{"id":"41496568-1aa2-4f32-8e6f-3d3d900d75fa","arxiv_id":"2509.02747","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For the interchange-and-contact process on Z^d, the critical infection rate lambda_c(v,p) tends to 1/(2dp) as the interchange rate v tends to infinity, for every fixed particle density p.","lead":"This paper introduces a model of epidemics among moving particles that swap positions at rate v, and proves that as v grows the infection threshold approaches 1/(2dp), where p is the particle density. A smart generalist should read it because it establishes the exact fast-motion limit for a new three-state interacting particle system and connects mean-field arguments with vacancy thinning.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Extinction side rests on Lemma B.1, an under-specified external decorrelation estimate; the asserted d≥1 extension is unproved, and the statement's unused δ needs clarification.","rationale":"The paper's main theorem is well structured and the two-sided renormalization is detailed. The survival side has a self-contained microscopic coupling; its equilibrium-thinning argument does not hinge on Lemma B.1 but on the independence of the target particle's trajectory from the containment flow, which is valid. The extinction side, however, depends on Lemma 3.2, and Lemma 3.2's proof is the only place where an external, non-obvious decorrelation fact is imported. Because the threshold value 1/(2dp) is exactly the thinning prediction, an error here would propagate to the main theorem. The unused δ in Lemma B.1 and the asserted 'easy' extension are presentation red flags, but they are checkable; numerical simulation can quickly confirm or refute the core decorrelation. I also noticed that Prop 5.1's proof applies Lemma 2.11 to an initial set A that may exceed the lemma's size restriction, but Prop 5.6 only needs the exact-size case and a subset reduction repairs this, so it is not the main concern. Overall, the argument is convincing modulo the external Lemma B.1; making it conditional on a verified proof/statement is the appropriate stress-test outcome.","tokens_in":60025,"tokens_out":38762,"duration_ms":357632,"concrete_test":"Two-step check: (1) Locate the exact statement proved in De Masi–Presutti, Section 6.7, and verify it gives the uniform total-variation bound in Lemma B.1 for all d≥1 (including the role of δ). (2) Numerically estimate the total-variation distance between the joint law of (Φ(x,0,t), Φ(y,0,t)) and the product of its marginals in the rate-1 interchange process on a large torus for d=1 and d=3, initial separations |x-y|=1,2,...,50, and t=10^2,...,10^5; if the sup over x≠y does not decay to 0, Lemmas B.1 and 3.2 fail, and the extinction proof has no valid thinning estimate.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central extinction argument reduces to Lemma 3.2 (up-and-down), whose proof in Appendix B uses Lemma B.1 to replace the joint law of two interchange-flow trajectories by the product of their marginals in total variation, uniformly over distinct starting points. Lemma B.1 is quoted from [18, Sec. 6.7] for d=1, with only the comment that the extension to d≥1 is easy; as printed it also contains an unused parameter δ>0, so the exact external statement to be cited is unclear. This is load-bearing: Lemma 3.3 uses Lemma 3.2 to conclude that a transmission attempt targets an occupied site with probability at most p1, yielding the effective rate 2dλp1 in the birth-and-death bound; Proposition 3.1, the bottom of the extinction renormalization, depends on it. If decorrelation fails or is not uniform in the starting points (the proof needs all pairs w,z inside a ball of radius ~√v log^4 v, including close pairs), the local thinning could be biased away from p, and the claimed threshold 1/(2dp) would not follow from the given argument. The paper provides no independent proof of Lemma B.1 or of the uniformity/dimension claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines the \"interchange-and-contact process\" on Z^d: each site is empty, occupied by a healthy particle, or occupied by an infected particle; infections spread at rate λ to healthy neighbours and particles swap states across edges at rate v. Starting from a single infection and with all other sites independently healthy with probability p or empty otherwise, the authors define λ_c(v,p) and prove Theorem 1.1, which states that lim_{v→∞} λ_c(v,p) = 1/(2dp). The proof is split into an extinction regime (2dpλ<1) and a survival regime (2dpλ>1). The extinction argument combines a microscopic \"up-and-down\" lemma (Lemma 3.2, proved in Appendix B) with a half-crossing renormalization; the survival argument couples the process to a branching random walk and then runs a multi-scale renormalization with a new deterministic-initial-state decoupling. The paper is dense and the presentation is generally rigorous, but a central microscopic estimate on the extinction side depends on an external decorrelation result whose stated form and application are not fully justified.","tokens_in":60255,"tokens_out":13130,"duration_ms":127025,"significance":"If valid, Theorem 1.1 is a strong and natural result: it gives the exact mean-field critical value 1/(2dp) for epidemics among fast-moving particles, unifying the effective threshold of the fast-stirring contact process with the thinning effect present in the contact process on dynamical percolation. The techniques are also of independent interest: the survival side introduces a careful coupling with a branching random walk, and the refined decoupling for deterministic initial configurations (Lemma 2.7) is a useful improvement over earlier stochastic-domination tools. The paper is largely self-contained and the renormalization arguments are written in detail, with explicit constants and error bounds; these are genuine strengths. However, the extinction side relies at a load-bearing point on Lemma B.1, a decorrelation statement quoted from the literature whose higher-dimensional form is merely asserted and whose application appears to require more than the cited result provides.","major_comments":[{"comment":"The proof of Lemma 3.2 derives (111) from (110) by applying Lemma B.1 at the intermediate time t = T - v^{-3/4}. This is not justified: Lemma B.1 only asserts total-variation decay as t→∞, while the allowed range of T in the statement is [v^{-1/2}, log v]. For T near v^{-1/2} (which is a range of positive probability in Lemma 3.3, since σ is exponential with rate up to order log^3 v), the time t is vanishing, not tending to infinity, so the decorrelation estimate cannot be applied uniformly over the claimed domain. This invalidates inequality (36) in Lemma 3.3, which is the step that produces the effective birth rate 2dλ p_1. In turn, Lemma 3.3 is used in Lemma 3.4 and Proposition 3.1, the bottom of the extinction renormalization. The authors should either prove a quantitative two-particle decorrelation bound that is uniform in t over the relevant range, or restructure the proof of Lemma 3.2 to treat small T separately (for example via local CLT estimates together with the density assumption (31)). Additionally, Lemma B.1 is cited from [18] for d=1 with the statement that the extension to d≥1 is easy, but no argument is supplied; since the manuscript's main theorem is for general d, this is a genuine gap. The unused parameter δ>0 in the statement of Lemma B.1 further obscures exactly which quoted assertion is being used.","section":"Appendix B, proof of Lemma 3.2"}],"minor_comments":[{"comment":"The parameter δ>0 appearing in the statement of Lemma B.1 plays no role in the displayed formula; please remove it or state the stronger version that is actually needed.","section":"Appendix B, Lemma B.1"},{"comment":"There is a typo in the phrase \"times t,T > 0 with t≤T,and and parameter p∈ [0,1]\": the duplicated \"and\" should be removed.","section":"Section 2.2, Lemma 2.7"},{"comment":"\"straightfoward\" should be \"straightforward\".","section":"Section 2.2.1"},{"comment":"In the bullet point beginning \"iff(j−1) =A\", the text reads \"(i,j ), so (i,j) gains the property\" but the intended clause is \"(i,j) is (N−1)-good, so it gains the property\"; please correct this omission.","section":"Section 6.1.3, proof of Lemma 6.4"},{"comment":"The assertion that the target vertex y at a transmission time t' is in equilibrium because its past trajectory avoids the containment flow is correct, but the sentence would benefit from a brief justification that membership in Ψ^A_s along the trajectory would force the endpoint y at time t' to lie in Ψ^A_{t'}.","section":"Section 5.3, paragraph after Definition 2.12"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is technically impressive and the main theorem is very likely correct, but the extinction side currently rests on a decoration estimate (Lemma B.1) that is not proved for d≥1 and is applied at time scales where the stated t→∞ version does not apply. I would recommend major revision with a request for a fully self-contained proof of Lemma 3.2, or a clearly stated quantitative substitute. If the authors can supply that, the paper would be a strong candidate for acceptance. I also note that the proof of Lemma 3.2 is the only place where the argument appears not to be self-contained; the rest of the renormalization and coupling arguments seem sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The headline result is a sharp limit for a new three-state epidemic model: on Z^d, with particles moving by interchange at rate v and infection rate λ, the critical threshold converges to 1/(2dp) as v→∞. That's exactly what you'd expect from combining the stirring limit 1/(2d) with the thinning factor 1/p from vacancies, and the paper makes it a theorem with a genuine two-sided proof. This is a substantive first result, not a routine extension.\n\nCredit where due: the model is natural, the theorem is what you'd want, and the proof is substantial. The extinction side uses a subcritical comparison plus a half-crossing renormalization; the survival side couples to a branching random walk and uses a refined sprinkling decoupling to handle deterministic initial configurations. The paper is long and heavy, but the main steps are in the text or appendices.\n\nSoft spots: the extinction side rests on Lemma 3.2, whose proof in Appendix B uses Lemma B.1. That lemma is quoted from De Masi–Presutti for d=1, with the line \"extension to d≥1 is easy,\" and as printed it contains a δ that does nothing. It's load-bearing because replacing the joint law of two interchange trajectories by a product of marginals is what forces the effective rate to be 2dλp. The uniformity over starting points is also important. I don't think this is an error—the statement is plausible and the scaling works (the intermediate time t satisfies vt→∞ under the hypotheses)—but it's under-specified. A referee should ask for a precise reference or a short proof, and the δ should be cleaned up.\n\nThere are also a few omitted proofs, e.g., Lemma 2.12, but they are marked as analogous to earlier results.\n\nOverall: this paper deserves a serious referee. The theorem is significant and the argument is serious. The main fix is to pin down Lemma B.1. I'd accept with revisions.","headline":"Sharp limit for a new epidemic model with moving particles and vacancies; main proof is solid but leans on an under-specified external lemma.","tokens_in":60779,"tokens_out":7992,"would_cite":true,"duration_ms":81976,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60K35","60J80","82C22"],"pacs":[],"model":"deepseek-v4-flash","headline":"Fast particle swaps set the epidemic threshold to 1/(2dp)","keywords":["interchange process","contact process","interchange-and-contact process","critical threshold","mean-field limit","fast stirring","renormalization","branching random walk"],"falsifier":"Simulate two interchange-flow trajectories started at neighboring sites on a large finite box in d=3 at rate v=1, and measure the total-variation distance between their joint law and the product of their marginals as time T grows; if the distance does not decay to zero, Lemma B.1 fails and the paper's microscopic estimates, hence its proof of the theorem, would lose their grounding.","tokens_in":59839,"feed_emoji":"🦠","tokens_out":4709,"duration_ms":44130,"temperature":0.7,"pith_summary":"This paper defines the interchange-and-contact process, an epidemic model on Z^d in which infected particles recover at rate 1, infect healthy neighbors at rate λ, and particles of any state swap along edges at rate v. Starting from one infected particle and otherwise independent healthy or vacant sites with density p, it studies the critical infection threshold λ_c(v,p). The main claim is that as v→∞, λ_c(v,p) converges to 1/(2dp): the same mean-field threshold as a contact process in which every transmission succeeds, discounted by the probability p that the target site is actually occupied. If true, the result says that extremely fast mixing makes the environment self-average completely, so the epidemic's critical point is governed by the product of infection rate, degree, and particle density.","feed_headline":"Fast particle swaps set epidemic threshold to 1/(2dp)","feed_subtitle":"Above the critical value, the infection survives; below it, it dies out, for any fixed particle density p.","key_machinery":"The argument is carried by three interlocking devices. First, the interchange flow Φ(x,s,t) tracks individual particles, and the paper uses it to define infection paths and a containment flow that bounds the spread of the infection. Second, an up-and-down lemma shows that, under a uniform local density bound, the particle reached by a transmission is nearly in equilibrium, which yields a branching-random-walk approximation with birth rate 2dλp. Third, renormalization schemes on space-time boxes—with horizontal decoupling coming from large-deviation bounds on random walks and vertical decoupling from a refined stochastic-domination coupling—show that bad boxes are exponentially rare across scales. The branching-random-walk approximation is anchored by a large-deviation propagation result for branching Brownian motion, obtained as the diffusive scaling limit of the branching random walk.","core_discovery":"On the paper's own terms, the discovery is a theorem: for every p∈(0,1] and every d≥1, lim_{v→∞} λ_c(v,p)=1/(2dp). Equivalently, whenever 2dpλ<1 the infection dies out almost surely for all sufficiently large v, and whenever 2dpλ>1 it survives with positive probability for all sufficiently large v. The proof derives this by approximating the set of infected particles by a branching random walk with death rate 1 and birth rate 2dλp, valid while infections are sparse; extinction and survival are then established through separate renormalization arguments that control rare spatial and temporal fluctuations in particle density. A key structural ingredient is that when the interchange rate is large, each transmission attempt sees the target location in near-equilibrium, so the only effect of the moving environment is to thin transmissions by the occupation probability p.","pith_inferences":["A natural extension not addressed in the paper is the rate of convergence: by analogy with existing refinements for the contact process with stirring, one may expect corrections of order 1/v in d≥3, (log v)/v in d=2, and v^{-1/3} in d=1.","The proof's reliance on asymptotic independence of interchange trajectories suggests a direct numerical check in d=3 could decide whether that decorrelation holds beyond the dimension in which the cited source proves it.","A testable prediction of the theorem is that on large finite tori with fast swapping, the critical threshold should display finite-size scaling around 1/(2dp), with the scaling window governed by the time the interchange process needs to mix the environment."],"forward_implications":["For fixed p, if 2dpλ<1, the infection dies out almost surely for all sufficiently large v.","For fixed p, if 2dpλ>1, the infection survives with positive probability for all sufficiently large v.","Taking p=1 recovers the known fast-stirring threshold 1/(2d) for the contact process with stirring.","The threshold depends on the environment only through the particle density p, so highly mobile populations behave like a mean-field epidemic with transmissions thinned by occupancy."],"supporting_citations":[{"why":"Establishes the fast-stirring contact process threshold 1/(2d), the result this paper extends to vacant sites.","marker":"[20]"},{"why":"Gives the contact-process-on-dynamic-percolation threshold 1/(p λ_c^CP), supplying the thinning phenomenon the paper reproduces with particle motion.","marker":"[38]"},{"why":"Supplies the sprinkling and decoupling method whose refined deterministic-initialization version is used for vertical decoupling.","marker":"[3]"},{"why":"Provides the space-time half-crossing renormalization scheme reused here for the extinction side.","marker":"[23]"},{"why":"Source for Lemma B.1, the asymptotic independence of interchange flow trajectories underlying the up-and-down lemma.","marker":"[18]"},{"why":"Supplies the branching random walk propagation estimates, used through the branching Brownian motion scaling limit in the survival analysis.","marker":"[9]"},{"why":"Develops a parallel refinement allowing decoupling couplings from deterministic initial configurations, cited as a closely related technique.","marker":"[14]"}],"fun_headline_variants":["Fast swaps drive epidemic threshold to 1/(2dp)","Moving particles: infection survives iff 2dpλ>1","High interchange rates set critical infection rate to 1/(2dp)","For swift particle shuffling, infection threshold becomes 1/(2dp)"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that two particles moving under the interchange flow become statistically independent as time passes, even in dimensions d≥2; the paper relies on this decorrelation to justify the equilibrium thinning factor p, citing it from an earlier source where the proof is written for d=1 and asserted to extend easily.","fun_headline_variants_meta":{"raw":{"variants":["Fast swaps drive epidemic threshold to 1/(2dp)","Moving particles: infection survives iff 2dpλ>1","High interchange rates set critical infection rate to 1/(2dp)","For swift particle shuffling, infection threshold becomes 1/(2dp)"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000641,"raw_usage":{"total_tokens":2921,"prompt_tokens":886,"completion_tokens":2035,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":502,"completion_tokens_details":{"reasoning_tokens":1962}},"tokens_in":502,"tokens_out":2035,"duration_ms":14247,"temperature":1.0,"reasoning_tokens":1962,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:33:43.320586+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate two interchange-flow trajectories started at neighboring sites on a large finite box in d=3 at rate v=1, and measure the total-variation distance between their joint law and the product of their marginals as time T grows; if the distance does not decay to zero, Lemma B.1 fails and the paper's microscopic estimates, hence its proof of the theorem, would lose their grounding.","supporting_citations":[{"cited_title":"Particle systems and reaction-diffusion equations.The Annals of Pro- bability, pages 289–333, 1994","cited_arxiv_id":null,"evidence_quote":"Establishes the fast-stirring contact process threshold 1/(2d), the result this paper extends to vacant sites."},{"cited_title":"The contact process with dynamic edges onZ","cited_arxiv_id":null,"evidence_quote":"Gives the contact-process-on-dynamic-percolation threshold 1/(p λ_c^CP), supplying the thinning phenomenon the paper reproduces with particle motion."},{"cited_title":"How can a clairvoyant particle escape the exclusion process?Ann","cited_arxiv_id":null,"evidence_quote":"Supplies the sprinkling and decoupling method whose refined deterministic-initialization version is used for vertical decoupling."},{"cited_title":"Results on the contact process with dynamic edges or under renewals.Electronic Journal of Probability , 27:1–31, 2022","cited_arxiv_id":null,"evidence_quote":"Provides the space-time half-crossing renormalization scheme reused here for the extinction side."},{"cited_title":"Springer Verlag, 1991","cited_arxiv_id":null,"evidence_quote":"Source for Lemma B.1, the asymptotic independence of interchange flow trajectories underlying the up-and-down lemma."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the branching random walk propagation estimates, used through the branching Brownian motion scaling limit in the survival analysis."}],"review_version":2}