{"id":"52245f50-8c07-4161-8295-5d4bdfbadef2","arxiv_id":"1908.02801","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"If the variety of biangular Gabor frames is path-connected and a Gabor mutually unbiased basis exists, then the intermediate value theorem yields a SIC; the authors prove this mechanism works in dimension 2.","lead":"The paper introduces biangular Gabor frames, a relaxation of equiangular frames, and shows a new path toward proving Zauner's conjecture about symmetric informationally complete measurements. It proves a conditional theorem using the intermediate value theorem and gives a new non-constructive proof that a SIC exists in dimension two.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3's hypothesis 'B_d is path-connected' is ill-posed: if B_d includes 0 (as an algebraic variety) it is automatically path-connected, and the proof's WLOG unit-norm step is unjustified; the intended projective/normalized object must be stated.","rationale":"The reader's verdict ACCEPT has the sole stated weakness being the unproved path-connectivity of B_d. The stress-test agrees that path-connectivity is load-bearing, but the issue is sharper than 'unproved': as stated, the hypothesis is ill-posed. Because B_d is closed under nonzero scalar multiplication and the defining polynomial conditions also admit 0, the affine variety is path-connected automatically via radial contraction to 0. Hence Lemma 3 would have a vacuous hypothesis and would imply known-open SIC existence for every prime d with a Gabor MUB, which cannot be right; the invalid link is the proof's 'WLOG unit norm' normalization. This does not destroy the paper's contribution: the IVT idea, Lemma 4, and the d = 2 computation are sound, and the fix is straightforward—replace B_d by C_d or B_d/C^×, state that all paths are confined to nonzero vectors, and adjust Problem 6 accordingly. Because the central lemma is stated incorrectly, acceptance should be conditional on this revision. The reader's weakest_assumption identifies the same general area but not the precise vacuity/normalization flaw.","tokens_in":6457,"tokens_out":15022,"duration_ms":185867,"concrete_test":"Check the paper's own definition of B_d at v = 0: the biangular equalities hold with α = β = 0. If so, form the radial path v(t) = (1-t)v0 for 0 ≤ t ≤ 1 from a Gabor MUB to 0; it lies in B_d, so B_d is path-connected automatically. Attempt to apply Lemma 3's 'WLOG unit norm' step to this path: at t = 1 it cannot be normalized, showing the stated proof is incomplete. Then rerun Corollary 5 with B_d replaced by the projective slice C_d throughout; if the SIC still follows without invoking the vacuous B_d path-connectivity, the paper needs only a revision of Lemma 3 and Problem 6 rather than a change of conclusion.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In Lemma 3, the proof selects a path in B_d from a Gabor MUB v0 to the trivial biangular vector v1 = 1/sqrt(d) 1 and then says 'Without loss of generality, it holds that ||v(t)||_2 = 1.' This reduction requires the path to avoid v = 0, but path-connectivity of B_d as defined does not supply that. B_d is the real algebraic variety of v in C^d for which G(v) is biangular; the biangular conditions are homogeneous and are satisfied by v = 0 (with α = β = 0), so 0 lies in B_d. Since G(cv) is biangular for every c in C^× and the radial segment (1-t)v stays in B_d, every nonzero point is connected to 0. Thus, under the literal definition, B_d is path-connected in every dimension, making the hypothesis of Lemma 3 vacuous and Problem 6 trivial. The proof's WLOG normalization cannot be applied to a path that passes through 0, and no argument is given that a normalized path exists. The intended hypothesis must be path-connectivity of the projective or normalized variety (e.g., C_d or B_d/C^×), but the paper states Lemma 3 and Problem 6 in terms of B_d. This is a real correctness gap in the central lemma, not merely an unproved open premise.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a non-constructive route to Zauner's conjecture. It introduces biangular Gabor frames, proves structural facts about them, and shows that if a Gabor mutually unbiased basis (MUB) exists and the variety of biangular Gabor frames is path-connected, then an intermediate-value argument yields a SIC. The paper proves the existence of a SIC in dimension 2 as a proof of concept, gives a lemma connecting path-connectivity of a normalized slice to path-connectivity of the full variety, and presents numerical experiments for d=2,4,5 that are explicitly described as illustrations. It isolates path-connectivity of the variety as the main open problem.","tokens_in":6774,"tokens_out":7959,"duration_ms":84896,"significance":"If the proposed program can be completed, it would provide a genuinely new, non-constructive approach to Zauner's conjecture and would complement the current Stark-units-based constructive program. The paper is self-contained, proves its elementary lemmas cleanly, and is appropriately careful to label the numerical experiments as heuristic rather than rigorous evidence. The main obstacle is correctly formulating the path-connectivity hypothesis: as written, the central lemma suffers from a correctness gap, though the gap appears repairable by restating the hypothesis in terms of the normalized slice or the punctured variety. The paper is a reasonable research contribution once that issue is resolved.","major_comments":[{"comment":"Under the literal definition of B_d, the hypothesis of Lemma 3 is vacuous and the 'Without loss of generality, it holds that ||v(t)||_2 = 1' step is unjustified. The zero vector satisfies the biangular conditions with alpha=beta=0, and every nonzero v in B_d is connected to 0 by the segment t |-> (1-t)v, so B_d is path-connected in every dimension. A path from the Gabor MUB v0 to v1 may therefore pass through 0, in which case the normalized curve v(t)/||v(t)|| is not defined and the formula Delta(t) = (1-(d+1)alpha(t))/d, which relies on unit norm, does not follow. The intended hypothesis must be path-connectivity of the normalized slice C_d, or of B_d \\ {0}, or at minimum the existence of a path between v0 and v1 that avoids 0. With such a corrected hypothesis the intermediate-value argument would go through, since scaling a path that avoids 0 by 1/||v(t)|| preserves biangularity.","section":"Section 2, Problem 6"},{"comment":"Problem 6 as stated is trivial: because B_d contains 0 and is a cone, B_d is path-connected for every dimension d. The intended open problem must be about the normalized or projective variety, e.g., whether C_d is path-connected. This is not merely a wording issue, because Lemma 3 requires a path that stays away from 0; Lemma 4 proves that C_d path-connected implies B_d path-connected, but the converse direction is not what the proof of Lemma 3 needs. The authors should restate Problem 6 and Lemma 3 in terms of the normalized slice C_d (or B_d^\\times) and make the avoidance of 0 an explicit part of the argument.","section":"Section 2, Problem 6"}],"minor_comments":[{"comment":"The first paragraph contains the typo 'Zuaner's conjecture'; this should be 'Zauner's conjecture'.","section":"Section 3"},{"comment":"The sentence 'For each d in {2,4,5}, we consider the numerical fiducial reported by Scott and Grassl (when d=3, the variety of SIC fiducials is already interesting)' is confusing, since the parenthetical does not explain why d=3 is excluded from the experiment; please clarify.","section":"Section 2"},{"comment":"The phrase 'B_d/C^\\times is at times one-dimensional' is informal; since C^\\times acts on B_d, a more precise formulation would specify whether the dimension refers to the quotient variety or to the slice C_d.","section":"Section 2"}],"recommendation":"major_revision","confidential_remarks":"The central issue is a correctable flaw in the statement of Lemma 3 and Problem 6 rather than an irreparable defect. The numerical work is properly labeled as illustrative, so I see no concern about overclaiming there. The paper's contribution is modest but potentially useful for future approaches to Zauner's conjecture; after the lemma and problem are restated with the normalized slice, the paper could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know up front. The core idea is genuinely new and attractive: relax SICs to biangular Gabor frames, observe that the variety is often one-dimensional, and use the intermediate value theorem to promote a Gabor MUB to a SIC. That is a fresh angle on Zauner, and the paper is honest about what is proven versus conjectural. Second, the reader's report missed a real flaw in the central lemma, and the stress-test is right. As defined, B_d is the real algebraic variety of all v for which G(v) is biangular. The zero vector satisfies the biangular conditions with α=β=0, so 0∈B_d. And because scaling a biangular vector by any nonzero scalar preserves biangularity, every nonzero point connects radially to 0. So B_d is path-connected in every dimension, and Problem 6 is trivially answered. The proof of Lemma 3 then hits a second problem: it says 'Without loss of generality, ||v(t)||=1', but a path in B_d may pass through 0, where normalization is impossible. The IVT could return t* with v(t*)=0, which is not a SIC. The paper's actual intended object is the normalized slice C_d, or B_d\\{0}, and Lemma 4 already does the right bridge: if C_d is path-connected, then any two nonzero points in B_d are connected by a path avoiding 0. With that correction, the IVT argument works, and the d=2 proof is fine. The numerical experiments are clearly labeled as illustrations, not proofs, and the paper is careful elsewhere, so this is a localized but load-bearing error in the write-up. The fix is simple and the underlying idea survives, but the version you showed me should not be accepted as-is; it needs a revision that restates Lemma 3 and Problem 6 in terms of C_d or B_d\\{0}. I would still send it to a serious referee, because the research direction is promising and the core argument is repairable. Worth a reading-group slot, and I'd also flag it in discussion of the stress-test note.","headline":"The biangular relaxation is a fresh idea, but Lemma 3's path-connectivity hypothesis is ill-posed as written: B_d contains 0 and is trivially path-connected, so the central lemma needs a corrected statement.","tokens_in":7271,"tokens_out":4877,"would_cite":false,"duration_ms":52637,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42C15","52C17"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proposes a non-constructive route to Zauner's conjecture: if the variety of biangular Gabor frames is path-connected, the intermediate value theorem produces a SIC.","keywords":["biangular Gabor frames","Zauner's conjecture","SIC-POVM","mutually unbiased bases","intermediate value theorem","path-connected variety","equiangular tight frames","non-constructive proof"],"falsifier":"In any dimension where the path-connectivity claim fails, for example a $d$ in which $C_d$ has two separate components one containing the trivial all-ones frame and the other containing every biangular frame with $\\alpha<1/(d+1)$, Lemma 3 would not apply; a concrete check is to compute the real irreducible components of $C_d$ for, say, $d=7$ and see whether one component meets both endpoint sets.","tokens_in":6264,"feed_emoji":"📐","tokens_out":9718,"duration_ms":92153,"temperature":0.7,"pith_summary":"This paper proposes a non-constructive route to proof of Zauner's conjecture, which says that for every dimension $d$ there are $d^2$ equiangular lines in $\\mathbb{C}^d$ (equivalently, a symmetric informationally complete measurement). The idea is to enlarge the search space to biangular Gabor frames: Gabor frames whose squared inner products take only two values, $\\alpha$ for nonzero translations and $\\beta$ for nonzero modulations. The paper proves that if a Gabor mutually-unbiased basis exists and the real variety of biangular Gabor frames is path-connected, then by the intermediate value theorem some frame on any connecting path has $\\alpha=\\beta$ and is therefore a SIC. It carries out this non-constructive argument in dimension $d=2$ and gives numerical experiments in dimensions $d=4$ and $d=5$ suggesting the needed path-connectivity. The remaining open problem is to prove path-connectivity of this variety in every dimension.","feed_headline":"Biangular frames could prove Zauner's conjecture","feed_subtitle":"Mutually unbiased bases connect to SICs through a path-connected variety; the intermediate value theorem does the rest.","key_machinery":"The central object is $B_d$, the real algebraic variety of all $v\\in\\mathbb{C}^d$ whose Gabor frame $G(v)=\\{M^\\ell T^k v\\}$ is biangular: squared inner products equal $\\alpha$ among nonzero translations and $\\beta$ among nonzero modulation-translations. The relation $\\alpha+d\\beta=\\|v\\|_2^4$ (Lemma 1) makes the two angle parameters dependent. The argument runs on the function $\\Delta(t)=\\beta(t)-\\alpha(t)$ along a path: the intermediate value theorem forces a point with $\\alpha=\\beta$, which by definition is an equiangular Gabor frame, i.e. a SIC. Gabor MUBs obtained from cubic-phase sequence constructions provide the starting point with $\\alpha=0$, and the all-ones vector provides the endpoint with $\\alpha=1$.","core_discovery":"The central claim is conditional but explicit: Lemma 3 states that if there exists a Gabor MUB in $\\mathbb{C}^d$ and the variety $B_d$ of biangular Gabor frames is path-connected, then there exists a SIC in $\\mathbb{C}^d$. The proof tracks the continuous gap $\\Delta(t)=\\beta(t)-\\alpha(t)$ along a path in $B_d$. At a Gabor MUB one has $\\alpha=0$ and $\\beta=1/d$, so $\\Delta=1/d$; at the trivial all-ones Gabor frame one has $\\alpha=1$ and $\\beta=0$, so $\\Delta=-1$. Along any path of biangular frames the sign must change, and a zero of $\\Delta$ is exactly a frame with equal angles, i.e. an equiangular tight frame, hence a SIC. Because biangular frames form a real algebraic variety, with the relation $\\alpha+d\\beta=\\|v\\|_2^4$, the authors recast Zauner's conjecture as a question of whether this variety connects the MUB point to the trivial point. They prove the $d=2$ case and pose the path-connectivity of $B_d$ as the key open problem.","pith_inferences":["The paper leaves implicit that path-connectivity, if established, would make SIC existence a topological consequence rather than an arithmetic miracle; one could then predict SICs in dimensions where algebraic constructions remain unknown, including non-prime-power dimensions.","A testable extension is to compute the real irreducible components of $C_d$ for dimensions beyond 5 and check whether every component contains both the trivial all-ones frame and a point with $\\alpha<1/(d+1)$; the numerical experiments suggest this may hold generically.","The same intermediate-value mechanism might apply to other families of frames in which two angle parameters swap order along a connected variety, potentially widening the route beyond Gabor frames."],"forward_implications":["Solving the paper's Problem 6 (path-connectivity of $C_d$ in every dimension) would yield a non-constructive proof of the SIC existence conjecture, because Gabor MUBs with $\\alpha=0$ are known in infinitely many prime dimensions.","The method works with a weaker endpoint hypothesis: any biangular frame with $\\alpha<1/(d+1)$ that is path-connected to one with $\\alpha>1/(d+1)$ produces a SIC, so explicit SICs are not needed at either endpoint (Problem 7).","A non-constructive proof would bypass the apparent need for explicit algebraic descriptions of fiducial coordinates, which currently seem to require deep number-theoretic conjectures.","The paper's numerical path-tracing procedure generates a one-parameter family of biangular frames passing through a SIC; the same scheme could be developed as a search heuristic for approximate SICs in larger dimensions."],"supporting_citations":[{"why":"States the conjecture that an equiangular tight frame of $d^2$ vectors exists in every dimension; this is the claim the paper aims to support.","marker":"[7]"},{"why":"Supplies the cubic-phase sequence construction whose Fourier transform generates Gabor MUBs in prime dimensions, providing the $\\alpha=0$ starting point for Lemma 3.","marker":"[20]"},{"why":"Gives the weighted 2-design lower bound whose equality characterizes fiducial vectors, anchoring the definition of SIC used throughout the paper.","marker":"[8]"},{"why":"Reports numerical SIC fiducials used as seeds and as the $d=2$ control in the path-connectivity experiments.","marker":"[11]"},{"why":"Demonstrates path-connectivity for the variety of unit norm tight frames using eigensteps, the closest precedent for Problem 6.","marker":"[21]"},{"why":"Provides a non-constructive symplectic geometry proof of connectivity for frame spaces, a technique the paper suggests could solve Problem 6.","marker":"[23]"}],"fun_headline_variants":["Path-connected biangular frames could prove Zauner's","Biangular Gabor frames offer new route to Zauner's","Intermediate value theorem bridges biangular frames and Zauner","Conditional proof of Zauner's conjecture via biangular frames","Biangular frames may imply Zauner's SIC existence"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the real variety of biangular Gabor frames ($B_d$, or the normalized slice $C_d$) is path-connected in every dimension; the paper proves this only for $d=2$ and offers numerical evidence for $d=4$ and $d=5$.","fun_headline_variants_meta":{"raw":{"variants":["Path-connected biangular frames could prove Zauner's","Biangular Gabor frames offer new route to Zauner's","Intermediate value theorem bridges biangular frames and Zauner","Conditional proof of Zauner's conjecture via biangular frames","Biangular frames may imply Zauner's SIC existence"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000884,"raw_usage":{"total_tokens":3779,"prompt_tokens":869,"completion_tokens":2910,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":485,"completion_tokens_details":{"reasoning_tokens":2829}},"tokens_in":485,"tokens_out":2910,"duration_ms":21432,"temperature":1.0,"reasoning_tokens":2829,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:33:42.067309+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In any dimension where the path-connectivity claim fails, for example a $d$ in which $C_d$ has two separate components one containing the trivial all-ones frame and the other containing every biangular frame with $\\alpha<1/(d+1)$, Lemma 3 would not apply; a concrete check is to compute the real irreducible components of $C_d$ for, say, $d=7$ and see whether one component meets both endpoint sets.","supporting_citations":[{"cited_title":"Grundz uge einer nichtkommutativen designtheorie,","cited_arxiv_id":null,"evidence_quote":"States the conjecture that an equiangular tight frame of $d^2$ vectors exists in every dimension; this is the claim the paper aims to support."},{"cited_title":"Complex sequences with low periodic correlations (corresp.),","cited_arxiv_id":null,"evidence_quote":"Supplies the cubic-phase sequence construction whose Fourier transform generates Gabor MUBs in prime dimensions, providing the $\\alpha=0$ starting point for Lemma 3."},{"cited_title":"Weighted complex projective 2-designs from bases: Optimal state determination by orthogonal measurements,","cited_arxiv_id":null,"evidence_quote":"Gives the weighted 2-design lower bound whose equality characterizes fiducial vectors, anchoring the definition of SIC used throughout the paper."},{"cited_title":"Symmetric informationally complete positive-operator-valued measures: A new computer study,","cited_arxiv_id":null,"evidence_quote":"Reports numerical SIC fiducials used as seeds and as the $d=2$ control in the path-connectivity experiments."},{"cited_title":"Connectivity and irreducibility of algebraic varieties of finite unit norm tight frames,","cited_arxiv_id":null,"evidence_quote":"Demonstrates path-connectivity for the variety of unit norm tight frames using eigensteps, the closest precedent for Problem 6."},{"cited_title":"Symplectic Geometry and Connectivity of Spaces of Frames","cited_arxiv_id":"1804.05899","evidence_quote":"Provides a non-constructive symplectic geometry proof of connectivity for frame spaces, a technique the paper suggests could solve Problem 6."}],"review_version":1}