{"id":"5312d30c-21d6-4ae6-b1be-20b7a8391f5e","arxiv_id":"2411.08184","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"For every graph, min{s+, s-} is at least 2m/chi_vec(G), resolving a conjecture of Wocjan, Elphick and Anekstein.","lead":"This paper proves a 2018 conjecture bounding the sum of squares of positive or negative adjacency matrix eigenvalues of a graph by a function of the vector chromatic number, and gives partial progress on a related Bollobas and Nikiforov conjecture. It also shows that rank-constrained vector chromatic number SDP solutions are NP-hard to compute.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection: the alleged diagonal-term omission in Lemma 20 is a notational convention, not an error.","rationale":"The reader's conditional verdict rests on a claimed error in Lemma 20: that the Motzkin-Straus inequality applied to the completely positive matrix W circ W yields a lower bound on diagonal terms plus nonedge terms, not on nonedge terms alone. This is correct as a statement about off-diagonal nonedges, but it is not the sum used in the paper. Throughout Section 3.4, the expression sum_{ij notin E} X_ij^2 is interpreted over all ordered pairs (i,j) such that {i,j} is not an edge, and since there are no loops, this includes i=j. This is forced by the claimed equivalence in Lemma 12: the feasibility constraint of program (8) is <I+bar A, X circ X> = 1, and <I+bar A, X circ X> = sum_i X_ii^2 + sum_{i≠j, ij notin E} X_ij^2. Thus the left side of the Lemma 12 inequality is exactly the constraint value, and no diagonal terms are missing. In Lemma 20, W circ W is completely positive because the w_i lie in the nonnegative quadrant, so inequality (5) applied to W circ W gives <I+bar A, W circ W> >= (1/omega)<J, W circ W>, which is precisely the claimed sum_{ij notin E} W_ij^2 >= (1/omega) sum_{i,j} W_ij^2 under the same convention. The complete-graph example that would refute the off-diagonal reading is not a counterexample here, because the diagonal is included on the left. I therefore find no load-bearing mathematical error in the proof of the asymptotic Bollobas-Nikiforov bound. The paper proves Theorem 1, Theorem 3, and Theorem 4 as stated, and the only suggestion is to make the 'ij notin E' convention explicit to avoid precisely this kind of misreading.","tokens_in":18455,"tokens_out":26880,"duration_ms":232916,"concrete_test":"Re-derive the displayed inequality in the proof of Lemma 20 from equation (5) with Z = W circ W. Verify that (5) yields <I+bar A, W circ W> >= (1/omega)<J, W circ W>, and that <I+bar A, W circ W> equals the paper's sum over ij notin E of W_ij^2 when diagonal pairs are counted. Then instantiate a complete graph with nonzero w_i in the quadrant and confirm the inequality holds: the left side equals the full sum, so no counterexample arises.","verdict_should_be":"ACCEPT","load_bearing_attack":"The reader's attack on Lemma 20 depends on reading the sum over 'ij notin E' as excluding diagonal pairs. In the manuscript, however, the sums in Lemma 12, Theorem 24, and the proof of Lemma 20 are over ordered pairs (i,j) with ij notin E, and a loop is not an edge, so i=j is included. This is not an ad hoc choice: Lemma 12 is asserted to be equivalent to sum_{ij notin E} X_ij^2 >= 1/(omega+50omega^{5/6}) sum_{i,j} X_ij^2, and the left side must equal the feasibility expression <I+bar A, X circ X> = sum_i X_ii^2 + sum_{i≠j, ij notin E} X_ij^2 for that equivalence to hold. With this convention, applying the Motzkin-Straus inequality (5) to the completely positive matrix W circ W gives exactly sum_{ij notin E} W_ij^2 >= (1/omega) sum_{i,j} W_ij^2, because <I+bar A, W circ W> is precisely the left-hand side. The diagonal terms are therefore not omitted; they are part of the left sum. Hence the proof of Lemma 20, and with it Theorem 24 and Theorem 4, is mathematically sound. The only real improvement would be for the authors to state explicitly that 'ij notin E' includes diagonal pairs, since the convention is otherwise ambiguous.","agreement_with_reader":"disagree"},"referee_report":null,"author_rebuttal":null,"desk_editor":null,"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-12T21:57:41.570580+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}