{"id":"de5d32b6-cd61-4ec8-a0ea-ffe3ab328152","arxiv_id":"2501.06812","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For any finite directed multigraph, the branching ratio of a node's input tree exists and equals the Perron eigenvalue of the subgraph of all nodes that can reach it.","lead":"This paper defines the branching ratio of a node's input tree as the root growth rate of the number of walks ending at that node, and proves it always exists and equals the dominant eigenvalue of the upstream subnetwork. It is a rigorous formalization of a quantity used in biological network analysis, giving the field a solid mathematical foundation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 6.1's lower bound is not established: global non-vanishing of S1 does not imply a per-residue positive lower bound, so the proof of δ(i)=ρ(i) has a gap.","rationale":"The reader's conditional verdict focuses on Theorem 8.1's independence claim and formula issues. My pass finds a more central issue in the proof of Theorem 6.1. The main theorem δ(i)=ρ(i) is very likely true — the standard imprimitive Perron–Frobenius theorem gives positive coefficients on each residue class — but the paper's written argument does not supply that fact. The jump from 'S1 not identically zero' to an eventually positive polynomial lower bound is the load-bearing step, because Theorem 7.2 and hence the general case rely on it via the Upstream Principle. I agree with the reader that the paper should be conditional, and I do not propose changing that verdict. My concern is not that the theorem is false; it is that the proof as written has a gap, and the requested revision should include the per-residue lower bound or a citation to the standard theorem.","tokens_in":20182,"tokens_out":18081,"duration_ms":189790,"concrete_test":"Carry out the omitted residue-class computation for Theorem 6.1. For an irreducible period-h matrix, fix r and isolate the block of A^{qh+r} that contributes to a_i(qh+r); write it as M_r(B_r)^q with B_r primitive and Perron eigenvalue ρ^h, and compute C_r = lim_{q→∞} (uA^{qh+r})_i / ρ^{qh+r}. For h=2 this coefficient for odd q is (1^T B w)/ρ with w the Perron eigenvector of CB; because B has no zero column and w>0, it is positive. If the same positivity is verified for every r in the general cyclic-block decomposition, the lower bound in Theorem 6.1 is repaired and Lemma 3.5 applies. If some nonnegative irreducible A yields C_r=0, that is a counterexample to the asserted lower bound and the written proof fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorem is proved by establishing a lower bound a_i(ℓ) ≥ Q(ℓ)ρ^ℓ and applying Lemma 3.5. In Theorem 6.1, the only argument for this lower bound is the claim that S1(ℓ) is not identically zero, after which the text asserts 'there exists an eventually positive polynomial Q(ℓ)' near Eq. (6.24). This inference is invalid: S1 is a finite sum of terms c_k ζ^{kℓ}ρ^ℓ, i.e. a quasi-polynomial, and a non-identically-zero quasi-polynomial can vanish on infinitely many ℓ (1+(-1)^ℓ is an example). If S1(ℓ)=0 on a full residue class modulo h, then on that class a_i(ℓ)=S_2(ℓ)=o(ρ^ℓ), so the hypotheses of Lemma 3.5 fail. Nothing in the contradiction argument using the Cesàro average (5.20) rules this out; it only proves that the sum over all ℓ is nonzero. The missing step is the standard block-cyclic Perron–Frobenius computation showing that for each residue r the dominant coefficient is positive. Since Theorem 7.2 inherits this lower bound through the Upstream Principle, the existence half of the headline result rests on this gap. This is separate from the reader-noted independence claim in Theorem 8.1.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines the branching ratio δ(i) of a node i in a finite directed multigraph as the root-limit lim_ℓ (a_i(ℓ))^{1/ℓ} of the number of walks terminating at i, and claims that this limit always exists and equals the Perron eigenvalue ρ(i) of the upstream subnetwork U(i). The proof proceeds by an upper bound from the generalized eigenspace decomposition, a lower bound obtained first for strongly connected graphs via Perron–Frobenius theory and a Cesàro-average argument, and then an Upstream Principle to pass to the general case. The paper also derives asymptotic refinements: in the irreducible case it asserts a_i(ℓ) ∼ C_r ρ^ℓ with constants depending on ℓ modulo the period, and in the general case it asserts a_i(ℓ) ∼ R_s(ℓ)ρ^ℓ with polynomials depending on ℓ modulo an lcm of periods. Motivation is drawn from biological networks and distributed systems.","tokens_in":20499,"tokens_out":12291,"duration_ms":127913,"significance":"If the main theorem is correct, the paper gives a clean and useful characterization: the branching ratio is a Perron eigenvalue of a canonically defined subgraph, reducing its computation to linear algebra. The definition is made independently of the eigenvalue conclusion, no parameters are fitted, and the Upstream Principle is a nice reduction. The asymptotic refinements in Section 8 are also potentially useful. However, the proof of the lower bound in the strongly connected case has a genuine gap, and two statements in Section 8 are internally contradicted by the paper's own examples. These issues are load-bearing for the claimed theorems and require substantive repair, although the underlying results appear likely to be true and fixable within the manuscript's framework.","major_comments":[{"comment":"The proof establishes only that S1(ℓ) is not identically zero. Since S1 is a finite sum c_k ζ^{kℓ}ρ^ℓ, it is a quasi-polynomial, and a nonzero quasi-polynomial can vanish on an entire residue class, for example 1 + (−1)^ℓ. The Cesàro-average contradiction rules out S1 ≡ 0 globally, not S1 ≡ 0 on a single residue class modulo h. Therefore the inference that 'there exists an eventually positive polynomial Q(ℓ)' with a_i(ℓ) ≥ Q(ℓ)ρ^ℓ does not follow. A per-residue argument using the cyclic class structure, or Perron–Frobenius applied to A^h, is needed. Because Theorem 7.2 inherits the lower bound through the Upstream Principle, the existence half of the main result is not proved as written.","section":"Section 6, proof of Theorem 6.1, around Eqs. (6.22)–(6.24)"},{"comment":"The assertion that the constants C_r are independent of i is false and is contradicted by the paper's own Example 1.1. That example is strongly connected and aperiodic (h = 1), and Eq. (1.2) gives a_1(ℓ) ∼ (1/√5)φ^ℓ while a_2(ℓ) ∼ (φ/√5)φ^ℓ, so the constants differ by node. Corollary 8.2 explicitly states that C may depend on i. The proof's appeal to the Upstream Principle is insufficient: U(i) = G for all i gives the same spectral data, but not the same coefficient of the Perron projection onto the i-th coordinate. Theorem 8.1 should be corrected to C_r = C_r(i), or the independence claim removed.","section":"Section 8.1, Theorem 8.1"},{"comment":"The sentence claiming that the Upstream Principle implies the polynomials R_k are the same for all nodes in the same SCC does not follow and is false by the same example: nodes 1 and 2 of Example 1.1 lie in the same strongly connected component and have different asymptotic constants. Equality of the upstream subnetwork U(i) = U(j) does not imply equality of the i-th and j-th coordinates of the generalized-eigenvector expansion of u. The statement 'but are the same for all nodes in the same SCC of G' should be removed or replaced by an explicit coordinate-wise computation.","section":"Section 8.2, proof of Theorem 8.3, final paragraph"}],"minor_comments":[{"comment":"All four displayed cases are labelled ℓ ≡ 0 (mod 4), which is presumably a typo for residues 0, 1, 2, 3; additionally the four expressions do not appear to match the claimed formula a_6(ℓ) = ⌈ℓ/4⌉ for the included values. Please correct the residue labels and verify the constants.","section":"Example 4.4, Eq. (4.19)"},{"comment":"The lines 'a_2(ℓ) = 0' and 'a_2(ℓ) = 2' appear to be typos. For the displayed adjacency matrix, node 2 has a self-loop and node 3 has no incoming edges, so the expected values are a_1(ℓ) = 3·2^ℓ, a_2(ℓ) = 2 for ℓ ≥ 1, and a_3(ℓ) = 0 for ℓ ≥ 1.","section":"Example 4.6"},{"comment":"The cross-reference 'As in Lemma 3.16' should be 'Lemma 3.6'.","section":"Section 3.2, after Eq. (3.17)"},{"comment":"Lemma 3.6 states that the polynomial P_λ(ℓ) is over R and 'clearly must be eventually positive', but the construction gives a complex polynomial. Remark 3.9 later addresses real and imaginary parts, but the wording in the lemma should be corrected for consistency.","section":"Lemma 3.6 and Remark 3.9"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and the central characterization is plausible and likely repairable. The proof gap in Theorem 6.1 is substantive, and the Section 8 independence claims are internally inconsistent with the paper's own examples. I would not reject on novelty grounds, but the authors need to supply the per-residue Perron–Frobenius argument and correct the asymptotic statements before the paper can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nQuick take on Boldi–Stewart, arXiv:2501.06812. The headline result—branching ratio δ(i) = lim (a_i(ℓ))^{1/ℓ} exists for every node and equals the Perron eigenvalue of the upstream subgraph—is correct in substance and, as far as I can tell, genuinely new in this exact form. They also give the right asymptotic refinements (residue-class constants for irreducible case, polynomials for general case). The biological/dynamical motivation is well explained, and the paper is clearly written. So it's worth reading.\n\nBut there are two problems.\n\nFirst, a real gap in the proof of Theorem 6.1, the existence half. They prove that S1(ℓ), the dominant spectral sum, is not identically zero, then immediately assert there is an eventually positive polynomial Q(ℓ) with a_i(ℓ) ≥ Q(ℓ)ρ^ℓ. That does not follow. S1 is a quasi-polynomial: a finite sum of terms c_k ζ^{kℓ}ρ^ℓ. It can be nonzero as a function yet vanish on an entire residue class modulo h (e.g. 1+(-1)^ℓ vanishes on odd ℓ). The Cesàro argument only shows the average over all ℓ is positive, not that each residue class contributes. Without a per-residue positive lower bound, Lemma 3.5 cannot be applied. The fix is standard—use the cyclic structure from Perron–Frobenius to show each residue class has a positive dominant coefficient—but it is not in the paper. As written, the proof of δ(i)=ρ(i) is incomplete.\n\nSecond, Theorem 8.1 states the constants C_r are independent of i. Their own Example 1.1 shows a_1(ℓ) ~ (1/√5)φ^ℓ and a_2(ℓ) ~ (φ/√5)φ^ℓ, different constants. So that claim is false as stated. The error is localized; the rest of Section 8 still stands if the independence assertion is dropped or qualified.\n\nMinor: Example 4.6 has a typo (a_2(ℓ) duplicated), and Example 4.4's formula has the same residue condition repeated. Nothing affecting the math.\n\nBottom line: the central result is likely true, the writing is honest (they even say the result is not a surprise), and the proof strategy is sound in outline. But the lower-bound gap is load-bearing and needs to be fixed before publication. It is fixable, not fatal. I'd send it to review, with instructions to the authors to supply the missing block-cyclic argument and correct Theorem 8.1.\n\nFor you: if you work on network synchrony or spectral graph theory, worth a look. I wouldn't cite it until the proof is patched.","headline":"True headline result, but the proof has a real lower-bound gap and Theorem 8.1 contradicts Example 1.1.","tokens_in":20961,"tokens_out":6156,"would_cite":false,"duration_ms":52177,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C50","05C20","15A18"],"pacs":[],"model":"deepseek-v4-flash","headline":"The branching ratio of every node in a finite directed multigraph exists and equals the Perron eigenvalue of the subgraph of nodes that feed into it.","keywords":["branching ratio","input tree","directed multigraph","Perron eigenvalue","upstream subnetwork","walk counts","Perron-Frobenius theorem","asymptotic growth"],"falsifier":"Build a network where the upstream subnetwork of node $i$ is the union of two strongly connected components with the same Perron eigenvalue $\\rho$ and periods $2$ and $3$, feeding into $i$. Compute $a_i(\\ell)=(uA^\\ell)_i$ for $\\ell=0,\\ldots,12$: the theorem predicts $(a_i(\\ell))^{1/\\ell}\\to\\rho$ and that $a_i(\\ell)/\\rho^\\ell$ is eventually periodic with period $6$. If the root limit differs from $\\rho$, or the period is not $6$, the central claim fails.","tokens_in":20021,"feed_emoji":"📈","tokens_out":8891,"duration_ms":75607,"temperature":0.7,"pith_summary":"The paper gives a general definition of the branching ratio of a node in a finite directed multigraph: the growth rate of the number of walks of length $\\ell$ that end at that node. It proves that the root limit $\\delta(i)=\\lim_{\\ell\\to\\infty}(a_i(\\ell))^{1/\\ell}$ always exists and equals the Perron eigenvalue $\\rho(i)$ of the upstream subnetwork $U(i)$, the subgraph induced by all nodes that can reach $i$. This matters because branching ratios are used in biology and distributed computing to quantify how information accumulates at a node, and earlier ratio-based definitions can fail to converge for perfectly ordinary networks. The paper also pins down the fine asymptotics: in the strongly connected case $a_i(\\ell)$ is asymptotic to $C_r\\rho^\\ell$ with $r\\equiv \\ell\\pmod h$, and in general to a polynomial $R_s(\\ell)\\rho^\\ell$ whose period is the least common multiple of the periods of the dominant strongly connected components.","feed_headline":"Every node's branching ratio exists and equals one eigenvalue","feed_subtitle":"Walk growth into any node of a directed multigraph has a limit: the Perron eigenvalue of the subnetwork feeding it.","key_machinery":"The load-bearing objects are the aggregate walk count $a_i(\\ell)=(uA^\\ell)_i$, with $u=(1,\\ldots,1)$, and the Perron eigenvalue $\\rho(i)$ of the upstream subnetwork $U(i)$. The proof splits $a_i(\\ell)=\\sum_\\lambda P_\\lambda(\\ell)\\lambda^\\ell$ into dominant terms with $|\\lambda|=\\rho$ and subdominant terms, then uses the Cesàro-average limit $\\lim_{k\\to\\infty}\\frac1k\\sum_{\\ell=0}^{k}\\rho^{-\\ell}A^\\ell = vw^T$ from the Perron–Frobenius theorem to rule out cancellation of the dominant terms. The Upstream Principle—any walk ending at an upstream node can be extended by a fixed walk to $i$—carries the lower bound from strongly connected components to arbitrary nodes.","core_discovery":"The central claim is that for every node $i$ of any finite directed multigraph, the sequence $(a_i(\\ell))^{1/\\ell}$ converges, and its limit $\\delta(i)$ is the largest real eigenvalue of the adjacency matrix of $U(i)$. This is proved first for strongly connected graphs, using the decomposition of $a_i(\\ell)$ into a finite sum of polynomials times powers of eigenvalues and the Cesàro-average form of the Perron–Frobenius theorem to show that the dominant spectral terms do not cancel in the aggregate walk count. The general case follows by the Upstream Principle: a lower bound on walk counts transfers from any upstream node to the node in question, so the branching ratio is the maximum of the branching ratios of the strongly connected components that feed into $i$.","pith_inferences":["Readers working with weighted or continuous-time networks can expect the same theorem to hold for non-negative real weights, since the Perron–Frobenius and decomposition arguments do not use integrality of edge multiplicities.","The asymptotic periodic refinement suggests that in biological networks with several upstream modules, the branching ratio alone is too coarse: short-time walk counts will oscillate with period $g$ even though the root limit is stable.","A practical numerical shortcut follows: compute $\\rho(i)$ directly from the upstream adjacency matrix instead of simulating walks; the subdominant spectral gap controls the error."],"forward_implications":["Every node has a well-defined branching ratio, even when the ratio $a_i(\\ell+1)/a_i(\\ell)$ fails to converge, as in the three-step doubling example.","The branching ratio is computable by linear algebra: it is the Perron eigenvalue of the upstream subnetwork, not of the whole graph.","Nodes in the same strongly connected component have the same branching ratio, because their upstream subnetworks coincide.","If the upstream subnetwork is acyclic the branching ratio is $0$, matching the convention that finite input trees grow at rate zero.","The asymptotic shape of $a_i(\\ell)$ is periodic in $\\ell$ with period equal to the lcm of the periods of the dominant strongly connected components; the constant prefactor becomes a polynomial in the general reducible case."],"supporting_citations":[{"why":"gives the standard formula $a_i(\\ell)=(uA^\\ell)_i$ that ties walk counts to powers of the adjacency matrix.","marker":"[4]"},{"why":"states the Perron–Frobenius theorem including the Cesàro-average limit used to prove the non-cancellation of dominant terms.","marker":"[16]"},{"why":"supplies the block upper-triangular form of the adjacency matrix of an upstream subnetwork over its strongly connected components, used in the general case.","marker":"[18]"},{"why":"provides another statement of the Perron–Frobenius theorem and spectral properties of non-negative matrices used throughout.","marker":"[21]"},{"why":"offers a further reference for the Perron–Frobenius theorem and matrix analysis facts cited for generalised eigenspaces.","marker":"[26]"},{"why":"defines input trees and graph fibrations, the combinatorial objects whose growth the branching ratio measures.","marker":"[7]"}],"fun_headline_variants":["Branching ratio: node growth limit equals Perron eigenvalue","Input tree growth limit exists, equals top upstream eigenvalue","For every node, branching ratio is the Perron eigenvalue of feeds","Directed multigraphs: node branching ratio converges to feeding eigen","Branching ratio proven: always the largest eigenvalue of upstream subgraph"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof relies on counting all walks that end at a node, summed over every possible starting node, so that the dominant eigenvalue contributions cannot cancel; if one fixed the starting node, the walk count could be zero for infinitely many lengths and the root limit might not exist.","fun_headline_variants_meta":{"raw":{"variants":["Branching ratio: node growth limit equals Perron eigenvalue","Input tree growth limit exists, equals top upstream eigenvalue","For every node, branching ratio is the Perron eigenvalue of feeds","Directed multigraphs: node branching ratio converges to feeding eigen","Branching ratio proven: always the largest eigenvalue of upstream subgraph"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000133,"raw_usage":{"total_tokens":1053,"prompt_tokens":779,"completion_tokens":274,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":395,"completion_tokens_details":{"reasoning_tokens":189}},"tokens_in":395,"tokens_out":274,"duration_ms":3368,"temperature":1.0,"reasoning_tokens":189,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:50:38.064842+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Build a network where the upstream subnetwork of node $i$ is the union of two strongly connected components with the same Perron eigenvalue $\\rho$ and periods $2$ and $3$, feeding into $i$. Compute $a_i(\\ell)=(uA^\\ell)_i$ for $\\ell=0,\\ldots,12$: the theorem predicts $(a_i(\\ell))^{1/\\ell}\\to\\rho$ and that $a_i(\\ell)/\\rho^\\ell$ is eventually periodic with period $6$. If the root limit differs from $\\rho$, or the period is not $6$, the central claim fails.","supporting_citations":[{"cited_title":"Boldi and S","cited_arxiv_id":null,"evidence_quote":"defines input trees and graph fibrations, the combinatorial objects whose growth the branching ratio measures."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the standard formula $a_i(\\ell)=(uA^\\ell)_i$ that ties walk counts to powers of the adjacency matrix."},{"cited_title":"Gantmacher","cited_arxiv_id":null,"evidence_quote":"states the Perron–Frobenius theorem including the Cesàro-average limit used to prove the non-cancellation of dominant terms."},{"cited_title":"Golubitsky and I","cited_arxiv_id":null,"evidence_quote":"supplies the block upper-triangular form of the adjacency matrix of an upstream subnetwork over its strongly connected components, used in the general case."},{"cited_title":"Lancaster and M","cited_arxiv_id":null,"evidence_quote":"provides another statement of the Perron–Frobenius theorem and spectral properties of non-negative matrices used throughout."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"offers a further reference for the Perron–Frobenius theorem and matrix analysis facts cited for generalised eigenspaces."}],"review_version":1}