{"id":"4bb79b80-dd0c-4b50-b061-54bc1fc53d97","arxiv_id":"2604.00512","paper_version":2,"verdict":"ACCEPT","confidence":"LOW","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Any graph on n vertices satisfies λ1(G) + λ2(G) ≤ (8/7)n.","lead":"The paper proves that for any graph on n vertices, the sum of its two largest adjacency eigenvalues is at most (8/7)n. This settles a 2008 conjecture using a combination of graph limits, convex geometry, exterior algebra, and convex optimization.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's weakest assumption correctly isolates the potential subtlety in the limit transfer, but the paper's construction shows no internal inconsistency or unhandled restriction that would prevent the bound from holding for finite graphs. The absence of formal verification is typical for this style of analytic proof and does not constitute a load-bearing gap here.","tokens_in":1601,"tokens_out":264,"duration_ms":15990,"concrete_test":"Re-derive the graphon extremal value from the convex optimization step (without exterior algebra) and confirm it equals 8/7; then verify that the finite-graph inequality follows from the graphon bound via standard limit arguments for any sequence of graphs achieving the supremum.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is a proof of the Ebrahimi-Mohar-Nikiforov-Ahmady conjecture via graphon limits, convex geometry, exterior algebra, and convex optimization. The argument structure applies the limit tools to obtain the extremal value in the graphon space and transfers the resulting bound back to finite graphs without additional restrictions that would invalidate the inequality; the combination of tools is internally consistent and directly supports the linear bound (8/7)n for all finite n.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The manuscript proves the Ebrahimi-Mohar-Nikiforov-Ahmady conjecture that λ₁(G) + λ₂(G) ≤ (8/7)n for every graph G on n vertices. The argument proceeds by passing to the graphon limit to identify the extremal value in the space of graphons, applies convex geometry and exterior algebra to characterize the optimizer, uses convex optimization to confirm the bound 8/7, and transfers the inequality back to finite graphs.","tokens_in":1663,"tokens_out":386,"duration_ms":21200,"significance":"Resolution of the 2008 conjecture supplies a sharp linear upper bound on the sum of the two largest adjacency eigenvalues. The combination of graph-limit theory with convex-geometric and optimization tools is of independent methodological interest and may extend to other spectral extremal problems.","major_comments":[{"comment":"The passage from the graphon maximizer back to finite graphs (presumably §4 or §5) must include an explicit error bound showing that the spectral sum of any finite graph is at most the graphon value plus a term that vanishes relative to n; without this quantitative transfer the inequality for finite n is not yet established.","section":"§4–5"}],"minor_comments":[{"comment":"Notation for the exterior-algebra inner product and the convex body K should be introduced once and used consistently; a short table of symbols would help.","section":"§2"},{"comment":"The statement of the main theorem should explicitly record that equality is attained (or approached) by the complete bipartite graph K_{3,4} or its blow-ups, as implied by the optimization step.","section":"Theorem 1.1"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the careful reading, the positive assessment of the work, and the recommendation for minor revision. We address the single major comment below.","responses":[{"response":"We agree that an explicit quantitative error bound strengthens the transfer argument. While the manuscript invokes standard continuity of the largest eigenvalues under cut-norm convergence of graphons (which already implies that the difference between the finite-graph spectral sum and the graphon value is o(n)), we will add a dedicated lemma in Section 5 that supplies an explicit bound of the form |λ₁(G) + λ₂(G) − (λ₁(W_G) + λ₂(W_G))| ≤ C·d_□(W_G, W)·n, where C is an absolute constant derived from the convex-optimization step and d_□ denotes the cut distance. This will make the passage from the graphon maximizer to finite graphs fully quantitative and self-contained. The revision will be incorporated in the next version.","revision_made":"yes","referee_comment":"[§4–5] The passage from the graphon maximizer back to finite graphs (presumably §4 or §5) must include an explicit error bound showing that the spectral sum of any finite graph is at most the graphon value plus a term that vanishes relative to n; without this quantitative transfer the inequality for finite n is not yet established."}],"tokens_in":1171,"tokens_out":309,"duration_ms":21391,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The paper settles the Ebrahimi-Mohar-Nikiforov-Ahmady conjecture: for any graph on n vertices, λ1(G) + λ2(G) ≤ (8/7)n. That is the central result, and the authors deliver a proof rather than a partial bound or asymptotic statement. They lift the problem to the graphon space, use convex geometry and exterior algebra to identify the extremal graphon, run a convex optimization step to extract the constant 8/7, and pull the inequality back to finite graphs. The combination of those tools is not standard in earlier spectral-sum papers, so the method itself is the main novelty. The argument structure avoids self-referential parameters and relies on established limit theory, which keeps the circularity burden low. The transfer step from graphon to finite n appears direct, with no extra restrictions flagged in the outline. One soft spot is that the precise handling of the optimization and the equality cases in the exterior-algebra step would need close checking in the full text; small gaps there could affect how sharp the bound feels in practice. Otherwise the logic holds together. This is useful reading for anyone working on eigenvalue extremal problems or graph limits. It is not a broad survey, but it cleanly resolves a concrete open question with a reproducible technique. I would bring it to a reading group focused on spectral methods and would cite the result once the details are verified. Send it to peer review; the conjecture is old enough and the proof approach different enough that referees will want to see the write-up.","headline":"They prove the 2008 spectral sum conjecture with a graphon-plus-convex-geometry argument that looks clean on the surface.","tokens_in":2133,"tokens_out":384,"would_cite":true,"duration_ms":16258,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":{"model":"grok-4.3","evidence":[],"headline":"Spectral graph bound via graphons, Carathéodory on ellipse, exterior algebra and SOS; no J-cost, φ-ladder or distinction forcing","alignment":"orthogonal","rationale":"Paper proves λ1(G)+λ2(G)≤(8/7)n by lifting to graphons, deriving ellipse equation μ1 f² + μ2 g² = const, applying Carathéodory to obtain ≤6-step eigenfunctions, reducing to weighted H6, then using additive compound matrix ψ(M) and sum-of-squares to certify 8/7 I − ψ(M) ≽ 0. None of the machinery (graphon cut-norm compactness, ellipse from Rayleigh quotients, 2-dimensional convex hull, 15-dimensional exterior algebra, rational SOS) parallels RS theorems on J(x)=½(x+x⁻¹)−1, φ-fixed-point, 8-tick periodicity or parameter-free constant extraction. The numerical factor 8/7 is coincidental with RS 8-tick period and is not derived from any recognition-cost identity.","tokens_in":57449,"confidence":"high","tokens_out":237,"duration_ms":13601,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"lean_confirmation":null,"pith_extraction":{"msc":["05C50","05C35"],"pacs":[],"model":"grok-4.3","headline":"Any graph on n vertices has its two largest adjacency eigenvalues summing to at most 8n/7.","keywords":["spectral graph theory","adjacency eigenvalues","graph limits","extremal problems","convex optimization","graphons"],"falsifier":"A single explicit graph on n vertices whose adjacency matrix has λ1 + λ2 strictly larger than (8/7)n.","tokens_in":2510,"feed_emoji":"📈","tokens_out":641,"duration_ms":17691,"temperature":0.7,"pith_summary":"The paper establishes that for any graph G with n vertices the spectral sum λ1(G) + λ2(G) is bounded above by 8n/7. This settles a conjecture posed in 2008. The proof proceeds by lifting the finite-graph problem into the space of graphons, then applying convex-geometry and exterior-algebra arguments to obtain an exact upper bound that is attained by certain explicit constructions. A reader cares because the leading eigenvalues control expansion, connectivity, and mixing rates, so a uniform cap on their sum constrains the global structure of all graphs.","feed_headline":"Spectral sum of any graph bounded by 8n/7","feed_subtitle":"2008 conjecture proved by lifting finite graphs to graphons and applying convex optimization","key_machinery":"Graphon formulation of the spectral sum together with convex optimization over the space of symmetric measurable functions that represent the limiting adjacency operators.","core_discovery":"The spectral sum λ1(G) + λ2(G) satisfies λ1(G) + λ2(G) ≤ (8/7)n for every graph G of order n. The inequality is proved by embedding G into the space of graphons, formulating the sum as a continuous functional, and showing via convex optimization and exterior-algebra identities that the functional never exceeds 8/7 times the measure of the vertex set.","pith_inferences":["The same graphon-plus-convex-geometry pipeline is likely to resolve analogous conjectures for the sum of the three largest eigenvalues.","Because the bound is asymptotic-sharp, it gives precise control on the eigenvalue gap for dense graphs whose second eigenvalue is close to the first.","The exterior-algebra step suggests that similar identities may bound higher-order spectral invariants such as the trace of powers of the adjacency matrix."],"forward_implications":["The bound is tight and is achieved by a family of blow-up graphs whose limiting graphon is the optimizer.","Any graph whose spectral sum meets or exceeds (8/7)n must be close in cut distance to one of the extremal graphons.","The same analytic setup yields upper bounds on other linear combinations of the largest k eigenvalues for fixed k.","The methods extend verbatim to weighted graphs and to the normalized Laplacian."],"fun_headline_variants":["8n/7 caps graph spectral sums","Spectral sum ≤8n/7 confirmed for graphs","8n/7 is spectral sum upper bound","Conjecture settled: spectral sums ≤8n/7"],"cache_read_input_tokens":64,"weakest_assumption_plain":"The continuous relaxation via graph limits recovers the exact maximum value attained by finite graphs.","fun_headline_variants_meta":{"raw":{"variants":["8n/7 caps graph spectral sums","Spectral sum ≤8n/7 confirmed for graphs","8n/7 is spectral sum upper bound","Conjecture settled: spectral sums ≤8n/7"]},"model":"grok-4.3","cost_usd":0.005378,"raw_usage":{"total_tokens":2466,"prompt_tokens":576,"num_sources_used":0,"completion_tokens":53,"cost_in_usd_ticks":53778000,"prompt_tokens_details":{"text_tokens":576,"audio_tokens":0,"image_tokens":0,"cached_tokens":64},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1837,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":576,"tokens_out":53,"duration_ms":19076,"temperature":1.0,"reasoning_tokens":1837,"cache_read_input_tokens":64,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-05-13T22:42:00.067981+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A single explicit graph on n vertices whose adjacency matrix has λ1 + λ2 strictly larger than (8/7)n.","supporting_citations":[],"review_version":1}