the large-deviation estimates at each scale, which is a known technical challenge. The proof of the induction step must be present in the full text; from the abstract alone this step is unverifiable.","section":"Abstract"}],"minor_comments":[{"comment":"The term 'alloy-type Anderson-Bernoulli model' is used without a definition. Clarify whether the single-site kernel is compactly supported or has exponential decay, and whether the Bernoulli variables are independent at each site of the underlying lattice.","section":"Abstract"}],"recommendation":"uncertain","confidential_remarks":"This review is based solely on the abstract because the full text was not provided. The claim is plausible and within the scope of math-ph, but I cannot verify the central technical steps. The key risk is the initial-scale estimate for alloy-type Bernoulli potentials; the authors should be asked to supply the full manuscript and, in particular, the precise lemma and proof for the uncertainty-principle step. I recommend an editor-level request for the complete text before substantive review."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"So the thing to know: this is an abstract-only submission, so the verdict is provisional. It claims a real extension of Bourgain's Bernoulli localization to alloy-type potentials with correlated, multi-valued random variables, on Z^d and R^d, with exponential long-range hopping. That is genuinely new. The authors are honest about leaning on Bourgain's multi-scale analysis and on Klopp's quantitative uncertainty principle for the initial-scale estimates. No new free parameters, no circularity.\n\nWhat the paper does well: it names the two ingredients exactly, and the claimed theorem is a natural next step that has presumably resisted prior attempts. The basic strategy—Bourgain for the induction, Klopp for the seed—is plausible.\n\nThe soft spot is precisely the adaptation of Klopp's estimate. Klopp's uncertainty principle typically needs a potential with a positive lower bound on a positive-density set. A Bernoulli alloy with a non-point kernel does not obviously provide that uniformly over the support of the large-deviation events. The abstract gives no conditions on the alloy kernel (support, sign, decay) and no indication of how the constants in the initial-scale estimate stay uniform. If the adaptation only works under extra hypotheses that are not stated in the theorem, the paper has a gap. That is the referee's job to check.\n\nI would not desk-reject this. It is a technical paper for the random Schrödinger community, and the core claim is plausible enough to warrant referee time. The referee should be asked to pin down the initial-scale lemma and to state the kernel conditions clearly. If that holds up, it's a solid contribution.\n\nFor me personally, I'd likely cite it if I work on Bernoulli localization, but that is not my immediate area. Bring it to a reading group if someone is tracking that literature.\n\nRecommendation: send to peer review, with an explicit request to verify the initial-scale estimate.","headline":"Plausible and genuinely new extension of Bourgain's localization to alloy-type Bernoulli potentials; the initial-scale estimate is the pressure point and needs referee scrutiny.","tokens_in":1277,"tokens_out":2965,"would_cite":false,"duration_ms":34223,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B44","47B80"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves Anderson localization near the spectral edge for an alloy-type Anderson-Bernoulli model with exponential long-range hopping, on both $\\mathbb{Z}^d$ and $\\mathbb{R}^d$.","keywords":["Anderson localization","Bernoulli random variables","alloy-type model","long-range hopping","multi-scale analysis","Floquet-Bloch theory","spectral edge","random Schrödinger operators"],"falsifier":"For a one-dimensional alloy-type Bernoulli model with exponential long-range hopping, compute the Lyapunov exponent at a near-edge energy; if it is zero, localization fails there and the theorem is wrong. Alternatively, simulate the finite-volume Green's function at the initial scale for many Bernoulli configurations and look for a configuration without exponential decay; even one would falsify the key estimate.","tokens_in":587,"feed_emoji":"🎲","tokens_out":5503,"duration_ms":60037,"temperature":0.7,"pith_summary":"The paper proves that an electron in a random alloy potential with Bernoulli (on/off) impurities is localized near the spectral edge, even when hopping has exponential long-range tails. This extends a 2004 multi-scale method devised specifically for Bernoulli randomness, whose original version handled shorter-range models. The proof anchors that method with initial-scale Green's function estimates derived from Floquet-Bloch theory and a quantitative uncertainty principle. A careful reader would care because Bernoulli potentials are not Hölder regular, so standard smooth-random-potential techniques do not apply; this closes a gap for a class of discrete and continuous random Schrödinger operators.","feed_headline":"Localization proven near edge for random alloy with long-range hops","feed_subtitle":"A multi-scale proof covers exponential hopping in both lattice and continuum settings.","key_machinery":"The central mechanism is a multi-scale induction for Green's functions: exponential decay of the resolvent is established on a growing sequence of length scales, and this yields eigenfunction localization. The load-bearing new input is the initial-scale estimate, the base case of the induction, which is derived from Floquet-Bloch theory and a quantitative uncertainty principle. The uncertainty principle supplies the control needed for the non-smooth Bernoulli potential; without it the base case would not get off the ground.","core_discovery":"The central claim is a theorem: for the alloy-type Anderson-Bernoulli model on $\\mathbb{Z}^d$ with exponential long-range hopping, Anderson localization holds near the spectral edge — the spectrum there consists of exponentially decaying eigenfunctions. The same result is proved for a continuum analogue on $\\mathbb{R}^d$. The proof is an extension of a 2004 multi-scale analysis for Bernoulli random variables, and the principal new work is the initial-scale Green's function estimate. That estimate is obtained by combining Floquet-Bloch theory with a quantitative uncertainty principle, which together control the finite-volume resolvent. The theorem thereby widens the class of singular random p","pith_inferences":["If the initial-scale estimates remain uniform, the same multi-scale induction might extend to super-exponentially decaying or stretched-exponential hopping; the paper only treats the exponential case.","The quantitative uncertainty principle is the pivotal input; replacing it by a simpler argument could make the method more transparent and possibly applicable to non-alloy potentials.","The theorem supports the picture that, for single-band Bernoulli alloys, any mobility edge would lie in the interior of the spectrum rather than near the edges.","A numerical test of the finite-volume Green's function in $d=1$ at the initial scale could quantify how large the exponential tail may be before the estimates degrade."],"forward_implications":["Localization near the spectral edge now covers alloy-type Bernoulli models with exponentially decaying, arbitrarily long-range hopping, not only finite-range or compactly supported potentials.","The continuum analogue on $\\mathbb{R}^d$ inherits the same edge-localization statement, so the result applies to random Schrödinger operators with alloy-type Bernoulli impurities and long-range interactions.","The Floquet-Bloch route to the initial-scale estimate may be reusable for other singular single-site distributions.","The method confirms that near-edge states remain exponentially localized even when hopping connects distant sites, provided the decay is exponential."],"supporting_citations":[],"fun_headline_variants":["Anderson localization holds for exponential long-range hopping","Near-edge localization proven for alloy-type Bernoulli model","New proof extends Bourgain's localization to long-range hops","Localization near edge covers both lattice and continuum alloys","Multi-scale analysis tackles long-range random alloy model"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The whole proof depends on the initial-scale Green's function estimates remaining uniform for the Bernoulli alloy with exponential long-range hopping; if those finite-volume bounds fail for some regime, the multi-scale induction has no starting point.","fun_headline_variants_meta":{"raw":{"variants":["Anderson localization holds for exponential long-range hopping","Near-edge localization proven for alloy-type Bernoulli model","New proof extends Bourgain's localization to long-range hops","Localization near edge covers both lattice and continuum alloys","Multi-scale analysis tackles long-range random alloy model"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000226,"raw_usage":{"total_tokens":1260,"prompt_tokens":652,"completion_tokens":608,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":396,"completion_tokens_details":{"reasoning_tokens":535}},"tokens_in":396,"tokens_out":608,"duration_ms":7598,"temperature":1.0,"reasoning_tokens":535,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T19:17:19.631622+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a one-dimensional alloy-type Bernoulli model with exponential long-range hopping, compute the Lyapunov exponent at a near-edge energy; if it is zero, localization fails there and the theorem is wrong. Alternatively, simulate the finite-volume Green's function at the initial scale for many Bernoulli configurations and look for a configuration without exponential decay; even one would falsify the key estimate.","supporting_citations":[],"review_version":1}