{"id":"1ad3d82d-a989-4acb-aad8-fafba7976470","arxiv_id":"2506.03247","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A new tensor RG map with a verified master function rigorously establishes stability of the high-temperature fixed point and yields new high-temperature bounds for the 2D Ising and XY models.","lead":"This paper introduces a new tensor renormalization group map, the 2x1 map, with a computer-assisted proof that it converges to the high-temperature fixed point for a wide class of lattice tensors. It provides rigorous bounds on the high-temperature phase of the 2D Ising and XY models.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The rigorous soundness of the key contraction check depends on unverified Julia code and on interval arithmetic on an infinite-dimensional tail; an independent re-derivation of the master function or a formal check of the code is needed.","rationale":"The reader's weakest_assumption is exactly the soundness of the computer-assisted verification. The central claim—Theorem 3.1 and its consequence Proposition 3.3—depends entirely on the correctness of the master function implementation and the interval arithmetic libraries. Since the paper provides no formal verification, no independent check, and even notes that the symbolic formula (2.99) was not checked by hand, the bug/rounding risk is real and load-bearing. The concern is not that computer assistance is inherently unreliable; it is that the paper's key numerical claim is a verified computation whose verification has not been independently confirmed. The concrete test would resolve this: an independent reimplementation and re-derivation. The verdict should remain CONDITIONAL, not REJECT, because the paper is honest and transparent about the code dependency and provides the code; the concern is about verification, not about the overall approach.","tokens_in":39176,"tokens_out":1247,"duration_ms":12173,"concrete_test":"Independently reimplement the master function M(pb) defined in Sections 2.6-2.9 (e.g. from the paper's diagrams alone, without referencing the authors' code), and verify that the iterates from pb^(0)_abcd = 0.02 satisfy pb^(15) <= 0.96 pb^(14) with verified interval arithmetic. Also independently re-derive the symbolic formula (2.99) for the leading part of M; if the formula or the contraction check does not reproduce the paper's claim, the central theorem fails.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 3.1 rests on the computer-assisted check that pb^(15) <= 0.96 pb^(14) for the master function M, obtained by a Julia implementation that is not independently verified and uses interval arithmetic libraries (ArbNumerics.jl, HCubature.jl) whose soundness is not established in the paper. The master function is defined through a long chain of contractions over infinite-dimensional sectors, with a non-analytic square-root combination (2.78), and the paper explicitly states that checking formula (2.99) by pencil and paper would be straightforward but tedious and was not done (Section 2.10). If the code has a bug, a wrong contraction, an unsound rounding mode, or an erroneous tail bound, the key inequality may fail and Theorem 3.1 as stated (with w_x=2.2, w_o=2, delta=0.02) could be false. A concrete re-derivation of the linearized master function, or an independent reimplementation of the full master function and its key basin check, would settle the concern.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a new anisotropic tensor renormalization group map, the 2x1 map, which coarse-grains a square lattice by a factor of two in one direction and then rotates by 90 degrees. The map is defined on four-leg tensors over infinite-dimensional Hilbert spaces, and the authors develop a graphical calculus that translates the RG steps into inequalities on tensor components. A finite-dimensional \"hat-tensor\" bounding box is introduced, together with a master function M that controls how this box evolves under RG. The main theorem (Theorem 3.1) states that, for reweighting parameters w_x=2.2 and w_o=2, every tensor whose 63 non-trivial sectors have Hilbert-Schmidt norm at most 0.02 flows to the high-temperature fixed point; consequently the free energy is analytic in the interior of this neighborhood (Proposition 3.3). The method is then applied to obtain explicit high-temperature bounds for the 2D Ising model (β≤0.12) and the XY model (β≤0.18995). The proofs are computer-assisted, using interval arithmetic implemented in a provided Julia notebook.","tokens_in":39386,"tokens_out":6743,"duration_ms":74695,"significance":"If the computer-assisted checks are sound, this is a valuable contribution to the rigorous tensor RG program. The 2x1 map is a genuinely new and relatively simple RG map, and the hat-tensor/master-function framework provides a concrete finite-dimensional control of an infinite-dimensional RG flow, with explicit basin-of-attraction sizes and model-specific bounds. The paper ships documented Julia code and uses interval arithmetic, which are strengths, and the bounds for the Ising and XY models are concrete, falsifiable predictions. The main limitation is that the central computer-assisted verification is not independently reproduced, and some supporting lemmas are stated without full proof.","major_comments":[{"comment":"The proof of Theorem 3.1 rests on the computer-assisted check that pb^(15) ≤ 0.96 pb^(14) for the master function; this check is performed only by the provided Julia code, and the paper explicitly states in §2.10 that the symbolic linearization (2.99) was not checked by hand. The correctness of the theorem therefore depends on the absence of implementation bugs and on the soundness of the interval arithmetic libraries (ArbNumerics.jl, HCubature.jl), neither of which is independently verified in the manuscript. Because this is the load-bearing step, I ask for an independent verification: for example, a pencil-and-paper derivation of at least the linearized master function (2.99), or a second, independent reimplementation of the full master function and of the check (3.2) in a different language or with a different interval arithmetic library.","section":"§3.1 and §2.10"},{"comment":"The master function M is proved monotonic and subhomogeneous by composing Propositions 2.5, 2.6, and 2.8. Proposition 2.6 is only sketched (the key quadrature estimate (2.78) is asserted to preserve subhomogeneity without a detailed proof), and Proposition 2.8 is stated without proof. These properties are essential for the Key Lemma 2.3, which converts the finite check (3.2) into infinite-time convergence. The proofs should be written out in full, at least for the non-trivial quadrature estimate (2.78) and for the normalization divisions by |N_1| and |N_2|.","section":"§2.7.6, Proposition 2.6, Proposition 2.8"},{"comment":"Lemma B.3, the counting identity |T_n| = 2n^2 + 3n + 1, is stated without proof and is used in Lemma B.4 to bound the tail of the XY tensor norm. This is an elementary counting argument, but since the XY result depends on it, a short proof should be included instead of being omitted.","section":"Appendix B, Lemma B.3"}],"minor_comments":[{"comment":"In Eq. (2.79), the notation pR should be pRκ; as written, the equation appears to define a hat-tensor for R rather than for Rκ.","section":"§2.7.6, Eq. (2.79)"},{"comment":"The sentence \"we represented the oo channel as a contraction of L_o and R_o\" uses L_o and R_o where L_oo and R_oo are meant; the subscript is missing.","section":"§2.7.4"},{"comment":"Reference [29] lists the title \"TensorSeries.jl,\" but the text in Appendix A refers to TaylorSeries.jl; the bibliography entry should be corrected.","section":"References"},{"comment":"The Kronecker delta δ_{i,i0} in Eq. (3.6) clashes with the radius δ = 0.02 used throughout Section 3; consider renaming one of these quantities to avoid confusion.","section":"§3.1, Eq. (3.6)"}],"recommendation":"major_revision","confidential_remarks":"This is a promising paper, and I did not find a clear mathematical error in the analytical framework. The main risk is the trustworthiness of the computer-assisted step: the paper explicitly acknowledges that the symbolic linearization (2.99) was not checked by hand, and the key contraction check (3.2) is only verified by the provided code. For a rigorous computer-assisted proof, an independent reimplementation or formal verification is typically expected. If the authors can supply such a check, or a detailed derivation of the linearized master function, I would support acceptance. The self-citations to [3,4,6,7] are used appropriately and I see no circularity."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, the 2x1 map and the hat-tensor/master-function machinery are a genuinely new way to turn an infinite-dimensional RG flow into a finite interval-arithmetic computation. That is the paper's real contribution, and it is well executed. Second, the main theorem—stability of the high-T fixed point in a 0.02 box—is not a new physical result; stability was already known from Kennedy-Rychkov and others. But the method is designed to be pushed toward critical fixed points, which makes it worth careful attention.\n\nWhat the paper does well: It is unusually self-aware. The graphical language is clear and the detailed definitions are mostly complete. It states explicitly where proofs are omitted (Lemma 2.4, B.3) and where symbolic linearization (Eq. 2.99) was not checked by hand. The code is public, documented, and short (~300 lines). The interval arithmetic use is standard for computer-assisted proofs, and the authors verify the final convergence condition (pb^(15) ≤ 0.96 pb^(14)) rather than assuming it. The Ising and XY applications are concrete and honestly compared with cluster expansion.\n\nSoft spots, in proportion. The biggest is exactly what the stress-test flags: the key contraction check relies on unverified Julia code and on the soundness of ArbNumerics and HCubature. That is a real but routine caveat for this field; it does not undermine the framework. A referee should ask for an independent reimplementation or a formal proof of the code for that one step. Second, the basin size δ=0.02 is obtained by manual tuning of w_x, w_o; that's fine, but it means the theorem is not robust to parameter choice—a different pair might give different δ. Minor. Third, (2.99) is central to the linearized contraction argument, and the paper says pencil-and-paper verification would be tedious and wasn't done. That's a fixable gap; it should be closed or independently checked.\n\nThe reader's take is about right: conditional, moderate confidence, soundness 7. I'd say the math structure is solid; the computational component is standard rigorous-computation practice. No circularity—the convergence condition is verified, not fit.\n\nWho this is for: researchers in rigorous RG, tensor networks, and computer-assisted proofs. The paper deserves a serious referee, and I would engage with it. My recommendation: send to peer review, with a referee asked to verify the code and (2.99). Would cite in my own work if I worked in this area.","headline":"A genuinely new computational framework for rigorous tensor RG; the high-T result is not new physics, but the method is the contribution and it deserves peer review.","tokens_in":39905,"tokens_out":2547,"would_cite":true,"duration_ms":31132,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B20","82B28","65G20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves, with a computer-assisted interval-arithmetic argument, that every tensor within 0.02 of the high-temperature fixed point in all 63 sectors flows to that fixed point under a new 2x1 tensor renormalization group map.","keywords":["tensor renormalization group","2x1 RG map","hat-tensor","master function","computer-assisted proof","interval arithmetic","high-temperature fixed point","lattice spin models"],"falsifier":"An independent re-implementation of the master function that fails to reproduce the componentwise bound $\\hat b^{(15)}\\le 0.96\\,\\hat b^{(14)}$ for $\\delta=0.02$, $w_x=2.2$, $w_o=2$ would collapse Theorem 3.1. So would a single tensor in $O_{0.02}$ whose normalized 2x1 iterates fail to converge to zero.","tokens_in":38977,"feed_emoji":"🧩","tokens_out":8314,"duration_ms":88653,"temperature":0.7,"pith_summary":"The paper establishes a rigorous, computer-assisted basin of attraction for the high-temperature fixed point of a new tensor renormalization group map. It introduces the 2x1 map, which coarse-grains a lattice by a factor of two in one direction and then rotates by 90 degrees, together with a graphical calculus that turns each step of the map into inequalities on tensor components. The central result is that any tensor whose 63 non-trivial sector norms are bounded by 0.02 flows to the fixed point under the normalized 2x1 map, and the free energy is analytic in that neighborhood. Applied to the 2D Ising and XY models, the method proves high-temperature phases for inverse temperatures $eta\\le 0.12$ and $\\beta\\le 0.18995$, respectively.","feed_headline":"All tensors within 0.02 flow to the high-T fixed point","feed_subtitle":"A 2x1 tensor RG map plus interval arithmetic turns renormalization into a rigorous phase proof.","key_machinery":"The central mechanism is the hat-tensor: a finite-dimensional tensor of nonnegative real numbers that bounds the Hilbert-Schmidt norm of each sector of the infinite-dimensional full tensor. The master function $M$ maps hat-tensors to hat-tensors by exactly mimicking the four steps of the 2x1 map, namely gauge transformation, disentangling and splitting, reconnection, and rotation. Because $M$ is monotonic and subhomogeneous, one rigorous check that an iterate decreases by a factor $\\lambda<1$ implies all future iterates decrease geometrically, via the Key Lemma 2.3, reducing infinite-dimensional RG control to a finite numerical computation.","core_discovery":"The central claim is Theorem 3.1: for reweighting parameters $w_x=2.2$ and $w_o=2$, the set $O_\\delta$ with $\\delta=0.02$ is a basin of stability, meaning the normalized 2x1 RG map can be iterated forever from any starting tensor in the set and every iterate converges to zero in the Hilbert-Schmidt norm. The proof is computer-assisted via a master function on finite-dimensional hat-tensors; iterating that function in interval arithmetic verifies the componentwise contraction $\\hat b^{(15)}\\le 0.96\\,\\hat b^{(14)}$, after which monotonicity and subhomogeneity force exponential decay forever. By Proposition 3.3 the free energy is analytic in the interior of the basin. The same framework yields rigorous high-temperature bounds for the Ising model ($\\beta\\le 0.12$, versus $\\beta_c\\approx 0.44$) and the XY model ($\\beta\\le 0.18995$, versus $\\beta_c\\approx 1.12$).","pith_inferences":["An implication the authors leave implicit: replacing the scalar reweighting factors $w_x,w_o$ with sector-resolved matrices, using the same $w\\,w^{-1}=1$ identity, should enlarge the provable basin beyond $\\delta=0.02$.","The structure of the proof suggests a testable extension: running the same master-function iteration on other two-dimensional models, such as Potts or clock models, should yield explicit high-temperature intervals with only a few lines of the provided code.","The authors' own sketch for critical fixed points points to a finite-dimensional head plus hat-controlled tail; the hat-tensor machinery here is exactly the control component such a construction would need.","Because the method bounds all sectors uniformly, it automatically proves stability against symmetry-breaking perturbations such as magnetic fields, a robustness the paper notes but does not exploit numerically."],"forward_implications":["Any tensor within the 0.02 box in all 63 sectors has a well-defined infinite RG trajectory converging to the high-temperature fixed point.","The free energy exists and is analytic on the interior of that basin, so all such tensors lie in the same high-temperature phase.","The Ising model at inverse temperatures $\\beta\\le 0.12$ and the XY model at $\\beta\\le 0.18995$ are rigorously inside their high-temperature phases.","The same master-function scheme can be rerun for other models or other reweighting parameters, producing checkable computer-assisted phase proofs.","The graphical language converts diagrammatic RG maps into componentwise inequalities, so future maps, including those aimed at critical fixed points, can be controlled in the same way."],"supporting_citations":[{"why":"Supplies the prior rigorous high-temperature tensor RG construction and the disentangler strategy that the 2x1 map adapts.","marker":"[3]"},{"why":"Sets tensor-network conventions, partition-function invariance under gauge and disentangler steps, and analyticity propositions reused here.","marker":"[4]"},{"why":"The accompanying computer code that evaluates the master function and performs the interval-arithmetic check behind the main theorem.","marker":"[8]"},{"why":"Classic result used to justify free-energy limits for finite-range lattice models.","marker":"[11]"},{"why":"Cluster-expansion criterion used as the comparison benchmark for the size of the high-temperature region.","marker":"[13]"},{"why":"Gives the two-state tensor representation of the Ising model from which the Ising analysis starts.","marker":"[14]"},{"why":"Provides the Bessel-function Fourier series used to bound the XY model hat-tensor.","marker":"[16]"},{"why":"Interval-arithmetic library used to make the master-function iterates rigorous in Section 3.","marker":"[30]"},{"why":"Adaptive cubature library used in the rigorous numerical bounds.","marker":"[32]"}],"fun_headline_variants":["2x1 map turns tensor RG into a rigorous proof","Tensor RG proven to converge for all 0.02 deviations","Computer-assisted proof: RG convergence basin of size 0.02","Rigorous high-T bounds for Ising and XY via tensor RG","New 2x1 map yields rigorous RG convergence check"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire proof rests on the computer check being sound: the accompanying code must implement the master function correctly, and the interval arithmetic must truly enclose all rounding errors, so the observed contraction of the bounding box after fifteen iterations is guaranteed for every tensor in the box.","fun_headline_variants_meta":{"raw":{"variants":["2x1 map turns tensor RG into a rigorous proof","Tensor RG proven to converge for all 0.02 deviations","Computer-assisted proof: RG convergence basin of size 0.02","Rigorous high-T bounds for Ising and XY via tensor RG","New 2x1 map yields rigorous RG convergence check"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001187,"raw_usage":{"total_tokens":4908,"prompt_tokens":964,"completion_tokens":3944,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":580,"completion_tokens_details":{"reasoning_tokens":3857}},"tokens_in":580,"tokens_out":3944,"duration_ms":30335,"temperature":1.0,"reasoning_tokens":3857,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:07:54.110648+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An independent re-implementation of the master function that fails to reproduce the componentwise bound $\\hat b^{(15)}\\le 0.96\\,\\hat b^{(14)}$ for $\\delta=0.02$, $w_x=2.2$, $w_o=2$ would collapse Theorem 3.1. So would a single tensor in $O_{0.02}$ whose normalized 2x1 iterates fail to converge to zero.","supporting_citations":[{"cited_title":"Ruelle,Statistical Mechanics: Rigorous Results","cited_arxiv_id":null,"evidence_quote":"Classic result used to justify free-energy limits for finite-range lattice models."},{"cited_title":"Cluster expansion for abstract polymer models,","cited_arxiv_id":null,"evidence_quote":"Cluster-expansion criterion used as the comparison benchmark for the size of the high-temperature region."},{"cited_title":"NIST Digital Library of Mathematical Functions","cited_arxiv_id":null,"evidence_quote":"Provides the Bessel-function Fourier series used to bound the XY model hat-tensor."},{"cited_title":"ArbNumerics.jl","cited_arxiv_id":null,"evidence_quote":"Interval-arithmetic library used to make the master-function iterates rigorous in Section 3."},{"cited_title":"HCubature.jl","cited_arxiv_id":null,"evidence_quote":"Adaptive cubature library used in the rigorous numerical bounds."}],"review_version":1}