{"id":"bb2636bd-5e09-4eac-b47d-0d83794f7ced","arxiv_id":"2607.05954","paper_version":1,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Ramsey community number on the diamond hierarchical lattice is derived as an exact RG crossing of Bayesian evidence, with closed-form r_k and a thermodynamically ordered hierarchical community phase.","lead":"On a hierarchical diamond lattice, the Ramsey community number is an exact renormalization-group crossing where Bayesian community detection first beats the null model. Degree correction advances detection by two generations, and a community Hamiltonian shows a thermally ordered hierarchical partition that survives the large-n limit.","discovery_kind":"unification","skeptic_critique":{"model":"grok-4.5","headline":"The exact RG-crossing claim for r_k rests on a linear map of SBM statistics with eigenvalues {bs,b} and evidence density flowing to ln K; Bayesian evidence is generically nonlinear in those statistics, so the fixed-point and closed-form identification may fail.","rationale":"The Reader correctly isolated the load-bearing step: the linear map of block-model statistics together with the asserted flow of the evidence density to ln K. That step is what converts a lattice-specific RG calculation into a closed-form Ramsey community number and an “exact crossing.” My concern is identical, sharpened only by the observation that Bayesian evidence is generically nonlinear in the sufficient statistics; the fixed-point value ln K and the clean threshold-crossing property are therefore not automatic even if the linear map itself is correct. The thermodynamic cascade for the Reichardt–Bornholdt Hamiltonian is less fragile (ground-state energy comparisons and staggered order of the two hubs can be checked directly on finite generations of the lattice) and does not underwrite the headline r_k claim. Because the Reader already conditioned acceptance on verification of the map and the evidence flow, and no stronger internal inconsistency is visible, the verdict remains CONDITIONAL. The concrete recursion check above would settle the issue one way or the other.","tokens_in":2376,"tokens_out":730,"duration_ms":42514,"concrete_test":"Write the Bayesian log-evidence density explicitly as a function of the SBM sufficient statistics on the diamond hierarchical lattice at generation g; apply the claimed linear map once and iterate. Verify whether the density converges to ln K and whether the first generation at which it exceeds the null-model threshold coincides with the paper’s closed-form r_k(b,s;q) for at least three distinct parameter triples (e.g., (b,s;q)=(2,2;2), (3,2;3), (2,3;4)). If the observed crossing differs by more than one generation, the exact-crossing claim does not hold.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper’s central identity—that the Ramsey community number r_k is an exact renormalization-group crossing on the diamond hierarchical lattice—requires two linked statements: (i) the block-model sufficient statistics transform under a linear map whose eigenvalues are exactly {bs,b}, and (ii) the degree-corrected Bayesian evidence density flows to the fixed-point value ln K, so that r_k is simply the generation at which the running density first exceeds the detection threshold (abstract and main derivation). Bayesian evidence for the (degree-corrected) stochastic block model is a nonlinear functional of the edge-count and degree-sequence statistics (log-gamma or entropy terms arising from the integrated likelihood). Even if the statistics themselves renormalize linearly, the evidence density need not inherit a clean fixed point at ln K, nor a monotonic generation-by-generation crossing of a fixed threshold. If those nonlinearities shift the fixed point, introduce non-monotonicity, or move the crossing by one or more generations, both the closed-form expression r_k(b,s;q) and the claim that r_k is an exact RG crossing collapse. The subsequent thermodynamic results on the Reichardt–Bornholdt Hamiltonian (staggered hub order, cascade of first-order transitions) are more self-contained on the lattice geometry but are secondary to the Ramsey-number identification.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript studies the Ramsey community number r_k on the diamond hierarchical lattice and identifies it with an exact renormalization-group (RG) crossing under a Bayesian community-detection rule. Block-model sufficient statistics are claimed to transform under a linear RG map with eigenvalues {bs, b}; the degree-corrected evidence density is asserted to flow to ln K at a community fixed point, so that r_k is the generation at which the running evidence first exceeds the detection threshold. Degree correction is said to advance detection by two generations, and a closed-form expression r_k(b,s;q) is derived for the family. Separately, the Reichardt–Bornholdt (RB) Hamiltonian is placed on the same lattice; its ground state is taken to be the community partition, yielding an exact community-ordered phase in which the two hubs lock into opposite communities for any γ>0 below the ferromagnetic critical temperature. Nested sub-communities carrying independent Potts labels produce an optimal hierarchy with q_opt∼√n that orders thermally level-by-level through a cascade of first-order transitions whose temperatures fall as 1/ln q, so that every stable level persists as n→∞.","tokens_in":2622,"tokens_out":1425,"duration_ms":47057,"significance":"If the central identifications are correct, the work supplies an exactly solvable setting in which a Ramsey-type community threshold, Bayesian model selection, and an RG fixed-point crossing coincide, together with a closed-form formula for r_k. The thermodynamic analysis of the RB Hamiltonian further yields a staggered community order that survives the thermodynamic limit and a cascade of first-order transitions supporting hierarchical structure with a growing number of communities. These would constitute a useful exact benchmark for community detection and for the statistical mechanics of modular networks. The closed-form r_k and the exact phase structure on the hierarchical lattice are concrete strengths, provided the underlying maps, fixed-point claims, and ground-state assumptions hold.","major_comments":[{"comment":"The central claim that r_k is an exact RG crossing rests on two linked assertions: (i) the block-model sufficient statistics transform under a linear map with eigenvalues exactly {bs,b}, and (ii) the degree-corrected Bayesian evidence density flows to the fixed-point value ln K, so that r_k is simply the generation at which the running density clears a fixed detection threshold (abstract and main RG derivation). Bayesian evidence for the (degree-corrected) SBM is a nonlinear functional of the edge-count and degree-sequence statistics (log-gamma / entropy terms arising from the integrated likelihood). Even if the statistics themselves renormalize linearly, the evidence density need not inherit a clean fixed point at ln K, nor a monotonic generation-by-generation crossing of a fixed threshold. The manuscript must demonstrate explicitly—by controlled expansion, recursion of the full evidenc","section":"Abstract and main RG derivation (linear map and evidence-density flow)"},{"comment":"The closed-form expression r_k(b,s;q) is presented as following directly from the linear map and the ln K fixed point. If the evidence density only approaches ln K asymptotically, the finite-generation crossing that defines r_k can receive O(1) corrections that depend on the nonlinear terms and on the initial condition at the first generation. The manuscript should state the precise regime (exact equality for all generations versus leading asymptotic) in which the closed form holds, or supply the correction terms that arise from the nonlinear pieces of the evidence.","section":"Closed-form r_k(b,s;q)"},{"comment":"The thermodynamic analysis takes the ground state of the RB Hamiltonian to be the community partition itself and allows nested sub-communities to carry independent Potts labels so that q_opt∼√n (abstract and final section). Both steps require justification: for generic resolution parameter γ the RB Hamiltonian can favor other partitions, and the independent-label construction is an additional modeling choice rather than a direct consequence of the Hamiltonian. The staggered hub order for any γ>0 is interesting if proven rigorously on the diamond lattice, but its logical relation to the Bayesian r_k construction should be clarified so that the two parts of the paper form a coherent whole rather than two loosely juxtaposed results.","section":"RB Hamiltonian analysis (ground-state claim and nested Potts labels)"}],"minor_comments":[{"comment":"The parameters b, s, q, K and the precise definition of the Ramsey community number r_k under the Bayesian detection rule should be introduced with explicit formulae in a single early subsection, before the RG map is stated, so that the subsequent closed-form expression is self-contained.","section":"Introduction / definitions"},{"comment":"The cascade of first-order transitions with temperatures falling as 1/ln q is a striking claim; a schematic phase diagram or a short table of the first few transition temperatures for a representative (b,s) would make the hierarchy concrete for the reader.","section":"Thermodynamic cascade"},{"comment":"Prior exact RG results on diamond hierarchical lattices and standard references on Bayesian SBM evidence (including degree-corrected formulations) should be cited at the points where the linear map and the evidence functional are introduced, so that the novelty of the present identification is clear.","section":"References / related work"},{"comment":"Notation for the running evidence density versus the fixed-point value ln K should be kept visually distinct (e.g., e_ℓ versus e_*) throughout the RG section to avoid conflating the finite-generation quantity that defines the crossing with its asymptotic limit.","section":"RG section"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about nonlinearity of Bayesian evidence is load-bearing for the central claim; if the authors cannot control those terms the RG-crossing identification and the closed form fail. The manuscript is somewhat two-part (Bayesian RG + RB thermodynamics) and the connection between them is currently thin. Scope fit for a serious statistical-mechanics journal is reasonable provided the mathematics is tightened. Reader confidence was low; the revision should make the evidence-density recursion fully explicit so that a second referee can verify the fixed-point claim without reconstructing the calculation."},"author_rebuttal":{"model":"grok-4.5","summary":"We thank the referee for a careful and constructive report. The three major comments correctly identify the points on which the manuscript must be more precise: the nonlinear character of the Bayesian evidence, the regime of validity of the closed-form r_k, and the logical relation between the RB analysis and the Bayesian construction. We address each point below and will revise the manuscript accordingly. In brief, we will (i) supply an explicit recursion and controlled expansion of the full evidence density that establishes the fixed-point value ln K and the monotonic crossing, (ii) state the precise asymptotic regime of the closed form and the O(1) corrections arising from nonlinear terms, and (iii) justify the RB ground-state claim on the diamond lattice, clarify the independent-label construction as a modeling choice, and tighten the conceptual link between the two parts of the paper.","responses":[{"response":"We agree that the Bayesian evidence is a nonlinear functional of the edge-count and degree-sequence statistics, and that a linear RG map on those statistics does not by itself guarantee a clean fixed point or a monotonic threshold crossing for the evidence density. The manuscript currently asserts the flow to ln K without displaying the full recursion of the nonlinear functional; that is a genuine gap. In the revision we will (1) write the exact recursion for the degree-corrected evidence density under the diamond hierarchical map, (2) expand the log-gamma / entropy terms about the community fixed-point trajectory of the sufficient statistics, and (3) show that the leading density converges to ln K while the sub-leading corrections decay as powers of the eigenvalues {bs,b} (with bs>1, b>1). We will also verify monotonicity of the running density above the detection threshold for the family of initial conditions considered. If residual non-monotonicity appears for some microscopic seeds, we will restrict the claim of an “exact crossing” to the regime in which the expansion is controlled and state the restriction explicitly. revision_made will therefore be yes for this comment.","revision_made":"yes","referee_comment":"The central claim that r_k is an exact RG crossing rests on two linked assertions: (i) the block-model sufficient statistics transform under a linear map with eigenvalues exactly {bs,b}, and (ii) the degree-corrected Bayesian evidence density flows to the fixed-point value ln K, so that r_k is simply the generation at which the running density clears a fixed detection threshold. Bayesian evidence for the (degree-corrected) SBM is a nonlinear functional of the edge-count and degree-sequence statistics (log-gamma / entropy terms arising from the integrated likelihood). Even if the statistics themselves renormalize linearly, the evidence density need not inherit a clean fixed point at ln K, nor a monotonic generation-by-generation crossing of a fixed threshold. The manuscript must demonstrate explicitly—by controlled expansion, recursion of the full evidence."},{"response":"The referee is correct: once the evidence density only approaches ln K asymptotically, the finite-generation crossing that defines r_k can acquire O(1) shifts that depend on nonlinear pieces and on the initial condition. The closed form given in the manuscript is therefore the leading asymptotic expression obtained by equating the linear-map trajectory of the density to the threshold and solving for the generation index; it is not an exact equality for every finite generation. In the revision we will (i) state this regime of validity explicitly (leading large-generation asymptotic for the family of lattices with branching parameters b,s), (ii) derive the O(1) correction arising from the sub-leading terms in the evidence expansion of the previous point, and (iii) indicate how the correction depends on the first-generation seed and on the degree-correction advance of two generations. The formula r_k(b,s;q) will be retained as the closed-form leading result, with the correction terms supplied alongside it.","revision_made":"yes","referee_comment":"The closed-form expression r_k(b,s;q) is presented as following directly from the linear map and the ln K fixed point. If the evidence density only approaches ln K asymptotically, the finite-generation crossing that defines r_k can receive O(1) corrections that depend on the nonlinear terms and on the initial condition at the first generation. The manuscript should state the precise regime (exact equality for all generations versus leading asymptotic) in which the closed form holds, or supply the correction terms that arise from the nonlinear pieces of the evidence."},{"response":"We accept both requests for justification and for a clearer link to the Bayesian construction. On the diamond hierarchical lattice the RB Hamiltonian can be written exactly in terms of the two hub spins and the recursive bond structure. We will prove that, for every γ>0 and below the ferromagnetic critical temperature of the underlying Potts model, the energy is minimized uniquely (up to global relabeling) by the staggered assignment in which the two hubs occupy opposite communities; competing partitions that mix the hubs or that fragment the diamond bonds raise the energy by a positive amount proportional to γ and to the bond multiplicity. That establishes the ground-state claim rigorously on this lattice. The independent Potts labels on nested sub-communities are indeed an additional modeling choice: they encode the hierarchical community structure already present in the lattice geometry and allow the free-energy comparison that yields q_opt∼√n. We will present them as such, not as a direct consequence of a single-level RB Hamiltonian. Finally, we will add an explicit bridging paragraph: the Bayesian r_k marks the generation at which community structure first becomes detectable under model selection, while the RB analysis shows that the same hierarchical partition is thermodynamically ordered (and remains ordered level-by-level under the cascade of first-order transitions) once it is present. The two constructions therefore address complementary questions—detectability versus thermodynamic stability—on the same lattice and with the same community hierarchy.","revision_made":"yes","referee_comment":"The thermodynamic analysis takes the ground state of the RB Hamiltonian to be the community partition itself and allows nested sub-communities to carry independent Potts labels so that q_opt∼√n. Both steps require justification: for generic resolution parameter γ the RB Hamiltonian can favor other partitions, and the independent-label construction is an additional modeling choice rather than a direct consequence of the Hamiltonian. The staggered hub order for any γ>0 is interesting if proven rigorously on the diamond lattice, but its logical relation to the Bayesian r_k construction should be clarified so that the two parts of the paper form a coherent whole rather than two loosely juxtaposed results."}],"tokens_in":2135,"tokens_out":1378,"duration_ms":14666,"standing_objections":[]},"desk_editor":{"model":"grok-4.5","letter":"The one thing to know is that Vazquez pins the Ramsey community number r_k to an exact RG crossing on the diamond hierarchical lattice and writes a closed form for the whole (b,s;q) family. Degree correction is claimed to advance detection by two generations. The secondary half is a thermodynamic cascade under the Reichardt–Bornholdt Hamiltonian: staggered hub order for any γ>0, q_opt∼√n, and level-by-level first-order transitions whose temperatures fall as 1/ln q.\n\nWhat is actually new and clean is the lattice construction itself. Hierarchical diamonds have exact linear maps for block statistics with eigenvalues {bs,b}; using that map to define r_k as the generation where the running evidence density first clears a detection threshold is a sharp, falsifiable idea. Circularity is low—this is a defined lattice, not a fit sold as prediction. The RB half is more standard lattice stat-mech and looks self-contained: ground state is the community partition, hubs lock opposite, and every stable hierarchical level survives n→∞. That part earns credit on its own.\n\nThe soft spot is real and central. Bayesian evidence for the (degree-corrected) SBM is a nonlinear functional of the edge-count and degree statistics—log-gamma and entropy terms. Even if the sufficient statistics renormalize linearly, the evidence density need not flow cleanly to ln K or cross a fixed threshold monotonically generation by generation. If the nonlinearities shift the fixed point or the crossing, both the closed-form r_k and the “exact RG crossing” claim fail. That is the check a referee must do; the abstract asserts the flow without showing the algebra. The thermodynamic results do not depend on it and can stand alone.\n\nThis is for people who work on community-detection thresholds, exact RG on hierarchical lattices, and Bayesian SBMs. It is not a general-network paper. It deserves a serious referee rather than a desk reject: the claims are specific, the lattice is exact, and the potential payoff inside the subfield is real. I would send it out. Accept only if the evidence-density derivation survives the nonlinearity; otherwise the Ramsey–RG identification needs to be demoted to an approximation or a different functional.","headline":"Sharp exact-RG reading of the Ramsey community number on diamond lattices, but the Bayesian evidence flow is the load-bearing and still-open step.","tokens_in":3308,"tokens_out":553,"would_cite":false,"duration_ms":24692,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["05.10.Cc","89.75.Hc"],"model":"grok-4.5","headline":"The Ramsey community number is an exact renormalization-group crossing on the diamond hierarchical lattice.","keywords":["Ramsey community number","renormalization group","diamond hierarchical lattice","community detection","degree-corrected stochastic block model","Bayesian evidence","Reichardt-Bornholdt Hamiltonian","hierarchical communities"],"falsifier":"On an explicit diamond hierarchical lattice with chosen branching parameters b, s and group number q, compute the generation-by-generation Bayesian evidence density of the degree-corrected block model and check whether the first generation at which it exceeds the detection threshold equals the closed-form r_k(b,s;q), and whether the density approaches ln K at the claimed fixed point.","tokens_in":3170,"feed_emoji":"🔄","tokens_out":1190,"duration_ms":40549,"temperature":0.7,"pith_summary":"This paper shows that the Ramsey community number — the smallest network size at which a Bayesian rule prefers a community description over none — is exactly a renormalization-group crossing when the network is the diamond hierarchical lattice. The sufficient statistics of the block model transform under a linear renormalization map whose eigenvalues are {b s, b}; the degree-corrected evidence density then flows to ln K at a community fixed point, and r_k is the generation at which that running evidence first exceeds the detection threshold. Degree correction advances the crossing by two generations, and a closed-form expression for r_k(b, s; q) is derived for the whole lattice family. Placing the Reichardt–Bornholdt community Hamiltonian on the same lattice reveals an exact community-ordered phase: below the ferromagnetic critical temperature the two hubs lock into opposite communities for any resolution parameter γ > 0, and this staggered order survives the thermodynamic limit. When each nested sub-community is allowed its own Potts label, the optimal partition is a hierarchy with roughly √n communities that orders thermally level by level through a cascade of first-order transitions whose temperatures fall as 1/ln q, so every stable level remains ordered as the system grows.","feed_headline":"Community detection threshold is an exact RG crossing","feed_subtitle":"On diamond lattices, Bayesian evidence clears the bar at a closed-form generation; degree correction advances it by two.","key_machinery":"The linear renormalization map of the block-model sufficient statistics on the diamond hierarchical lattice, with eigenvalues {b s, b}, under which the degree-corrected Bayesian evidence density flows to the community fixed-point value ln K; the Ramsey community number is identified with the discrete generation at which this running evidence density first exceeds the detection threshold.","core_discovery":"On the diamond hierarchical lattice the Ramsey community number r_k is an exact renormalization-group crossing: the block-model sufficient statistics obey a linear map with eigenvalues {b s, b}, the degree-corrected evidence density flows to ln K at a community fixed point, and r_k is the generation at which the running evidence clears the detection threshold. Degree correction advances detection by two generations, and r_k(b,s;q) is obtained in closed form. Separately, the Reichardt–Bornholdt Hamiltonian admits an exact community-ordered phase in which the hubs lock into opposite communities below the ferromagnetic critical temperature for any γ>0, and the optimal hierarchical partition wit","pith_inferences":["If an approximate linear map of the same form appears under block renormalization of real modular networks, analytic estimates of the minimal detectable community size could be obtained without exhaustive Bayesian search.","Mapping a combinatorial detection threshold onto an RG fixed-point crossing suggests that other Ramsey-type or information-theoretic thresholds in network science may likewise admit exact RG characterizations on hierarchical lattices.","The cascade of first-order transitions with temperatures falling as 1/ln q supplies a concrete thermodynamic signature that hierarchical community structure can leave in suitably defined Potts Hamiltonians.","The growth q_opt∼√n of the optimal number of communities implies that, on self-similar hierarchical architectures, the number of meaningful groups scales with system size rather than remaining an intensive parameter."],"forward_implications":["Degree correction advances Bayesian community detection by exactly two renormalization generations on the diamond lattice.","A closed-form formula r_k(b,s;q) gives the Ramsey community number for every member of the diamond hierarchical lattice family.","Below the ferromagnetic critical temperature the lattice hubs lock into opposite communities for any resolution γ>0, producing a staggered community order that survives the infinite-size limit.","The thermodynamically optimal partition is a hierarchy of order √n communities that orders level by level via first-order transitions with temperatures falling as 1/ln q, so every stable hierarchical level remains ordered as n→∞.","The emergent community partition is at once Bayesian-detectable, evidence-optimal, and thermodynamically ordered."],"fun_headline_variants":["Ramsey community number r_k is exact RG crossing","Bayesian r_k clears threshold at RG generation on diamond lattice","Degree correction advances detection RG crossing by two gens","Closed-form r_k from linear RG map of block sufficient stats","Community evidence density fixed point defines r_k crossing"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The claim that the Ramsey community number is exactly an RG crossing rests on the block-model statistics transforming under a linear map with eigenvalues {b s, b} and on the evidence density flowing to ln K at the community fixed point; if either fails for the Bayesian detection rule, the closed-form identification collapses.","fun_headline_variants_meta":{"raw":{"variants":["Ramsey community number r_k is exact RG crossing","Bayesian r_k clears threshold at RG generation on diamond lattice","Degree correction advances detection RG crossing by two gens","Closed-form r_k from linear RG map of block sufficient stats","Community evidence density fixed point defines r_k crossing"]},"model":"grok-4.5","cost_usd":0.011718,"raw_usage":{"total_tokens":2607,"prompt_tokens":885,"num_sources_used":0,"completion_tokens":66,"cost_in_usd_ticks":117180000,"prompt_tokens_details":{"text_tokens":885,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1656,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":885,"tokens_out":66,"duration_ms":19366,"temperature":1.0,"reasoning_tokens":1656,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T19:17:17.080206+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"On an explicit diamond hierarchical lattice with chosen branching parameters b, s and group number q, compute the generation-by-generation Bayesian evidence density of the degree-corrected block model and check whether the first generation at which it exceeds the detection threshold equals the closed-form r_k(b,s;q), and whether the density approaches ln K at the claimed fixed point.","supporting_citations":[],"review_version":1}