{"id":"a1f4c949-0e90-4862-870e-b8593ae59e6d","arxiv_id":"1908.10015","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Under a dissipativity condition, a stochastic differential equation with quasi-periodic coefficients has a unique random quasi-periodic path, a unique quasi-periodic measure, and a unique ergodic invariant measure on a lifted cylinder.","lead":"This math paper defines random quasi-periodic paths and quasi-periodic probability measures for stochastic differential equations, and proves when a dissipative SDE has exactly one of each. It also shows the corresponding lifted system on a cylinder has a unique, ergodic invariant measure, a building block for modeling phenomena with two independent rhythms such as daily and seasonal temperature cycles.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.16's uniqueness proof uses a contraction estimate with a constant independent of x, which is false; the resulting lower bound and Fatou step are unjustified.","rationale":"The reader's weakest_assumption concerns the existence of a well-behaved periodic lift in Conditions 3.2-3.3. That is a legitimate scope concern, but it is an explicit hypothesis of the theorem rather than an internal inconsistency: quasi-periodicity alone is not claimed to imply the lift conditions. The proof gap I identified is more load-bearing because it directly affects the stated central claim of a unique, ergodic invariant measure on the cylinder. The uniform-in-x contraction estimate used to prove uniqueness of the invariant measure is demonstrably false for the paper's own O-U example. The conclusion may still be repairable by adding a Lyapunov integrability argument for arbitrary invariant measures, and for this reason I do not advocate rejecting the paper; the appropriate verdict remains CONDITIONAL, the same as the reader's. I marked agreement_with_reader as 'disagree' because the reader's weakest_assumption is not the concern I find most load-bearing, even though the reader's rationale independently points to a different omitted-proof issue (Lemma 4.4).","tokens_in":33450,"tokens_out":33444,"duration_ms":345855,"concrete_test":"Specialize Section 3.11 to the Ornstein-Uhlenbeck case S=0, sigma=I, A=aI with r1=r2=0. The explicit solution gives E|K^{0,0}(0,-T,x)-tilde_phi^{0,0}(0)|^2 = e^{-2aT}|x|^2. For any fixed T and any proposed uniform constant C, choose |x|>C to contradict the claimed bound ||K-tilde_phi||_2 <= C e^{-cT}. This check settles whether the estimate in Theorem 3.16's proof is valid; alternatively, if the proof is repaired by adding a Lyapunov argument showing every invariant measure has finite second moment, verify that the revised Fatou step is explicitly written out.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Theorem 3.16 (page 28), the authors state that K^{r1,r2}(t,s,x) satisfies ||K^{r1,r2}(t,s,x)-tilde_phi^{r1,r2}(t)||_2 <= C exp(-(alpha-beta^2/2)(t-s)) with C = C(alpha,beta,tilde_M) independent of x. This is not true for general dissipative SDEs. In the O-U example of Section 3.11, take S=0, sigma=I, A=aI, and r1=r2=0. Then K^{0,0}(0,-T,x) = e^{-aT}x + integral_{-T}^0 e^{av} dW_v, while tilde_phi^{0,0}(0) = integral_{-infinity}^0 e^{av} dW_v, so the L2 distance is exactly e^{-aT}|x|. No constant independent of x can bound this. Consequently, the error term in the lower bound (3.58) is actually C(1+|x|^2)/delta^2 times an exponentially decaying factor, and passing to the limit inside the integral in (3.59) requires integrability of |x|^2 with respect to the unknown invariant measure nu. The paper does not establish this integrability before using Fatou's lemma; it can be obtained from the dissipativity Lyapunov inequality, but that argument is absent. Thus the proof of uniqueness of the invariant measure is incomplete as written.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces notions of random quasi-periodic paths for random dynamical systems and quasi-periodic measures for Markovian semigroups, and applies them to SDEs with quasi-periodic coefficients. Assuming a two-variable periodic lift of the coefficients satisfying uniform one-sided dissipativity, Lipschitz, boundedness and Hölder conditions, it proves existence and uniqueness of uniformly L^p-bounded random quasi-periodic paths and quasi-periodic entrance measures in M_p. It then lifts the flow to the cylinder [0,τ1)×[0,τ2)×R^d, proves tightness of the Cesàro averages and the Feller property, and obtains an invariant measure. The paper further claims uniqueness and ergodicity of this invariant measure, and gives sufficient conditions for the existence of densities and for the Fokker-Planck equation. A multidimensional Ornstein-Uhlenbeck example with explicit quasi-periodic path and measure is worked out.","tokens_in":33645,"tokens_out":10504,"duration_ms":108321,"significance":"If the main theorems are correct, this is a valuable first systematic framework for random quasi-periodicity, parallel to the existing theory of random periodic paths and periodic measures. The explicit dissipativity and regularity assumptions make the existence results checkable, and the OU example gives concrete formulas that help validate the definitions. The paper also contains useful technical components: L^p contraction estimates, Kolmogorov continuity arguments, tightness, and the Feller property. The claimed uniqueness and ergodicity of the lifted invariant measure is the capstone of the paper; since the proof of that claim is currently incomplete, the significance is conditional on a repair. The main theorems also depend on the strong premise, stated in Conditions 3.2 and 3.3, that the quasi-periodic coefficients admit a two-variable periodic lift satisfying uniform dissipativity; the paper does not discuss how to verify this beyond the linear example.","major_comments":[{"comment":"The estimate displayed before (3.58), asserting that ||K^{r1,r2}(t,s,x)-tilde_phi^{r1,r2}(t)||_2 is bounded by C exp(-(alpha-beta^2/2)(t-s)) with C=C(alpha,beta,tilde_M) independent of x, is false. In the Ornstein-Uhlenbeck example with S=0, sigma=I, A=aI and r1=r2=0, one has K^{0,0}(0,-T,x)=e^{-aT}x+int_{-T}^0 e^{av} dW_v while tilde_phi^{0,0}(0)=int_{-infty}^0 e^{av} dW_v, so the L^2 distance equals e^{-aT}|x|. Consequently the term subtracted in (3.58) must carry a factor depending on |x|, typically 1+|x|^2, and passing to the limit inside the integral in (3.59) via Fatou's lemma requires a second-moment bound on the arbitrary invariant measure nu. The paper does not establish int |x|^2 nu(dx)<infty before using Fatou's lemma. The uniqueness conclusion of Theorem 3.16 is therefore not supported as written. This gap appears fixable by adding a Lyapunov/dissipativity argument for the invariant measure and by tracking the x-dependence in the contraction estimate, but the missing argument is load-bearing.","section":"Theorem 3.16, proof, equations (3.58)-(3.59)"},{"comment":"The proof invokes equation (3.33) of Lemma 3.7 to obtain L^2 continuity of tilde_phi(t,s) and hence weak continuity of tilde_rho(t,s). However Lemma 3.7 is stated under Conditions 2.1, 3.2, 3.3 and 3.4, with the additional threshold assumptions such as p gamma >= 2 kappa. Theorem 3.16 assumes only Conditions 3.2 and 3.3 together with alpha > beta^2/2, and does not assume Condition 3.4 or the large-moment condition. Thus the continuity needed for the unique ergodicity argument is not established under the hypotheses of Theorem 3.16. The authors should either prove the required L^2 continuity under Conditions 3.2-3.3 alone or add the missing assumptions to the statement of Theorem 3.16.","section":"Theorem 3.16, proof, Birkhoff step"}],"minor_comments":[{"comment":"There is a typo in 'there exists a uniquen invariant probability measure'; it should read 'a unique invariant probability measure'.","section":"Proof of Theorem 3.16, first paragraph"},{"comment":"The identity 'tilde_mu_s = hat_mu_{s1,s2}' is not meaningful as written; it should be 'tilde_mu_s = hat_mu_{s,s}' or the variables s1 and s2 should be set equal to s in the definition of hat_mu.","section":"Section 3.3, after equation (3.45)"},{"comment":"The notation tilde P in Lemma 4.4 and in the proof of Theorem 4.5 is ambiguous: it appears to denote the measure with density equal to the stochastic exponential, but this is not defined explicitly. Please introduce a clear notation for the Girsanov-changed measure.","section":"Lemma 4.4 and Theorem 4.5"},{"comment":"The paper would benefit from a discussion of how the uniform dissipativity of the two-variable lift can be verified for a given quasi-periodic pair (b,sigma). The OU example constructs the lift explicitly, but for general quasi-periodic coefficients the existence of a lift with a positive uniform alpha is a nontrivial premise that strongly restricts the class of equations covered.","section":"Conditions 3.2 and 3.3"}],"recommendation":"major_revision","confidential_remarks":"The paper introduces an interesting and timely framework, and the non-uniqueness/ergodicity gap in Theorem 3.16 is precisely the advertised capstone result. The central existence results appear sound, and the flaw is likely repairable by adding a second-moment bound for invariant measures and by either proving the needed continuity under weaker assumptions or adjusting the hypotheses. I would not accept the paper in its current form; after a careful revision addressing the two major comments, it could be suitable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis paper is worth reading: it gives the first definitions of random quasi-periodic paths and quasi-periodic measures for SDEs, and proves existence and uniqueness under a dissipativity condition. The machinery — the two-shift reparameterized SDE, the cylinder lift, the explicit O-U example — is genuinely new and mostly executed with care. Theorems 3.5 and 3.9 appear solid: contraction estimates, Lp bounds, tightness and the Feller property all check out.\n\nBut the proof of Theorem 3.16, uniqueness of the invariant measure on the cylinder, has a real gap. The authors assert that ||K^{r1,r2}(t,s,x) - tilde_phi^{r1,r2}(t)||_2 is bounded by C exp(-(alpha-beta^2/2)(t-s)) with C independent of x. That is false. In the O-U case with zero drift and identity noise, the distance is e^{-a(t-s)}|x|. So the error term in (3.58) is actually C(1+|x|^2)/delta^2 times a decaying exponential, and passing through Fatou in (3.59) requires the unknown invariant measure to have finite second moment. The paper never proves that. I suspect the claim is true and the gap is fixable via a standard Lyapunov estimate, but as written the uniqueness proof is incomplete.\n\nThe other soft spot is Section 4: Lemma 4.4, which is load-bearing for the density and Fokker-Planck results, is not proved; the paper defers to a self-cited arXiv preprint. That is worth flagging to a referee, though not fatal if the preprint is correct.\n\nMinor issues: some Ito estimates omit localization details, and Theorem 4.7 bundles uniqueness of the infinite-horizon Fokker-Planck problem into an iff, which is awkward. These are minor.\n\nOverall: the paper makes a real conceptual contribution, and the existence/uniqueness results for quasi-periodic paths and measures are likely correct. The invariant measure uniqueness needs a repaired proof, and the density section is conditional on an external lemma. This should go to peer review — a serious referee should engage with it — but the authors need to address the gap before publication.","headline":"New concepts and solid existence proofs, but the uniqueness proof for the invariant measure on the cylinder contains a false x-independent estimate and an unjustified Fatou step.","tokens_in":34253,"tokens_out":5439,"would_cite":true,"duration_ms":53414,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60H10","37H10","60J60","37A30"],"pacs":[],"model":"deepseek-v4-flash","headline":"Under a dissipativity condition, a stochastic differential equation with quasi-periodic coefficients has a unique random quasi-periodic path whose law is a quasi-periodic measure; lifting to a cylinder yields a unique ergodic invariant…","keywords":["quasi-periodic measures","random quasi-periodic paths","stochastic differential equations","invariant measures","ergodicity","Fokker-Planck equation","random dynamical systems","Markovian semigroups"],"falsifier":"Find a quasi-periodic drift and diffusion that satisfy the original one-sided dissipativity on the diagonal but for which every two-variable periodic lift violates the uniform dissipativity Condition 3.3, and check whether the SDE still has two distinct uniformly $L^p$-bounded random paths; if it does, the theorem's premise is not necessary, and if no such path exists at all, the premise is doing the work.","tokens_in":33153,"feed_emoji":"🎲","tokens_out":6421,"duration_ms":57538,"temperature":0.7,"pith_summary":"This paper introduces random quasi-periodic paths for random dynamical systems and quasi-periodic measures for Markovian semigroups, then proves that for stochastic differential equations whose drift and diffusion are quasi-periodic in time there is a unique uniformly $L^p$-bounded random quasi-periodic path, provided a two-parameter periodic lift of the coefficients is uniformly dissipative and the noise is not too strong relative to the drift. The law of this path is the unique quasi-periodic entrance measure, and its density, when it exists, satisfies the Fokker-Planck equation. Lifting the flow to the cylinder $[0,\\tau_1) \\times [0,\\tau_2) \\times \\mathbb{R}^d$ turns the quasi-periodic measure into a family of measures whose time average converges to a unique invariant measure, given by the double average of $\\delta_{s_1} \\times \\delta_{s_2} \\times \\tilde{\\rho}_{s_1,s_2}$, and this measure is ergodic. The paper matters because it provides a mathematical language for systems that combine randomness with two incommensurate periodicities, such as temperature or energy demand with daily and annual cycles.","feed_headline":"Quasi-periodic SDEs get unique ergodic measures","feed_subtitle":"A new theorem proves a unique random quasi-periodic path and an explicit invariant measure for noisy systems with two clocks.","key_machinery":"The central object is the two-variable periodic lift $(\\tilde{b}, \\tilde{\\sigma})$ of the quasi-periodic coefficients, and the parameterised SDE $K^{r_1,r_2}$ it generates: for each pair $(r_1,r_2)$, $K^{r_1,r_2}(t,s,x)$ solves the SDE with coefficients $\\tilde{b}(t+r_1,t+r_2,x)$ and $\\tilde{\\sigma}(t+r_1,t+r_2,x)$. The key identity (1.3), $K^{r_1,r_2}(t+r,s+r,x,\\theta_{-r}\\omega) = K^{r_1+r,r_2+r}(t,s,x,\\omega)$, gives the family a shift-invariance that the original time-dependent SDE lacks. Under the dissipativity condition, each $K^{r_1,r_2}$ has a unique $L^p$ pull-back limit, and periodicity of the lift in $r_1$ and $r_2$ makes that limit a random quasi-periodic path. Lifting the flow to the cylinder then produces a cocycle whose time averages are tight; Birkhoff ergodic theory and the minimality of the two-torus rotation identify the limiting invariant measure.","core_discovery":"The central discovery is that quasi-periodicity of the coefficients can be converted into a family of autonomous-like SDEs indexed by two parameters, and the pull-back limit produces a unique random quasi-periodic path. More precisely, assuming the coefficients admit a two-variable periodic lift satisfying a uniform one-sided dissipativity condition with exponent $\\alpha$ and Lipschitz noise with constant $\\beta$, and $\\alpha > (p-1)\\beta^2/2$ for some $p \\ge 2$, the SDE has a unique uniformly $L^p$-bounded random quasi-periodic path; its law is the unique quasi-periodic probability measure in $M_p$. When the lifted flow is considered on the cylinder $[0,\\tau_1) \\times [0,\\tau_2) \\times \\mathbb{R}^d$, the averages of the lifted quasi-periodic measures are tight and converge to the unique invariant measure $\\frac{1}{\\tau_1\\tau_2}\\int_0^{\\tau_1}\\int_0^{\\tau_2} \\delta_{s_1} \\times \\delta_{s_2} \\times \\tilde{\\rho}_{s_1,s_2}\\, ds_1\\, ds_2$, which is ergodic.","pith_inferences":["The explicit quasi-periodic Ornstein-Uhlenbeck example suggests a concrete statistical test: fit a quasi-periodic mean-reverting model to temperature or electricity-price data and compare the empirical long-run distribution with the predicted Gaussian quasi-periodic measure; the paper does not run such tests.","The construction should extend to any finite number of incommensurate periods, as the authors state, so the two-period case is a template for multi-period random quasi-periodicity; a natural next step is to check whether the invariant measure remains ergodic with three or more periods.","The uniqueness of the random quasi-periodic path is obtained before any perfection argument, which suggests that the core existence result is insensitive to pathwise regularity; one could attempt to weaken Condition 3.4 by proving continuity of the limiting path with a weaker growth condition.","Since the invariant measure is ergodic on the cylinder, observable long-time averages of smooth functions of the lifted process converge to the double average; this gives a quantitative prediction that could be tested numerically in the quasi-periodic OU model."],"forward_implications":["Every SDE satisfying Conditions 3.2 and 3.3 with $\\alpha > (p-1)\\beta^2/2$ has a unique uniformly $L^p$-bounded random quasi-periodic path, so quasi-periodic forcing does not destroy uniqueness of the long-time random behaviour.","The law of this path is the unique quasi-periodic measure in $M_p$, giving a direct bridge from pathwise random quasi-periodicity to measure-valued quasi-periodicity.","The lifted semigroup on the cylinder has a unique invariant measure, so the long-time statistical behaviour is ergodic and can be written explicitly as an average of $\\delta_{s_1} \\times \\delta_{s_2} \\times \\tilde{\\rho}_{s_1,s_2}$.","When the diffusion coefficient is invertible and the coefficients are smooth, the quasi-periodic measure has a density that satisfies the Fokker-Planck equation, and entrance measures correspond one-to-one to solutions of the infinite-horizon Fokker-Planck problem.","The abstract Theorem 3.18 extends the invariant-measure conclusion to any Feller Markovian semigroup with a tight family of averaged quasi-periodic entrance measures."],"supporting_citations":[{"why":"Supplies the metric dynamical system and measure-preserving shift $\\theta_t$, plus the cocycle framework used to define random paths and the lift to the cylinder.","marker":"[1]"},{"why":"Introduces random periodic solutions and random periodic measures, the single-period ancestor of the quasi-periodic concepts defined here.","marker":"[27]"},{"why":"Establishes random periodic processes, periodic measures and ergodicity, the direct predecessor whose methods are extended to two periods.","marker":"[12]"},{"why":"Provides the density equivalence lemma (Lemma 4.4) and the BMO martingale estimates used to prove absolute continuity and the Fokker-Planck equation.","marker":"[13]"},{"why":"Supplies the minimal rotation and Birkhoff ergodic theory on the two-torus, used to identify the unique ergodic invariant measure on the cylinder.","marker":"[25]"},{"why":"Gives the criterion that a unique invariant measure is ergodic, applied in Theorem 3.16.","marker":"[8]"}],"fun_headline_variants":["Unique ergodic measure for quasi-periodic SDEs","Quasi-periodic paths: unique and ergodic","Dissipative SDEs admit unique quasi-periodic measure","Explicit invariant measure for quasi-periodic SDEs","Two-frequency SDEs: unique path and ergodic measure"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction works only if the quasi-periodic coefficients admit a two-variable periodic lift that is uniformly dissipative (one-sided Lipschitz with a positive $\\alpha$) and H\\\"older continuous, so that the parameterised family of SDEs is well behaved; quasi-periodicity of the original coefficients alone does not guarantee such a lift.","fun_headline_variants_meta":{"raw":{"variants":["Unique ergodic measure for quasi-periodic SDEs","Quasi-periodic paths: unique and ergodic","Dissipative SDEs admit unique quasi-periodic measure","Explicit invariant measure for quasi-periodic SDEs","Two-frequency SDEs: unique path and ergodic measure"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001099,"raw_usage":{"total_tokens":4559,"prompt_tokens":895,"completion_tokens":3664,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":3582}},"tokens_in":511,"tokens_out":3664,"duration_ms":28767,"temperature":1.0,"reasoning_tokens":3582,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:55:51.991898+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a quasi-periodic drift and diffusion that satisfy the original one-sided dissipativity on the diagonal but for which every two-variable periodic lift violates the uniform dissipativity Condition 3.3, and check whether the SDE still has two distinct uniformly $L^p$-bounded random paths; if it does, the theorem's premise is not necessary, and if no such path exists at all, the premise is doing the work.","supporting_citations":[{"cited_title":"Arnold, Random Dynamical Systems , Springer-Verlag Berlin Heidelberg New York, 1998","cited_arxiv_id":null,"evidence_quote":"Supplies the metric dynamical system and measure-preserving shift $\\theta_t$, plus the cocycle framework used to define random paths and the lift to the cylinder."},{"cited_title":"Zhao and Z.H","cited_arxiv_id":null,"evidence_quote":"Introduces random periodic solutions and random periodic measures, the single-period ancestor of the quasi-periodic concepts defined here."},{"cited_title":"Feng and H.Z","cited_arxiv_id":null,"evidence_quote":"Establishes random periodic processes, periodic measures and ergodicity, the direct predecessor whose methods are extended to two periods."},{"cited_title":"Existence of Geometric Ergodic Periodic Measures of Stochastic Differential Equations","cited_arxiv_id":"1904.08091","evidence_quote":"Provides the density equivalence lemma (Lemma 4.4) and the BMO martingale estimates used to prove absolute continuity and the Fokker-Planck equation."},{"cited_title":"Walters, An Introduction to Ergodic Theory, Graduate Tests in Mathem atics, 79, Springer-Verlag New York, 1982","cited_arxiv_id":null,"evidence_quote":"Supplies the minimal rotation and Birkhoff ergodic theory on the two-torus, used to identify the unique ergodic invariant measure on the cylinder."},{"cited_title":"Da Prato and J","cited_arxiv_id":null,"evidence_quote":"Gives the criterion that a unique invariant measure is ergodic, applied in Theorem 3.16."}],"review_version":1}