{"id":"0f586c4f-d1c6-45ba-b876-9b0f5c71257f","arxiv_id":"2611.16338","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Every finite simple cubic graph on n vertices has total edge irregularity strength n/2+1, confirming the Ivančo-Jendrol' conjecture for the entire class of cubic graphs.","lead":"Every finite simple cubic graph, in which every vertex has exactly three neighbors, can be labeled so that all edge sums are distinct using n/2+1 labels, which is the smallest possible. The result settles a longstanding conjecture about these graphs and gives an explicit recipe for building such labels.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"The swap step assumes existence of a vertex u in U with all three neighbors in U when dU > 2, but the argument only guarantees average degree > 2, not that any vertex has degree 3 inside U.","rationale":"The reader's weakest_assumption focused on the existence of a subtree of prescribed order, which is standard and true. The more serious concern is the swap step's existence claim, which is not justified and is central to the reduction. The paper does not provide a proof that such vertices u and v always exist when a > s. Without this, the algorithm may not terminate, and the main theorem is not established. The concern is load-bearing because the entire reduction from a > s to a ≤ s depends on it. The concrete test on a small cubic graph would reveal whether the claim holds in all cases or if a counterexample exists. I recommend CONDITIONAL acceptance pending a justification or modification of the swap step.","tokens_in":8427,"tokens_out":640,"duration_ms":6263,"concrete_test":"Construct a cubic graph on n=6 vertices (e.g., the triangular prism) and run the described algorithm: choose initial U as described, then attempt the swap step. Check whether at each iteration with a > s there exists a vertex u in U with all three neighbors in U and a vertex v in V(G)\\U with all three neighbors outside U, with u and v nonadjacent. If the algorithm fails to find such a pair for some cubic graph, the proof is incomplete.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In the iterative reduction step (after [b56]), the authors write: 'It follows that we can find a vertex u ∈ U (resp. v ∈ V(G)\\U) all three of whose neighbors are in U (resp. V(G)\\U).' This is a non-sequitur. The average degree of the induced subgraph G[U] is dU = 2a/|U| = 2a/(s-1). When a > s, we have dU > 2s/(s-1) > 2, but average degree > 2 does not imply any vertex has degree 3 inside U. For a cubic graph, vertices in U can have 0,1,2,3 neighbors in U. The average being >2 only forces that there is at least one vertex with degree at least 3 (i.e., degree 3) in G[U], but the swap step requires a vertex u ∈ U with all three neighbors in U *and* a vertex v ∈ V(G)\\U with all three neighbors outside U. The latter condition is symmetric but also not guaranteed by average degree >2 in the complement. Moreover, the swap step also requires that u and v are not adjacent, which is not argued. If such a pair does not exist, the reduction cannot proceed and the algorithm may stall before reaching a ≤ s. This is a gap in the constructive proof.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper claims to prove that every finite simple cubic graph on n vertices has total edge irregularity strength n/2+1, confirming the Ivančo-Jendrol' conjecture for cubic graphs. The proof constructs a vertex labeling with labels 1 and s=n/2+1, using a subset U of size n/2, then iteratively adjusts U via swaps to achieve a certain triple (a,b,c) satisfying the admissibility lemma, with a final adjustment for one exceptional case.","tokens_in":8823,"tokens_out":2962,"duration_ms":19372,"significance":"If correct, this would be a significant result, settling the conjecture for all cubic graphs, complementing Pfender's large-graph result with an explicit construction for all orders. The paper provides a deterministic algorithm, which is valuable.","major_comments":[{"comment":"The claim that 'we can find a vertex u in U (resp. v in V(G)\\U) all three of whose neighbors are in U (resp. V(G)\\U)' is not justified. The average degree dU > 2 only guarantees existence of a vertex with degree at least 3 in G[U], but not that all three neighbors are in U. In a cubic graph, a vertex in U could have degree 3 in G[U] (i.e., all three neighbors in U) or degree 2 in G[U] (two neighbors in U, one outside) or degree 1 or 0. The average being >2 forces at least one vertex with degree 3 in G[U], but the symmetric condition for V(G)\\U is not guaranteed by the same average degree argument because the average degree in the complement is also dU > 2, so similarly there exists a vertex in V(G)\\U with all three neighbors in V(G)\\U. However, the swap step also requires that u and v are not adjacent, which is not argued. Moreover, the algorithm requires that such a pair exists simultaneously; the existence of a vertex u in U with all neighbors in U and a vertex v in V(G)\\U with all neighbors in V(G)\\U does not guarantee they are non-adjacent. If they are adjacent, swapping would change the count differently. The paper does not address this. This is a gap in the constructive proof.","section":"Section 3, after equation [b56]","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The construction of U uses a subtree S of a spanning tree T of component G_{k+1} with exactly n/2 - ρ vertices. The paper states 'pick any subtree S of T having n/2 - ρ vertices.' It is a known fact that every tree on N vertices contains a subtree of every order from 1 to N, but this is not stated. While this is a standard fact, the paper should at least mention it or provide a brief justification. However, this is a minor issue; the major gap is the swap step.","section":"Section 3, initial construction of U","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"In the exceptional case (a,b,c) = (s-2, s+1, s-2), the paper changes the label of a vertex u from 1 to 2. It claims that 'we can find a vertex u in U that has at most one neighbor in U' because dU < 2. This is correct: average degree < 2 implies existence of a vertex with degree at most 1 in G[U]. However, the subsequent edge label assignments for the two subcases (0 or 1 neighbor in U) need to be checked for consistency with the admissibility lemma or directly verified. The tables provided seem plausible, but the paper does not verify that all edge labels are in [1,s] and that the weight intervals are disjoint. For the first subcase (0 neighbors), the table assigns edge labels for E2s as [3,s], but the number of edges in E2s is s-2, so the labels [3,s] have size s-2? Actually [3,s] has s-2 elements (3,4,...,s). That's fine. But the weight intervals: E2s weights are [2s+3, 3s]? Let's compute: if edge label is l, weight = 2s + l? Actually for E2s, both endpoints have label s, so weight = s + l + s = 2s + l. So if l ranges from 3 to s, weight ranges from 2s+3 to 3s. That's disjoint from previous interval [2s, 2s+2]? The previous interval is from Es+2: weight = 1 + l + s? Wait, need to check carefully. The paper's table for the first subcase: Es+2 edge labels [s-2, s], weights [2s, 2s+2]. For l=s-2, weight = 1 + (s-2) + s = 2s-1? Actually the vertex labels: u is changed to 2, so the edge in Es+2 has endpoints: one with label 2, one with label s? Actually Es+2 is defined as edges where sum of endpoint labels = s+2. Since u has label 2, its neighbor in U? Wait, the subcase is u has 0 neighbors in U, so all three neighbors are in V(G)\\U with label s. So edges incident to u are between label 2 and label s, sum = s+2, so they belong to Es+2. The other edges in Es+2? Possibly none. The table says Es+2 edge labels [s-2, s], weights [2s, 2s+2]. But if edge label is s-2, weight = 2 + (s-2) + s = 2s? Actually 2 + (s-2) + s = 2s. If edge label is s, weight = 2 + s + s = 2s+2. So that matches. Then E2s weights: edge label l, weight = s + l + s = 2s + l. For l from 3 to s, weight from 2s+3 to 3s. So disjoint from [2s,2s+2]. Good. However, the paper does not verify that the edge labels for E2s are within [1,s] and that the number of labels matches the count. The table says E2s edge labels [3,s], which has size s-2, matching c=s-2. But is the label 1 used?","section":"Section 3, final adjustment case","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The inequality a >= n/2 - 1 = s-2 is derived assuming S is a tree. But S is a subtree of a spanning tree, so it is a tree. The number of edges in S is (n/2 - ρ) - 1. The sum includes edges from components G_i for i=1..k, each cubic, so |E(G_i)| = 3n_i/2. The sum of those is 3ρ/2. So a >= 3ρ/2 + (n/2 - ρ - 1) = n/2 + ρ/2 - 1. Since ρ < n/2, this is at least n/2 - 1? Actually n/2 + ρ/2 - 1 >= n/2 - 1 because ρ/2 >= 0. So a >= s-2. That's correct. But note that a is the number of edges with both endpoints in U, which includes edges inside each G_i and inside S, but also possibly edges between components? Since components are disconnected, there are no edges between different G_i. Also, S is a subtree of G_{k+1}, so edges inside S are counted. However, there might be edges in G_{k+1} that are not in S but have both endpoints in S? Since S is a subtree, it includes all vertices of S but only some edges; there could be additional edges in G_{k+1} between vertices of S that are not in the spanning tree T. The paper uses |E(S)| as the number of edges in the subtree S, but the actual number of edges in G[U] could be larger because G_{k+1} may have extra edges not in T. The paper's lower bound a >= ... uses |E(S)|, which is a lower bound on the number of edges inside U from component G_{k+1}. However, the paper says 'a ≥ ... |E(S)| = ...' which is valid because a counts all edges inside U, which includes at least the edges of S. So the inequality is correct. But the paper should clarify that S is a subtree, not necessarily an induced subgraph, and that the edges of S are a subset of edges of G[U]. This is fine.","section":"Section 3, after [b55]","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The lemma states an if-and-only-if characterization. The necessity is clear. The sufficiency construction assigns edge labels in intervals. The paper should verify that the intervals for edge labels are within [1,s] and that the weight intervals are disjoint. The proof does this, but there is a minor typo: in the definition of β, it says β = max{1, α+b+1-s}. Then the edge labels for Es+1 are from [α, α+b-1]. The largest label is α+b-1. The paper claims α+b-1 = max{b, a+b+1-s} ≤ s. But α = max{1, a+2-s}. So α+b-1 = max{b, a+b+1-s}? Let's check: if α = 1, then α+b-1 = b. If α = a+2-s, then α+b-1 = a+b+1-s. So yes. The condition a+b ≤ 2s-1 ensures a+b+1-s ≤ s. And b ≤ s from a,b,c ≤ s. So α+b-1 ≤ s. Similarly for β. This is correct.","section":"Section 2, Lemma 2.1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper states that the final distribution (a,b,c) is one of (s, s-3, s), (s-1, s-1, s-1), (s-2, s+1, s-2). This follows from a ∈ {s-2, s-1, s} and the equations 3|U| = 2a+b and 3|V(G)\\U| = 2c+b, with |U| = n/2, |V(G)\\U| = n/2, and a=c. But the paper should verify that b is determined uniquely by a. From 3n/2 = 2a+b, so b = 3n/2 - 2a = 3s-3 - 2a. For a=s, b=3s-3-2s = s-3. For a=s-1, b=3s-3-2(s-1)=s-1. For a=s-2, b=3s-3-2(s-2)=s+1. So indeed. The paper should state this explicitly.","section":"Section 3, after [b60]","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"In the exceptional case, the paper changes the label of u from 1 to 2. This changes the vertex labeling to use three labels: 1, 2, s. The admissibility lemma only covers vertex labelings using {1,s}. So the extension to a total edge irregular labeling is done directly via tables, not via Lemma 2.1. The paper should state this explicitly. It does, but the transition could be clearer.","section":"Section 3, final adjustment","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper uses the notation s = n/2 + 1. For n even, s is integer. For n odd? Cubic graphs have even number of vertices because sum of degrees is 3n, which must be even, so n is even. So s is integer. Good.","section":"General","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"The paper claims to prove the Ivančo-Jendrol' conjecture for all cubic graphs via a constructive labeling algorithm. The proof has a gap in the swap step that the authors do not address.","headline":"The paper claims to prove the Ivančo-Jendrol' conjecture for all cubic graphs via a constructive labeling algorithm. The proof has a gap in the swap step that the authors do not address.","tokens_in":9338,"tokens_out":133,"duration_ms":2476},"pith_extraction":null,"created_at":"2026-11-13T06:43:41.807280+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":null,"supporting_citations":[],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}