{"id":"88d4774d-d78a-4854-8498-34eda6f41d64","arxiv_id":"1908.06241","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"This paper constructs graphon-valued stochastic processes as scaling limits of Moran-type population models, driven by Wright-Fisher and Fleming-Viot diffusions.","lead":"This paper builds a mathematical framework for random networks that change over time, and shows that certain population genetics models produce such evolving networks with well-defined limits. The result connects graph limit theory with population genetics.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Example 4.3's H^m is discontinuous, so the paper's flagship non-block Fleming-Viot example falls outside Theorem 3.4's hypotheses.","rationale":"The reader's conditional verdict is reasonable, and my stress-test does not move it. The main convergence theorems are proved conditionally on (H1), and the proof of Theorem 3.4 is internally coherent under that assumption. However, the paper's most important illustrative construction, Example 4.3, violates (H1): the defined H^m is a step function in the graphon coordinate, not an element of C([0,1],[0,1]), so Theorem 3.4 does not apply to it as written. This is a concrete technical flaw that undermines the advertised claim that the framework produces general non-block graphon diffusions. I also agree with the reader's separate concern that the Markov property asserted in Section 4.2 requires invertibility of the map from the population state to the graphon, which is not established and can fail in simple cases such as r ≡ 1. Neither issue invalidates the central conditional convergence theorems, but together they justify keeping the verdict conditional rather than accepting the paper's full set of claims as they stand.","tokens_in":17650,"tokens_out":14487,"duration_ms":139890,"concrete_test":"Verify whether H^m(s;u) as defined in Example 4.3 belongs to C([0,1],[0,1]) in the simple case m = 1, Y_0 = Y_1 = 1/2, f(x,y) = xy, c = 1/4. Compute H^1(s;u) explicitly: it equals 0 for u < 1/2 and 1/2 for u ≥ 1/2, so it has a jump at u = 1/2 and therefore fails (H1). Alternatively, if the intention was linear interpolation between the grid values H^m(s; ℓ/(m+1)), recompute the example with that definition and check whether the displayed quantile formula is actually the correct representative of that interpolated function; it is not, because the formula gives a step function, not a continuous interpolation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 3.4 assumes (H1), which requires H^m and H to be random elements of D([0,∞), L) with L = C([0,1],[0,1]). In Example 4.3, the paper defines H^m(s;u) = ∫ I[f(bar F^m(s;u), bar F^m(s;v)) ≥ c] Z^m(dv). For finite m, Z^m is atomic with atoms at {0,...,m}/(m+1), so bar F^m(s;·) is a step function; composing it with f and integrating against Z^m yields a step function in u, not a continuous function. A concrete instance: take m = 1, Y_0 = Y_1 = 1/2, f(x,y) = xy, and c = 1/4. Then H^1(s;u) = 0 for u < 1/2 and H^1(s;u) = 1/2 for u ≥ 1/2, a jump at u = 1/2. Thus H^m ∉ C([0,1],[0,1]), so the example does not satisfy (H1). The same discontinuity persists in the claimed limit H(s;u) whenever the Fleming-Viot measure Z(s) has atoms, which is typical (e.g., Dirichlet-process states are purely atomic). Consequently Theorem 3.4 cannot be applied to the paper's own Example 4.3 as written. This does not refute the theorem, but it removes the advertised demonstration of graphon dynamics beyond block models and shows that the continuity requirement in (H1) is too restrictive for natural Z-dependent fitness landscapes. In addition, the Markov-property assertion in Section 4.2 is unsupported: it requires an invertible map G with tilde h(s) = G(Y(s)), but for r ≡ 1 the map sends every state to the same complete graphon, so the asserted 'some invertible map' is false without extra assumptions on r and H.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a framework for graphon-valued stochastic processes and constructs a natural class of such processes from population genetics. After establishing a metric characterization of weak convergence in the space of graphon-valued cadlag paths in terms of subgraph densities (Theorem 3.1 and Corollary 3.2), the authors consider a Moran model with n individuals and m+1 types, forming a random graph at each time by connecting two individuals i and j with probability r(H^{m,n}(s; type_i), H^{m,n}(s; type_j)), where H^{m,n} is a dynamically evolving fitness landscape. Theorem 3.3 states that, under joint convergence of the type-frequency process and the fitness landscape process, the graph processes converge to a block graphon whose boundaries move according to the Wright-Fisher diffusion. Theorem 3.4 passes to the Fleming-Viot limit by letting the number of types tend to infinity, yielding a limiting graphon driven by the Fleming-Viot diffusion. The paper gives three examples, including a claimed non-block example based on an indicator-threshold fitness landscape, and discusses possible generalizations. The proofs use McDiarmid's concentration inequality, Skorokhod embedding, and standard tightness criteria.","tokens_in":18094,"tokens_out":6077,"duration_ms":57870,"significance":"If the results are correct, the paper provides a useful and explicit route from finite population-genetic graph dynamics to diffusive graphon limits, and its subgraph-density characterization of weak convergence for graphon-valued processes is a valuable technical tool. The main convergence theorems are proved in a self-contained way with standard concentration and tightness arguments, and no free parameters are fitted. The principal weaknesses are that the advertised non-block Fleming-Viot example does not satisfy the continuity hypothesis of Theorem 3.4 as stated, and the Markov property asserted in Section 4.2 is not established; both concern load-bearing parts of the paper's narrative.","major_comments":[{"comment":"Example 4.3 claims that the fitness landscapes H^m defined by H^m(s;u) = integral of I[f(barF^m(s;u), barF^m(s;v)) >= c] Z^m(dv) satisfy hypothesis (H1). This is false for finite m: because Z^m is atomic and barF^m is a step function, H^m is a step function in u and is not an element of L = C([0,1],[0,1]). For a concrete instance, take m=1, Y_0=Y_1=1/2, f(x,y)=xy, and c=1/4; then H^1(s;u)=0 for u<1/2 and H^1(s;u)=1/2 for u>=1/2, a jump at u=1/2. The claimed limit H(s;u) has the same discontinuity whenever the Fleming-Viot measure Z(s) has atoms, which is typical. Consequently Theorem 3.4 does not apply to Example 4.3 as written, and the statement in the example that the H^m satisfy (H1) is incorrect. This does not invalidate Theorem 3.4, but it removes the paper's advertised demonstration of graphon dynamics beyond block models and indicates that the continuity requirement in (H1) is too restrictive for the natural Z-dependent fitness landscapes considered here.","section":"Section 4.1, Example 4.3"},{"comment":"The assertion that the limiting graphon dynamics in Theorems 3.3 and 3.4 are Markov processes, based on the formula tilde h(s) = G(Y(s)) for 'some invertible map G', is not supported and is false as stated. If r is identically 1, then G sends every state Y(s) to the same complete graphon, so no invertible map exists. Even when G is not constant, the generator computation tilde L phi = L(phi o G) o G^{-1} presupposes that the graphon state determines the underlying diffusion state Y(s), and this invertibility is exactly what is not established. The paper's own Example 1.2 shows that a function of a Markov process need not be Markov under its own filtration. The Markov property should either be proved directly from the dynamics or stated under explicit additional conditions on r and H; as written, the claim overreaches.","section":"Section 4.2"}],"minor_comments":[{"comment":"The statement of Theorem 3.1(iii) quantifies over 'all k >= 1' but then uses d in the display; the intended condition is for all d >= 1.","section":"Theorem 3.1(iii)"},{"comment":"In the proof near the strong separation argument, 'f_{F_i}(h')' should be 't_{F_i}(h')' in the displayed inequalities; the same typo appears once more in the following sentence.","section":"Proof of Theorem 3.1"},{"comment":"The displayed formula for h^m(s;x,y) has awkward parentheses: 'r(m+1) barF^m(s;x), (m+1) barF^m(s;y)' should be 'r((m+1)barF^m(s;x), (m+1)barF^m(s;y))'.","section":"Example 4.1"},{"comment":"The event notation An,m is introduced as 'An,m' but later rendered as 'A_{m,n}'; the notation should be unified.","section":"Section 5.2"}],"recommendation":"major_revision","confidential_remarks":"The central convergence theorems appear sound under their hypotheses, and the proof techniques are appropriate. The main problems are local but load-bearing: Example 4.3 violates the continuity hypothesis of Theorem 3.4, and the Markov-property claim in Section 4.2 is not justified. Both are fixable in a revision, so I do not recommend rejection. There are no concerns about attribution or citation practice."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper delivers what the title says: a general method for building graphon-valued stochastic processes as scaling limits of Moran and Fleming-Viot models, and it fills a real gap left by Crane (2016) and Cerny-Klimovsky. The main theorems are the new material: Theorem 3.1 gives a practical weak-convergence criterion for graphon processes in terms of subgraph densities, and Theorems 3.3 and 3.4 construct the limits for finite-type and continuum-type population models. The proofs are careful, and I did not find a hole in the central convergence argument; the use of McDiarmid and tightness is standard and appropriate.\n\nThe soft spots are at the edges, but real. Example 4.3, which is advertised as showing the framework goes beyond block models, defines H^m(s;u) = ∫ I[f(Fbar^m(s;u),Fbar^m(s;v)) ≥ c] Z^m(dv). For finite m, Z^m is atomic and Fbar^m is a step function, so H^m is a step function, not continuous. That violates (H1), which requires H^m ∈ C([0,1],[0,1]). So Theorem 3.4 cannot be applied to the paper's own example as written. The stress-test note gets this right; it's the only place where I'd say the paper oversells. The theorem may still be true under weaker continuity, but the proof as written needs that assumption, and the example needs repair.\n\nAlso, Section 4.2 asserts the limiting graphon processes are Markov because h~ = G(Y) for some invertible map G. The invertibility is not established, and in fact fails for r ≡ 1 — the map collapses everything to the complete graphon. The formal generator computation is therefore a heuristic. This is a minor-to-moderate flaw: the Markov property may well hold for many r, but it should be either proved or labeled a conjecture.\n\nOn citation pattern and originality: no issue. The change-of-reference-measure observation is attributed to Athreya-Rollin (2016), which is appropriate. The work is not a restatement of known results.\n\nOverall: the core theorems are worth knowing and the paper deserves a serious referee. It needs revision in Example 4.3 and a softening of the Markov assertion. I would recommend a conditional accept or a major-revision decision, not a desk reject.","headline":"A solid construction of graphon-valued diffusions from population genetics, with a flawed flagship example and an unproven Markov claim; the core theorems are sound.","tokens_in":18588,"tokens_out":1681,"would_cite":true,"duration_ms":15036,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C80","60J68","60K35"],"pacs":[],"model":"deepseek-v4-flash","headline":"Multi-type Moran random graphs converge weakly to graphon-valued diffusion processes driven by Wright-Fisher and Fleming-Viot dynamics.","keywords":["graphons","graphon dynamics","Moran model","Wright-Fisher diffusion","Fleming-Viot diffusion","Skorohod topology","dense random graphs","population genetics"],"falsifier":"Simulate the $m=1$ two-type model of Example 4.1 with connection matrix $[[\\alpha,\\delta],[\\delta,\\beta]]$ and compare the empirical subgraph densities of the evolving graph at large $n$ to those of the two-block graphon whose boundary is the Wright-Fisher diffusion $Y(s)$ and whose block heights are $\\alpha$, $\\beta$, and $\\delta$; a systematic mismatch at the $n\\to\\infty$ scaling would falsify Theorem 3.3.","tokens_in":17499,"feed_emoji":"🧬","tokens_out":10861,"duration_ms":97885,"temperature":0.7,"pith_summary":"This paper constructs a class of graphon-valued stochastic processes by taking a finite population whose individuals change type by Moran resampling and, at every time, connecting each pair of individuals with a probability that depends on the current fitness values of their two types. The central claim is that, when time is rescaled by population size and the population size goes to infinity, the random graph process converges weakly in the graphon space to a limiting process whose edge weights are given by the composition of the type-connection function, the fitness landscape, and the quantile transform of the type distribution. With a fixed number of types this limit is driven by the Wright-Fisher diffusion; when the number of types also goes to infinity, the limit is driven by the Fleming-Viot diffusion. This matters because it shows that diffusion-like Markov processes on graphon space arise naturally from discrete resampling dynamics, and that such processes can lie beyond the locally-bounded-variation dynamics obtainable from infinitely exchangeable arrays.","feed_headline":"Moran-model random graphs converge to graphon-valued diffusions","feed_subtitle":"Connection probabilities come from a type-connection matrix; in the population limit they converge to a graphon diffusion.","key_machinery":"The central object is the graphon-valued process defined by composing a type-connection kernel with a fitness landscape and the quantile function of the type distribution: $h(s;x,y)=r(H(s; \\bar F(s;x)), H(s; \\bar F(s;y)))$. Here $r:[0,1]^2 \\to [0,1]$ is continuous, $H(s;\\cdot)$ is a continuous fitness landscape on the type space, and $\\bar F(s;\\cdot)$ is the right-continuous generalised inverse of the cumulative type-distribution function; this last map reparameterises the vertex space by the evolving type measure. The proof machinery is Theorem 3.1, which identifies weak convergence in the Skorohod space of graphons with convergence of subgraph-density processes $t_F$, together with McDiarmid's concentration inequality to control the random edges around their conditional mean and Ethier-Kurtz tightness criteria to upgrade finite-dimensional convergence to process convergence.","core_discovery":"The paper's discovery is the explicit scaling limit of a time-evolving dense random graph built from a multi-type Moran model. Fix $m+1$ types and $n$ individuals; at rescaled time $s$, vertices $i$ and $j$ are connected with probability $r(H^{m,n}(s; \\tau_i^{m,n}(s)), H^{m,n}(s; \\tau_j^{m,n}(s)))$, where $\\tau_i^{m,n}(s)$ is the type of individual $i$ normalised to $[0,1]$, $H^{m,n}$ is a fitness landscape process, and $r$ is a continuous type-connection kernel. Theorem 3.3 states that, under the joint convergence $(Y^{m,n},H^{m,n}) \\Rightarrow (Y^m,H^m)$, the graph process converges to the graphon process with representative $h^m(s;x,y) = r(H^m(s; \\bar F^m(s;x)), H^m(s; \\bar F^m(s;y)))$, where $\\bar F^m$ is the quantile transform of the Wright-Fisher type measure. Theorem 3.4 lets the number of types tend to infinity and obtains $h(s;x,y) = r(H(s; \\bar F(s;x)), H(s; \\bar F(s;y)))$ with the Fleming-Viot limit $Z$ in place of $Z^m$. The paper also proves a general criterion (Theorem 3.1) saying that weak convergence of graphon processes is equivalent to convergence of all finite-dimensional subgraph-density processes, and uses that criterion as the route to the scaling limits.","pith_inferences":["The same quantile-composition formula suggests a general recipe for turning any Fleming-Viot martingale problem into a graphon diffusion: choose a continuous connection kernel $r$ and a fitness landscape $H$, and define $h$ by this composition; the paper's examples show this yields non-block limits, but the full class of graphons reachable this way is not characterized.","If the Markov generator computation in Section 4.2 is made rigorous, the state-to-graphon map $G$ would give an explicit generator on graphon space for any type-diffusion whose state-to-graphon map is invertible; proving that invertibility is the natural next step.","The construction should extend to state-dependent resampling rates and to other measure-valued diffusions beyond Wright-Fisher and Fleming-Viot, provided the joint convergence assumption $(H1)$ holds; this would yield graphon dynamics whose speed is modulated by population diversity.","The formula can be tested computationally: simulate the finite Moran graph, compute its subgraph densities, and compare them to the integrals over $Z^m$ of $r(H^m(\\bar F^m(\\cdot)), H^m(\\bar F^m(\\cdot)))$; a systematic mismatch as $n$ grows would indicate a breakdown of the joint convergence assumption."],"forward_implications":["For each fixed $m$, the limiting graphon process is a moving $(m+1)\\times(m+1)$ block graphon: the boundaries of the blocks evolve by the Wright-Fisher diffusion, and the height of each block is the connection probability evaluated on the fitness landscape.","As the number of types goes to infinity, the block structure is lost in general; the Fleming-Viot limit can produce graphon diffusions not confined to the stochastic block model class.","Every finite-dimensional subgraph density of the finite graph process converges to a functional of the type-diffusion, so standard network statistics such as edge densities, triangle counts, and motif frequencies follow diffusive dynamics in the limit.","Markovianity on the graphon level does not require a Markov lift to infinitely exchangeable arrays; a non-Markov array process can project to a Markov graphon process, so graphon dynamics are more flexible than the exchangeable-array framework alone would suggest."],"supporting_citations":[{"why":"Supplies the Moran-to-Wright-Fisher and Fleming-Viot scaling limits used as the type dynamics in Theorems 3.3 and 3.4.","marker":"Dawson (1993, Section 2)"},{"why":"Provides the weak-convergence, tightness, and continuous-mapping criteria used throughout the proofs of Theorems 3.1, 3.3 and 3.4.","marker":"Ethier and Kurtz (1986)"},{"why":"Supplies the bounded-differences concentration inequality used to control fluctuations of subgraph densities around their conditional mean.","marker":"McDiarmid (1989)"},{"why":"Establishes the graphon convergence theorem that underlies the graphon metric and the law of large numbers for dense graphs.","marker":"Lovász and Szegedy (2006)"},{"why":"Supplies the density result for subgraph-count polynomials used to show that $t_F$ processes determine weak convergence on graphon space.","marker":"Diao et al. (2015)"},{"why":"Provides an alternative proof of graphon convergence via exchangeable arrays referenced in Theorem 2.1.","marker":"Diaconis and Janson (2008)"}],"fun_headline_variants":["Wright-Fisher resampling drives graphon-valued process limits","Moran model graphs converge to Fleming-Viot-driven graphons","Population limit yields graphon diffusions from type-connection kernels","From Moran to graphon: scaling limits of evolving random graphs","Large-population limit to graphon diffusions via fitness landscapes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the joint convergence of the fitness landscape and type-count processes to their Wright-Fisher or Fleming-Viot limits, together with continuity of the connection kernel and fitness landscape; if these fail, the limiting graphon formula is not justified.","fun_headline_variants_meta":{"raw":{"variants":["Wright-Fisher resampling drives graphon-valued process limits","Moran model graphs converge to Fleming-Viot-driven graphons","Population limit yields graphon diffusions from type-connection kernels","From Moran to graphon: scaling limits of evolving random graphs","Large-population limit to graphon diffusions via fitness landscapes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000211,"raw_usage":{"total_tokens":1465,"prompt_tokens":1045,"completion_tokens":420,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":661,"completion_tokens_details":{"reasoning_tokens":334}},"tokens_in":661,"tokens_out":420,"duration_ms":4327,"temperature":1.0,"reasoning_tokens":334,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:52:01.220726+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the $m=1$ two-type model of Example 4.1 with connection matrix $[[\\alpha,\\delta],[\\delta,\\beta]]$ and compare the empirical subgraph densities of the evolving graph at large $n$ to those of the two-block graphon whose boundary is the Wright-Fisher diffusion $Y(s)$ and whose block heights are $\\alpha$, $\\beta$, and $\\delta$; a systematic mismatch at the $n\\to\\infty$ scaling would falsify Theorem 3.3.","supporting_citations":[],"review_version":1}