{"id":"cce84394-921f-4491-8de2-5a488100c06c","arxiv_id":"2608.13332","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For complex balanced reaction networks with arbitrarily high polynomial growth, a suitable transport noise yields unique global classical solutions with arbitrarily high probability and enhanced exponential dissipation of spatial fluctuations.","lead":"Adding a carefully chosen random stirring to a reacting and diffusing chemical system can stop it from blowing up, giving smooth solutions that last forever with probability as close to 1 as one wants. The same stirring can also force spatial fluctuations to decay exponentially at any prescribed rate.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper's motivating ATP and combustion networks actually possess boundary equilibria with positive conservation vector, so they are not covered by the no-boundary-equilibria assumption of Theorem 3.4.","rationale":"The reader identified the no-boundary-equilibria assumption as the weakest assumption, and I agree that this condition is load-bearing. However, the reader did not notice that the paper's own motivating examples fail this assumption. This is not a critique of the proof architecture itself; the conditional theorem for networks that do satisfy (C2) may well be correct. But the exposed failure of the examples weakens the paper's main advertised scope: the abstract and Section 1.3 claim coverage of realistic reactions with arbitrary polynomial growth, whereas the ATP and combustion examples both admit explicit boundary equilibria with strictly positive conservation-law coordinates. The fix is not merely to add a remark; either the examples must be replaced by networks genuinely without boundary equilibria, or Theorem 3.4 must be reformulated with a mass-set restriction K that avoids the boundary-equilibrium mass values. Because the central proof remains conditional on a structural hypothesis that is narrower than the paper's presentation suggests, conditional acceptance with corrected examples and explicit scope restrictions is the appropriate outcome.","tokens_in":58941,"tokens_out":19969,"duration_ms":215254,"concrete_test":"Instantiate the ATP network from Section 1.3 with n=3 and the explicit conservation matrix Q above. Substitute v*=(0,a,b,c,0,d) with a,b,c,d>0, verify R(v*)=0 and that all five entries of Qv* are positive. Repeat the analogous computation for O2+2N2 ⇌ 2N+2NO with v*=(0,0,a,b). If both checks succeed, Section 1.3's examples violate condition (C2). A further useful check is to search for a genuinely high-growth complex balanced network that satisfies (C2), or to explicitly restrict the compact set K in Theorem 3.4 to initial mass vectors lying outside every boundary-equilibrium cone.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section 1.3 claims that the ATP synthesis reaction (ADP+Pi+nH_A ⇌ ATP+H2O+nH_B) and the combustion reaction (O2+2N2 ⇌ 2N+2NO) satisfy the no-boundary-equilibria hypothesis of Theorems 3.4 and 3.5. This is false. For ATP, take the conservation matrix Q with rows (1,0,0,1,0,0), (0,1,0,0,1,0), (0,0,1,0,0,1), (0,1,0,1,0,0), and (0,0,1,n,0,0), all orthogonal to the reaction vector (-1,-1,-n,1,1,n). The boundary point v*=(0,a,b,c,0,d) with positive a,b,c,d satisfies R(v*)=k1·0·a·b^n - k2·c·0·d^n = 0, so f(v*)=0, while Qv*=(c,a,b+d,a+c,b+nc) has all entries positive. Thus v* is a boundary equilibrium in a positive stoichiometric compatibility class, so condition (C2) used in Theorem 3.4 fails for every compact K containing this mass vector. For combustion, v*=(0,0,a,b) gives Qv*=(a,b,a+b)>0 for the displayed Q and both mass-action monomials vanish, again giving a boundary equilibrium. Hence the advertised high-polynomial-growth examples are outside the stated hypothesis. The theorem may still be correct for networks genuinely satisfying (C2), but the claimed applications in Section 1.3 are unsupported as written, and Remark 3.6 does not repair the examples because they do not satisfy its alternative uniform positive lower bound either.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies a stochastic reaction-diffusion system with transport noise on the torus, for nonlinearities arising from mass-action chemical reaction networks. The main result, Theorem 3.4, asserts that for a complex balanced network with no boundary equilibria, one can choose a noise intensity and finitely supported, normalized noise coefficients such that, for all admissible L^q initial data with the relevant conservation vector in a compact set, the unique (p,kappa,delta,q)-solution is global in time with probability at least 1-epsilon and is classical in space, in C_t^{1/2-}C_x^infty. Theorem 3.5 asserts that the same mechanism yields enhanced dissipation of spatial fluctuations at an arbitrarily prescribed exponential rate on a high-probability event. The proof combines uniform entropy-entropy dissipation estimates (Section 4), a close-to-equilibrium global well-posedness result (Theorem 5.1), a scaling limit for stochastic equations with cutoff (Theorem 6.3), and quantitative estimates for deterministic and stochastic convolutions (Section 7). Section 1.3 advertises the ATP synthesis and methane-combustion networks as concrete high-growth examples to which the theorems apply.","tokens_in":59251,"tokens_out":12648,"duration_ms":132030,"significance":"If Theorems 3.4 and 3.5 are correct, they constitute a substantial advance in the regularization-by-noise literature for reaction-diffusion systems with superquadratic nonlinearities, and the enhanced-dissipation statement goes beyond the passive-scalar setting. The proof architecture is coherent and contains several nontrivial ingredients of independent interest, including the uniform entropy-entropy dissipation inequality of Theorem 4.1, the theta-uniform close-to-equilibrium estimate of Theorem 5.1, and the sharp stochastic-convolution estimate of Lemma 7.8. The paper is also honest in identifying the technical role of the no-boundary-equilibria condition. However, the claimed applications in Section 1.3 are not covered by the stated hypotheses: the ATP and combustion networks possess boundary equilibria in positive stoichiometric compatibility classes. Since the abstract and informal Theorem 1.1 present the result without this caveat, the advertised scope is materially overstated. The conditional theorems may still be correct for networks that genuinely satisfy the no-boundary-equilibria hypothesis, but the motivating examples require substantial revision.","major_comments":[{"comment":"The ATP and combustion examples do not satisfy the no-boundary-equilibria assumption of Theorem 3.4. For the ATP network, take the conservation matrix Q with rows (1,0,0,1,0,0), (0,1,0,0,1,0), (0,0,1,0,0,1), (0,1,0,1,0,0), and (0,0,1,n,0,0), which are all orthogonal to the reaction vector gamma=(-1,-1,-n,1,1,n). For v*=(0,a,b,c,0,d) with a,b,c,d>0, the mass-action rate is R(v*)=k1*0*a*b^n - k2*c*0*d^n = 0, so v* is a boundary equilibrium, while Qv*=(c,a,b+d,a+c,b+nc) has all entries positive. Moreover v*+tgamma is strictly positive for small t>0 and Q(v*+tgamma)=Qv*, so the mass vector M=Qv* belongs to Q R^l_{>0}. Thus the hypothesis 'for each M in Q R^l_{>0}, there are no boundary equilibria v8 with Qv8=M' fails for every compact K containing this M. The combustion network has the same defect: with v*=(0,a,0,b), a,b>0, one has R(v*)=0 and the conservation vector (b,2a+b) is positive for the rows (2,0,0,1) and (0,2,1,1). This is not a cosmetic issue: the no-boundary condition is used through Theorem 4.1(C2) to obtain the exponential decay (6.36) in Proposition 6.5, which is essential for Corollary 6.7 and for both main theorems. Remark 3.6 does not repair the examples because no uniform positive lower bound for the corresponding deterministic high-diffusivity solutions is proved or referenced for these networks.","section":"Section 1.3, Theorem 3.4, Theorem 4.1(C2)"},{"comment":"The proof of Theorem 6.3 is presented only as a sketch and delegates the central compactness and convergence argument to [1, Theorem 6.1]. Since this scaling limit is the bridge from the cutoff stochastic equation to the deterministic equation with enhanced diffusivity, and since Theorem 3.4 directly relies on it, the manuscript should provide a complete proof of the claimed modifications or state precisely which assertions are imported from [1] and verify them in the present setting. In particular, the theorem's assumption of existence and uniqueness of weak solutions to (6.22) is not proved inside Theorem 6.3; the later appeal to Proposition 6.5 and [1, Corollary 5.5] should be made explicit in the statement. As written, a referee cannot check the validity of (6.25) without reconstructing the full argument from [1].","section":"Section 6.2, Theorem 6.3"}],"minor_comments":[{"comment":"The sentence 'single reversible reaction with disjoint species on the two sides ... therefore ... the relevant positive stoichiometric compatibility classes contain no boundary equilibria' is false in general; the ATP calculation in the major comments is a concrete counterexample. The passage should be rewritten or supported by a correct example.","section":"Section 1.3"},{"comment":"Reference [59] appears to contain typesetting artifacts: 'K. Groger' should be 'K. Gröger' and 'R. Hiinlich' should be 'R. Hünlich' (or the spelling used in the original publication).","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The central concern is the mismatch between the advertised examples and the no-boundary-equilibria hypothesis. I see no indication of duplicate publication; the reliance on [1,7,8] is appropriate and acknowledged. If the authors cannot provide valid high-growth examples satisfying condition (C2), the abstract and Section 1.3 should be revised to avoid overclaiming, while the conditional theorems themselves may remain sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The central conditional theorem is substantial: for complex balanced networks satisfying the no-boundary-equilibria condition, transport noise upgrades finite-time delayed blow-up to global classical solutions with high probability, and gives enhanced dissipation of spatial fluctuations at an arbitrarily prescribed rate. The machinery is coherent: uniform entropy-entropy dissipation over compact mass sets, close-to-equilibrium global well-posedness, and the high-diffusivity deterministic limit are assembled into a real proof of the half-line result. Credit where due: the entropy-based passage from finite to infinite horizon is a genuine new step, and Lemma 4.2 / Theorem 4.1 is carefully done.\n\nBut there are two soft spots, one serious. First, Section 1.3's advertised examples do not satisfy the paper's own hypothesis. For ATP, the boundary point (0,a,b,c,0,d) with positive a,b,c,d has zero reaction rate and positive conserved values; for combustion, (0,0,a,b) is likewise a boundary equilibrium in a positive stoichiometric compatibility class. So the claim that Theorems 3.4 and 3.5 apply to ATP synthesis and methane combustion is unsupported. This is not a minor typo: it removes the two concrete high-growth applications from the scope, and Remark 3.6 does not rescue them because no uniform positive lower bound is established for those systems. Second, Theorem 6.3, the scaling limit that carries much of the weight, is only sketched and delegates the core argument to [1, Theorem 6.1]. The surrounding estimates look sound in Sections 4 and 5, and I do not see a circular dependency: the reliance on [1,7,8] is on published work, and the new contribution is the entropy upgrade to infinite time.\n\nThe reader's conditional verdict is fair. I would take the main theorem seriously as a theorem about networks that genuinely satisfy (C2), but the paper currently over-advertises. For a stochastic PDE or chemical network audience, this is worth reading and refereeing, not desk-rejecting. Recommendation: send it to referees, and insist that the authors either repair Section 1.3 with valid examples or explicitly state that the advertised networks are outside the theorem's assumptions, and ask for a self-contained or fully verified proof of Theorem 6.3. With those fixes, I would expect the result to stand.","headline":"The global-in-time result is real under its no-boundary-equilibria hypothesis, but the ATP and combustion examples advertised in Section 1.3 violate that hypothesis, so the application section needs repair.","tokens_in":59816,"tokens_out":3940,"would_cite":true,"duration_ms":40824,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60H50","60H15","35K57","35B65"],"pacs":[],"model":"deepseek-v4-flash","headline":"Transport noise forces global classical solutions for complex balanced reaction–diffusion networks with arbitrary polynomial growth, with probability arbitrarily close to 1.","keywords":["regularization by noise","global well-posedness","enhanced dissipation","stochastic reaction–diffusion equations","complex balanced reaction networks","transport noise","entropy dissipation","maximal Lp(Lq)-regularity"],"falsifier":"For a concrete complex balanced network without boundary equilibria, evaluate the infimum of $D(v)/E(v\\mid v_8)$ over smooth positive concentration fields $v$ on the torus with conserved mass vector $Q\\bar v = M$ and entropy $E(v\\mid v_8) \\le N$, with $M$ ranging over the compact set $K$. If this infimum is $0$ for some $M\\in K$, the uniform entropy–entropy dissipation inequality used in Proposition 6.5 is false, and the exponential convergence step on which Theorem 3.4 rests would collapse.","tokens_in":58688,"feed_emoji":"🌀","tokens_out":10094,"duration_ms":95144,"temperature":0.7,"pith_summary":"The paper claims that a specific physically motivated random stirring—transport noise—can prevent blow-up in reaction–diffusion systems for which global smooth solutions are not known in the deterministic setting. For any complex balanced chemical reaction network without boundary equilibria, and for reactions with polynomial growth of any order, one can choose the noise intensity and finitely supported Fourier-mode coefficients so that, from every admissible nonnegative initial concentration, the unique solution exists for all time and is classical in space with probability as close to 1 as desired. The same construction makes spatial fluctuations decay to the instantaneous spatial average at any prescribed exponential rate, with high probability and finite moments of the random prefactor. If the proof is right, turbulent transport is not only a modeling device but a mechanism that regularizes a class of chemical kinetics whose deterministic global regularity is open.","feed_headline":"Stirring random flows stop blow-up in reaction–diffusion systems","feed_subtitle":"Tuned transport noise yields global classical solutions with probability as close to 1 as desired.","key_machinery":"The load-bearing identity is the Itô–Stratonovich correction (2.12): under the divergence-free Fourier vector fields $\\sigma_{k,\\alpha}$, the Stratonovich transport noise is exactly equivalent to an Itô noise plus an extra dissipative term $\\nu\\Delta v_i$, with the orthogonality identity (2.13) giving the explicit coefficient. This converts random stirring into a quantitative diffusion enhancement. Around that sits the entropy–entropy dissipation machinery: the relative entropy $E(v\\mid v_8)$ and its dissipation $D(v)$, with the inequality $D(v) \\ge \\lambda E(v\\mid v_8)$ made uniform over compact mass sets by the no-boundary-equilibria condition. The proof then runs a scaling-limit argument: stochastic RDEs with a cutoff converge, over finite intervals, to deterministic RDEs with increased diffusivity; those deterministic equations are globally well-posed and exponentially convergent by the entropy inequality; and a close-to-equilibrium stochastic stability theorem (Moser-type iteration plus a spectral gap for the linearized operator) upgrades finite-time control to global control with high probability.","core_discovery":"On its own terms the paper proves Theorem 3.4: fix a complex balanced reaction network, assume no stoichiometric compatibility class (set of states with a fixed conserved mass vector) contains a boundary equilibrium, fix $N\\ge 1$, $\\varepsilon\\in(0,1)$, and a compact set $K$ of conserved mass vectors. Then there exist a noise intensity $\\nu>0$ and finitely supported, radially symmetric coefficients $\\theta$ with the following property: for every nonnegative initial datum with $L^q$ norm at most $N$ and mass vector in $K$, the unique $(p,\\kappa,\\delta,q)$-solution of the stochastic reaction–diffusion system is global with $P(\\tau=\\infty)>1-\\varepsilon$ and has paths in $C^{1/2-,\\infty}_{\\mathrm{loc}}((0,\\tau)\\times\\mathbb{T}^d;\\mathbb{R}^\\ell)$, so it is classical in space. This covers networks of arbitrary polynomial growth $h$, for which deterministic global well-posedness is unavailable. The companion Theorem 3.5 asserts that the same noise can enhance dissipation of the spatial fluctuations to any prescribed exponential rate, in the sense that $\\|v(t,\\cdot)-\\bar v(t,\\cdot)\\|_{L^2}\\le D e^{-\\chi t}\\|v_0\\|_{L^2}$ on a set of probability $>1-\\varepsilon$, with $\\mathbb{E}[D^b]<\\infty$.","pith_inferences":["One extension the paper leaves implicit is that any mechanism guaranteeing uniform positivity of the deterministic high-diffusivity solution, not only the no-boundary-equilibria condition, should unlock the same global result; Remark 3.6 already identifies the sufficient condition.","Because the noise coefficients are finitely supported, the construction is directly testable in simulation: run the stochastic RDE with the Fourier-mode noise (6.26) and check that the survival probability and the exponential homogenization estimate hold at the predicted rates.","The enhanced-dissipation theorem concerns decay toward the spatial mean, not toward the reaction equilibrium; for stirred combustion models this suggests that turbulent mixing can suppress spatial hot spots at essentially arbitrary rates while the mean composition still follows the deterministic reaction kinetics.","A neighbouring conjecture that might be approachable by the same tools is the Global Attractor Conjecture for complex balanced networks with boundary equilibria: the entropy method needs only a positive lower bound on the dissipation ratio, so stochastic transport could quantify convergence in cases where the deterministic dynamics remain unsettled."],"forward_implications":["For concrete networks such as ATP synthesis with $n\\ge 3$ protons or the combustion-type single reversible reaction $\\mathrm{O}_2 + 2\\mathrm{N}_2 \\rightleftharpoons 2\\mathrm{N} + 2\\mathrm{NO}$, global classical solutions follow with high probability even though deterministic global well-posedness is open.","The constructed transport noise leaves conservation laws and $L^q$ energy estimates unchanged, so the gain is purely in controlling blow-up and homogenization, not in adding mass dissipation.","Alongside global existence, the same noise yields quantitative enhanced dissipation: spatial fluctuations decay at any prescribed rate $\\chi$ with high probability, which is stronger than the pure-diffusion decay rate when $\\chi$ exceeds the diffusion eigenvalues.","For networks with boundary equilibria, the proof still works whenever the deterministic high-diffusivity system stays uniformly bounded away from zero, so the boundary-equilibria obstruction is removable in that case."],"supporting_citations":[{"why":"Supplies the entropy–entropy dissipation inequality and exponential convergence to equilibrium for complex balanced networks without boundary equilibria, which Theorem 4.1 makes uniform over compact mass sets.","marker":"[46]"},{"why":"Derives the entropy dissipation functional $D(v)$ in the form used to establish the uniform entropy–entropy dissipation estimate.","marker":"[37]"},{"why":"Provides the finite-time scaling-limit machinery for reaction–diffusion systems with transport noise in $L^p(L^q)$ spaces, upgraded here to the global horizon.","marker":"[1]"},{"why":"Gives the quantitative convergence rates for the scaling limit of SPDEs with transport noise used for the enhanced-dissipation estimates.","marker":"[52]"},{"why":"Supplies the quantitative stochastic enhanced-dissipation estimates, including exponential decay of fluctuations and moment bounds, adapted to the nonlinear reaction setting.","marker":"[76]"},{"why":"Provides the spectral-gap close-to-equilibrium stability theory for deterministic reaction–diffusion systems, extended here to stochastic RDEs with bounded noise.","marker":"[98]"},{"why":"Establishes local well-posedness, positivity, and instantaneous spatial regularity for the stochastic RDEs with transport noise, giving the solution concept that Theorem 3.4 upgrades to global.","marker":"[7]"},{"why":"Provides the stochastic maximal $L^p(L^q)$-regularity estimates used in the cutoff and scaling-limit arguments.","marker":"[9]"}],"fun_headline_variants":["Transport noise stops blow-up in reaction-diffusion","Random stirring yields global classical solutions","Noise tames blow-up for reaction networks","Stirring noise prevents singularities in reaction systems"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that no boundary equilibrium—a state where at least one species has zero concentration—sits in any stoichiometric compatibility class within the allowed range of conserved masses. This is what forces the entropy–entropy dissipation constant to be strictly positive and drives the deterministic high-diffusivity system exponentially to equilibrium; if such a boundary equilibrium exists, the proof only goes through when the deterministic solution stays uniformly bounded away from zero.","fun_headline_variants_meta":{"raw":{"variants":["Transport noise stops blow-up in reaction-diffusion","Random stirring yields global classical solutions","Noise tames blow-up for reaction networks","Stirring noise prevents singularities in reaction systems"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00064,"raw_usage":{"total_tokens":2962,"prompt_tokens":974,"completion_tokens":1988,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":590,"completion_tokens_details":{"reasoning_tokens":1930}},"tokens_in":590,"tokens_out":1988,"duration_ms":17664,"temperature":1.0,"reasoning_tokens":1930,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:31:24.335362+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a concrete complex balanced network without boundary equilibria, evaluate the infimum of $D(v)/E(v\\mid v_8)$ over smooth positive concentration fields $v$ on the torus with conserved mass vector $Q\\bar v = M$ and entropy $E(v\\mid v_8) \\le N$, with $M$ ranging over the compact set $K$. If this infimum is $0$ for some $M\\in K$, the uniform entropy–entropy dissipation inequality used in Proposition 6.5 is false, and the exponential convergence step on which Theorem 3.4 rests would collapse.","supporting_citations":[{"cited_title":"Fellner and B.Q","cited_arxiv_id":null,"evidence_quote":"Supplies the entropy–entropy dissipation inequality and exponential convergence to equilibrium for complex balanced networks without boundary equilibria, which Theorem 4.1 makes uniform over compact mass sets."},{"cited_title":"Desvillettes, K","cited_arxiv_id":null,"evidence_quote":"Derives the entropy dissipation functional $D(v)$ in the form used to establish the uniform entropy–entropy dissipation estimate."},{"cited_title":"Flandoli, L","cited_arxiv_id":null,"evidence_quote":"Gives the quantitative convergence rates for the scaling limit of SPDEs with transport noise used for the enhanced-dissipation estimates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the quantitative stochastic enhanced-dissipation estimates, including exponential decay of fluctuations and moment bounds, adapted to the nonlinear reaction setting."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the spectral-gap close-to-equilibrium stability theory for deterministic reaction–diffusion systems, extended here to stochastic RDEs with bounded noise."}],"review_version":1}