betaFeedback

WorksMathematics

WAIVES: Weak-form Asymptotically-unbiased Instrumental-Variable Estimation for SDEs

Abstract

Weak-form methods have transformed inference for deterministic dynamical systems. By integrating the governing equations against smooth, compactly supported test functions, they trade noise-amplifying differentiation for noise-attenuating integration, yielding fast and robust estimates. Extending this idea to stochastic differential equations (SDEs) has proven harder, and a consistent single-trajectory formulation in the time domain has remained out of reach. Applied naively in the time domain, the weak form makes the regressors and the errors functionals of the same noise, creating endogeneity that leaves weak-form least squares biased and inconsistent. We resolve it with instrumental variables, a classical idea from econometrics. Weak-form integrals over nearby disjoint time windows prove to be valid instruments, correlated with the regressors yet uncorrelated with the error, restoring consistency and asymptotic normality. The resulting generalized-method-of-moments estimator, WAIVES (Weak-form Asymptotically-unbiased Instrumental-Variable Estimation for SDEs), requires only a linear solve, with no likelihood evaluations, filtering, or simulation. It is therefore an order of magnitude faster than exact-likelihood methods where those exist, and it applies unchanged where they do not, with standard errors and calibrated confidence intervals supplied by the moment conditions. It also separates measurement noise from the intrinsic randomness of the dynamics. Across nine example systems, spanning linear, bistable, chaotic, excitable, state-dependent-diffusion, non-polynomial, hypoelliptic, and high-dimensional regimes (up to 120 dimensions), WAIVES recovers nonlinear drift and diffusion with negligible bias and near-nominal interval coverage. We close with two real-data illustrations: a 167-year North Atlantic sea-surface-temperature record and an optical-tweezer calibration.

Introduction

Data-driven inference of governing equations, from equation discovery with SINDy [1] and WSINDy [2] to parameter estimation, has become a central tool across the quantitative sciences [1, 3, 4, 5]. For deterministic ordinary differential equations (ODEs), weak-form estimation has proven especially powerful [2, 6, 7, 8]. Rather than estimating derivatives pointwise from noisy data, the weak-form approach relies on multiplying the governing equation by a smooth, compactly supported test function and integrating by parts, thus converting differentiation of the data into differentiation of the (known) smooth test function 1. The weak-form estimation of nonlinear dynamics (WENDy) framework [8, 9] exploits this to recover model parameters by linear regression, resulting in inference that is ultra-fast and robust to additive and multiplicative measurement noise.

The extension of the WENDy framework to stochastic differential equations (SDEs) has, to our knowledge, not previously been attempted. Many systems of scientific interest are not merely observed with noise but driven by it. Ion channels in neurons, climate oscillators, ecological populations near extinction, and molecular reaction networks are among the examples in which randomness enters into the dynamics themselves, and not just into the measurement process. For such systems, SDEs and stochastic partial differential equations (SPDEs) are more faithful models than ODEs and PDEs. Recovering the drift and diffusion of an SDE from sampled trajectories is the stochastic analogue of WENDy’s goal and the objective of this paper. SPDEs are left for future work.

Inference for discretely observed diffusions is difficult because the transition density of the SDE is almost never available in closed form, so the exact likelihood of a discretely sampled path cannot be evaluated directly (see, e.g., [10] and the references therein). A substantial body of work addresses this challenge. Some approaches use simulated-likelihood or data-augmentation schemes to account for the unobserved portions of the trajectory between observations [11, 12]. These methods, however, can be computationally demanding, as they require repeated simulation or sampling of the latent paths, and data-augmentation schemes can suffer from poor MCMC mixing due to strong dependence between the latent paths and the diffusion parameters. Other approaches construct closed-form approximations to the transition density [13]. Although these expansions can be highly accurate, practical implementations rely on finite truncations, whose accuracy and computational cost depend on the order of the expansion and the complexity of the underlying model. Finally, exact-simulation methods eliminate discretization error entirely [14, 15]. However, they typically require first transforming the SDE to one with unit diffusion coefficient (the Lamperti transform). Because this transformation generally exists only for scalar diffusions, and for a restricted class of multivariate ones, these methods do not extend to general multivariate systems. Despite their different approaches, these methods are generally developed for settings in which the state is observed exactly at discrete times, without additional measurement noise.

When the state is observed with measurement error, the unknown parameters must be estimated within a state-space formulation that couples them to a nonlinear filtering problem. This is handled through sequential Monte Carlo [16, 17, 18], particle MCMC [19], iterated filtering [20], and ensemble Kalman–Bucy schemes [21], jointly inferring the drift and diffusion parameters together with the latent path and yielding consistent, and in some cases unbiased, estimators even in high-dimensional or partially observed regimes. However, the performance varies with the method and the model class; for example, ensemble Kalman–Bucy schemes are exact only for linear Gaussian dynamics and become approximations without convergence guarantees once the drift or diffusion is nonlinear. A parallel line of work instead parameterizes the drift and diffusion by neural networks [22, 23, 24]; such models are flexible and accommodate irregular sampling, but they require large training sets and are not directly interpretable as governing equations. In all but trivial cases, these methods are also highly computationally intensive.

Within the governing equation discovery setting, sparse-regression and related methods have been adapted to learn the form of stochastic dynamics (and consequently the associated parameters), typically by fitting drift and diffusion fields to empirical estimates of conditional moments or to Fokker–Planck statistics [25, 26, 27]. Weak-form ideas have recently entered this setting as well, for mean-field equations estimated from interacting-particle data [28] and for mean-field Fokker–Planck equations learned from sparse ecological experiments [29]. A weak-collocation approach integrates the Fokker–Planck equation against Gaussian kernels placed in state space and evaluates the resulting integrals as sums over observed samples. This has been done for aggregate snapshot data [30] and, through ergodic averaging, for a single stationary trajectory split into segments [31]. In all of these weak-form methods, the weak form acts on the density of the process, so the stochastic-integral term is removed by the averaging over samples before any regression is performed. In exchange, the temporal structure of the increments is not used. The strong-form approaches, in turn, rely on finite differencing of the sampled path, which amplifies measurement noise and makes parameter inference more challenging. Many of these approaches are also based on empirical estimates of conditional moments or densities, obtained from single-step increments, by binning, or by smoothing. Binning generally discards the within-bin dynamics, and the recovered parameters show bias that grows with the noise level and with coarser sampling [32, 33, 34]. In contrast, WAIVES retains the weak form in the time variable, acting on a single trajectory. This preserves both the temporal information in the increments and the computational efficiency and speed of compactly supported test functions, the same properties behind the weak form’s success for deterministic systems. However, this formulation introduces a difficulty of its own, which we address next.

Applied naively to SDEs, weak-form regression in time fails. When the governing SDE is integrated against a test function in time, the resulting regression carries a stochastic-integral error that is correlated with its own regressors, because both are functionals of the same Brownian path over the test-function support. This is the endogeneity problem of classical regression [35], and it makes the weak-form least-squares estimator biased and inconsistent, regardless of how many test functions are used or how long the trajectory grows. In fact, as we show in § 3, for the Ornstein–Uhlenbeck process the least-squares estimate converges to zero regardless of the record length.

In this work, we propose to address the endogeneity issue directly, using instrumental variables (IV) [35, 36]. IVs are not new to system identification, where they have long been used to obtain consistent estimates of dynamical models from noisy data [37, 38, 39], with recent extensions to nonlinear continuous-time models through auxiliary-model instruments [40] or sample-splitting of the record [41]. Both of these, however, address deterministic output-error models, where the only disturbance is measurement noise, and thus do not carry over to SDEs. Once the driving noise enters the dynamics itself, no noise-free auxiliary trajectory and no subsample independent of the Brownian path exist. Instrumental variable ideas have also appeared against measurement noise alone. For example, in strong-form stochastic SINDy, whose single-step regression is consistent in the absence of measurement noise, the drift term candidates evaluated one time step earlier can serve as instruments against that noise [27]. That construction, however, does not address the endogeneity of the weak form, which is present even for perfectly observed paths, as we show in the next section. The instruments we construct below do.

In principle, an instrument can be any quantity that is correlated with the regressor, but uncorrelated with the error. The difficulty lies in finding an instrument for which both properties hold strongly. We divide the record into short sub-intervals (windows) and construct instruments from weak-form integrals over nearby windows disjoint from the target window on which the regressors are built. Because Brownian increments over disjoint intervals are mutually independent, a weak-form integral computed on one window (against a test function supported on that window) is uncorrelated with the stochastic-integral error of a non-overlapping window, while remaining correlated with the regressors whenever the two windows lie within the memory of the process. The resulting estimator, which we call WAIVES (Weak-form Asymptotically-unbiased Instrumental-Variable Estimation for SDEs), is shown to be consistent, asymptotically unbiased (with the bias vanishing in the joint limit of many windows and fine sampling), asymptotically normal, and it requires neither a path likelihood nor a Radon–Nikodym derivative. WAIVES thus belongs to the moment-based tradition for diffusions, alongside martingale estimating functions [42] and generator-based generalized method of moments (GMM) [43], which derive their moment conditions from the model’s transition or generator structure and presuppose exactly observed states. WAIVES builds them from the test functions and the noisy observations alone.

Because it avoids the finite-difference estimates on which strong-form methods rely, it inherits the weak-form noise robustness and solver-free efficiency, while still handling diffusions that are observed only at discrete times, with noise, and possibly through a linear sensor map. We demonstrate this on the nine example systems of Table 1, from scalar diffusions to a 120-dimensional Lorenz-96 model, and on two real data sets, the 167-year Atlantic Multidecadal Oscillation index (a North Atlantic sea-surface-temperature record) and an optical-tweezer calibration record.

SystemModelChallenge probedRepresentative recovery
Ornstein–Uhlenbeck (OU)dX=κ(μ−X) dt+σ dWtdX=\kappa(\mu-X)\,dt+\sigma\,dW_tlinear baseline with exact likelihood; calibrates the cost of the weak form (τ=1/κ\tau=1/\kappa)κ^=1.497 (0.126)\widehat{\kappa}=1.497\ (0.126) vs 1.5<br>σ^=0.800 (0.020)\widehat{\sigma}=0.800\ (0.020) vs 0.8
Double-well (DW)dX=(αX−βX3) dt+σ dWtdX=(\alpha X-\beta X^3)\,dt+\sigma\,dW_tbistability: local relaxation (τ=0.5 s\tau=0.5\,\mathrm{s}) vs rare hopping (≈13 s\approx13\,\mathrm{s})α^=1.003 (0.149)\widehat{\alpha}=1.003\ (0.149) vs 1<br>σ^=0.498 (0.032)\widehat{\sigma}=0.498\ (0.032) vs 0.5
Stochastic Lorenz-63 (L63)dx=a(y−x) dt+σ dW1dx=a(y-x)\,dt+\sigma\,dW_1; dy=(x(b−z)−y) dt+σ dW2dy=(x(b-z)-y)\,dt+\sigma\,dW_2; dz=(xy−cz) dt+σ dW3dz=(xy-cz)\,dt+\sigma\,dW_3chaos: decorrelation (τ≈0.14 s\tau\approx0.14\,\mathrm{s}) far inside the Lyapunov time; coefficients spanning 2.7 to 28b^=28.027 (0.355)\widehat{b}=28.027\ (0.355) vs 28<br>σ^y=1.075 (0.183)\widehat{\sigma}_y=1.075\ (0.183) vs 1
Stochastic FitzHugh–Nagumo (FHN)dV=(V−V3/3−W+I) dt+σV dB1dV=(V-V^3/3-W+I)\,dt+\sigma_V\,dB_1; dW=ϵ(V+a−bW) dt+σW dB2dW=\epsilon(V+a-bW)\,dt+\sigma_W\,dB_2slow–fast excitability; a,ba,b enter only through products with ϵ≈0.08\epsilon\approx0.08a^=0.684 [0.61,0.77]\widehat{a}=0.684\ [0.61,0.77] vs 0.7<br>σ^W=0.302 (0.038)\widehat{\sigma}_W=0.302\ (0.038) vs 0.3
Cox–Ingersoll–Ross (CIR)dX=κ(μ−X) dt+νX dWtdX=\kappa(\mu-X)\,dt+\nu\sqrt{X}\,dW_tstate-dependent diffusion through the diffusion term candidate h(x)=xh(x)=x−κ^=−2.10 (0.75)-\widehat{\kappa}=-2.10\ (0.75) vs −2-2<br>ν^=1.00 (0.04)\widehat{\nu}=1.00\ (0.04) vs 1
Lorenz-96 (L96)dxi=[(xi+1−xi−2)xi−1−xi+F] dt+σ dWidx_i=[(x_{i+1}-x_{i-2})x_{i-1}-x_i+F]\,dt+\sigma\,dW_idimension: 40 and 120 coupled componentsF^=7.99 (0.04)\widehat{F}=7.99\ (0.04) vs 8 (d=120d=120)<br>σ^=0.501 (0.002)\widehat{\sigma}=0.501\ (0.002) vs 0.5
Kuramotodθi=[ω+KN∑jsin⁡(θj−θi)] dt+σ dWid\theta_i=[\omega+\frac{K}{N}\sum_j\sin(\theta_j-\theta_i)]\,dt+\sigma\,dW_inetwork coupling; trigonometric drift term candidates and their debiasingK^=2.01 (0.10)\widehat{K}=2.01\ (0.10) vs 2.0<br>σ^=0.300 (0.002)\widehat{\sigma}=0.300\ (0.002) vs 0.3
Underdamped Langevindq=p dtdq=p\,dt; dp=[q−q3−γp] dt+σ dWdp=[q-q^3-\gamma p]\,dt+\sigma\,dWdegenerate noise: no Brownian forcing on qqpp in q˙\dot q: 1.00 (0.03) vs 1<br>σ^p=1.013 (0.011)\widehat{\sigma}_p=1.013\ (0.011) vs 1
Lorenz-63, observed through CCas above, observed as Y=CX+ηY=CX+\eta with non-identity CCindirect observation, handled by the deconvolution of Remark 2xx in y˙\dot y: 28.01 (0.67) vs 28<br>σ^y=0.412 (0.072)\widehat{\sigma}_y=0.412\ (0.072) vs 0.4

Table 1. The nine example systems, each chosen to probe one challenge for a weak-form SDE estimator. For each system the last column reports the mean,(std) over M=1000M=1000 replications of one representative drift parameter (the one most sensitive to the listed challenge), and, below it, the diffusion estimate for the same component. Diffusion entries use the lag-one quadratic-variation form, except the slope form for the Langevin at γ=0.5\gamma=0.5. The Cox–Ingersoll–Ross entry lists the drift rate −κ^-\widehat{\kappa} above ν^\widehat{\nu}, and the FitzHugh–Nagumo ratio parameter a^\widehat{a} is given as its mode with a 95% Fieller interval. Full drift and diffusion estimates for all nine systems appear in Table 2 and Table 3. Although each row is designed to stress a different failure mode, the same estimator under the same default rules recovers every reported parameter with negligible bias.

The weak form for SDEs

Model

We consider the state-space model in which a latent process obeys an Itô SDE and is observed discretely

The deterministic weak form

under additive Gaussian error,

State:dXt=f(Xt,θ) dt+σ(Xt,ν) dWt,Obs.:Ytn=Xtn+ηn,ηn∼iidN(0,R),\begin{aligned} \text{State:}\quad&\mathrm{d}X_t = f(X_t,\theta)\,\mathrm{d}t + \sigma(X_t,\nu)\,\mathrm{d}W_t,\\ \text{Obs.:}\quad&Y_{t_n} = X_{t_n} + \eta_n,\qquad\eta_n \overset{\mathrm{iid}}{\sim} \mathcal{N}(0,R), \end{aligned}

where {Wt}\{W_t\} is a standard Brownian motion in Rd\mathbb{R}^d and the state Xt∈RdX_t \in\mathbb{R}^d. The observations are taken at the discrete instants tn=nΔtt_n = n\Delta t for n=0,…,Nn = 0,\ldots,N, so that the record spans [0,T][0,T] with T=NΔtT = N\Delta t. The observation errors ηn∈Rdy\eta_n \in\mathbb{R}^{d_y} have covariance R∈Rdy×dyR \in\mathbb{R}^{d_y \times d_y}, which is assumed unknown. Throughout we assume dy=dd_y = d, but Remark 2 treats the more general linear observation model Ytn=CXtn+ηnY_{t_n} = C X_{t_n} + \eta_n with dy≥dd_y \ge d. The errors {ηn}\{\eta_n\} are assumed to be independent of {Wt}\{W_t\}, and the drift is assumed to be linear in the unknown parameters:

f(x,θ)=∑j=1pθjgj(x),(1)f(x,\theta) = \sum_{j=1}^{p} \theta_j g_j(x), \tag*{(1)}

with known drift term candidates {gj}j=1p\{g_j\}_{j=1}^{p} such as monomials or trigonometric functions, and the unknown parameter vector θ=(θ1,…,θp)T∈Rp\theta= (\theta_1,\ldots,\theta_p)^{\mathsf{T}} \in\mathbb{R}^p.

The diffusion is parameterized in the same way. We take σ=σ(Xt,ν)\sigma= \sigma(X_t,\nu) and, under the diagonal-noise assumption used throughout (each state component driven by its own independent Brownian motion), expand the squared diffusion of component ii in known diffusion term candidates {hl}l=1q\{h_l\}_{l=1}^{q},

σi2(x,ν)=∑l=1qνilhl(x),(2)\sigma_i^2(x,\nu) = \sum_{l=1}^{q} \nu_{il} h_l(x), \tag*{(2)}

References

  1. [1]Steven L. Brunton, Joshua L. Proctor, and J. Nathan Kutz. Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proc. Natl. Acad. Sci. U.S.A., 113(15):3932–3937, 2016.DOI
  2. [2]Daniel A. Messenger and David M. Bortz. Weak SINDy: Galerkin-based data-driven model selection. Multiscale Model. Simul., 19(3):1474–1497, 2021.DOI
  3. [3]Michael Schmidt and Hod Lipson. Distilling free-form natural laws from experimental data. Science, 324(5923):81–85, 2009.DOI
  4. [4]Samuel H. Rudy, Steven L. Brunton, Joshua L. Proctor, and J. Nathan Kutz. Data-driven discovery of partial differential equations. Sci. Adv., 3(4):e1602614, 2017.
  5. [5]Steven L. Brunton and J. Nathan Kutz. Data-Driven Science and Engineering: Machine Learning, Dynamical Systems, and Control. Cambridge University Press, 2 edition, 2022.
  6. [6]Daniel A. Messenger and David M. Bortz. Weak SINDy for partial differential equations. Journal of Computational Physics, 443:110525, 2021.DOI
  7. [7]Patrick A. K. Reinbold, Daniel R. Gurevich, and Roman O. Grigoriev. Using noisy or incomplete data to discover models of spatiotemporal dynamics. Physical Review E, 101:010203, 2020.DOI
  8. [8]David M. Bortz, Daniel A. Messenger, and Vanja Dukić. Direct estimation of parameters in ODE models using WENDy: Weak-form estimation of nonlinear dynamics. Bull. Math. Biol., 85(11):110, 2023.DOI
  9. [9]Nicholas Rummel, Daniel A. Messenger, Stephen Becker, Vanja Dukic, and David M. Bortz. WENDy for nonlinear-in-parameters ODEs. SIAM Journal on Scientific Computing, 2026. in press.arxiv.org/abs/2502.08881
  10. [10]Helle Sørensen. Parametric inference for diffusion processes observed at discrete points in time: a survey. Int. Stat. Rev., 72(3):337–354, 2004.DOI
  11. [11]Asger Roer Pedersen. A new approach to maximum likelihood estimation for stochastic differential equations based on discrete observations. Scand. J. Stat., 22(1):55–71, 1995.
  12. [12]Gareth O. Roberts and Osnat Stramer. On inference for partially observed nonlinear diffusion models using the Metropolis–Hastings algorithm. Biometrika, 88(3):603–621, 2001.DOI
  13. [13]Yacine Aït-Sahalia. Maximum likelihood estimation of discretely sampled diffusions: a closed-form approximation approach. Econometrica, 70(1):223–262, 2002.
  14. [14]Alexandros Beskos, Omiros Papaspiliopoulos, Gareth O. Roberts, and Paul Fearnhead. Exact and computationally efficient likelihood-based estimation for discretely observed diffusion processes. J. R. Stat. Soc. B, 68(3):333–382, 2006.DOI
  15. [15]Alexandros Beskos, Omiros Papaspiliopoulos, and Gareth O. Roberts. Monte Carlo maximum likelihood estimation for discretely observed diffusion processes. Ann. Stat., 37(1):223–245, 2009.DOI
  16. [16]Andrew Golightly and Darren J. Wilkinson. Bayesian inference for nonlinear multivariate diffusion models observed with error. Comput. Stat. Data Anal., 52(3):1674–1693, 2008.DOI
  17. [17]Nikolas Kantas, Arnaud Doucet, Sumeetpal S. Singh, Jan Maciejowski, and Nicolas Chopin. On particle methods for parameter estimation in state-space models. Stat. Sci., 30(3):328–351, 2015.
  18. [18]Alexandros Beskos, Dan Crisan, Ajay Jasra, Nikolas Kantas, and Hamza Ruzayqat. Score-based parameter estimation for a class of continuous-time state space models. SIAM J. Sci. Comput., 43(4):A2555–A2580, 2021.arxiv.org/abs/2008.07803
  19. [19]Christophe Andrieu, Arnaud Doucet, and Roman Holenstein. Particle Markov chain Monte Carlo methods. J. R. Stat. Soc. B, 72(3):269–342, 2010.DOI
  20. [20]Edward L. Ionides, Carlos Bretó, and Aaron A. King. Inference for nonlinear dynamical systems. Proc. Natl. Acad. Sci. U.S.A., 103(49):18438–18443, 2006.pubmed.ncbi.nlm.nih.gov/17121996
  21. [21]Dan Crisan, Pierre Del Moral, Ajay Jasra, and Hamza Ruzayqat. Log-normalization constant estimation using the ensemble Kalman–Bucy filter with application to high-dimensional models. Adv. Appl. Probab., 54(4):1139–1163, 2022.
  22. [22]Belinda Tzen and Maxim Raginsky. Neural stochastic differential equations: deep latent Gaussian models in the diffusion limit, 2019. arXiv:1905.09883.arxiv.org/abs/1905.09883
  23. [23]Xuechen Li, Ting-Kam Leonard Wong, Ricky T. Q. Chen, and David Duvenaud. Scalable gradients for stochastic differential equations. In Proc. 23rd Int. Conf. Artif. Intell. Stat. (AISTATS), volume 108 of Proc. Mach. Learn. Res., pages 3870–3882. PMLR, 2020.arxiv.org/abs/2001.01328
  24. [24]Patrick Kidger, James Foster, Xuechen Li, Harald Oberhauser, and Terry Lyons. Neural SDEs as infinite-dimensional GANs. In Proc. 38th Int. Conf. Mach. Learn. (ICML), volume 139 of Proc. Mach. Learn. Res., pages 5453–5463. PMLR, 2021.arxiv.org/abs/2102.03657
  25. [25]Lorenzo Boninsegna, Feliks Nüske, and Cecilia Clementi. Sparse learning of stochastic dynamical equations. J. Chem. Phys., 148(24):241723, 2018.DOI
  26. [26]Jared L. Callaham, Jean-Christophe Loiseau, Georgios Rigas, and Steven L. Brunton. Nonlinear stochastic modelling with Langevin regression. Proc. R. Soc. A, 477(2250):20210092, 2021.DOI
  27. [27]Mathias Wanner and Igor Mezić. On higher order drift and diffusion estimates for stochastic sindy. SIAM Journal on Applied Dynamical Systems, 23(2):1504–1539, 2024.arxiv.org/abs/2306.17814
  28. [28]Daniel A. Messenger and David M. Bortz. Learning mean-field equations from particle data using WSINDy. Physica D: Nonlinear Phenomena, 439:133406, 2022.DOI
  29. [29]Seth Minor, Bret D. Elder, Benjamin Van Allen, David M. Bortz, and Vanja Dukic. Weak-form learning for mean-field partial differential equations: An application to insect movement. SIAM Journal on Life Sciences, 1(3):444–484, 2026.
  30. [30]Liwei Lu, Zhijun Zeng, Yan Jiang, Yi Zhu, and Pipi Hu. Weak collocation regression method: Fast reveal hidden stochastic dynamics from high-dimensional aggregate data. J. Comput. Phys., 502:112799, 2024.
  31. [31]Yan Jiang, Zhijun Zeng, Wuyue Yang, Liu Hong, Pipi Hu, and Yi Zhu. STWCR: Weak collocation regression for revealing hidden stochastic dynamics from single trajectory data. Appl. Math. Comput., 521:129958, 2026.
  32. [32]Mario Ragwitz and Holger Kantz. Indispensable finite time corrections for Fokker–Planck equations from time series data. Physical Review Letters, 87:254501, 2001.
  33. [33]Frank Böttcher, Joachim Peinke, David Kleinhans, Rudolf Friedrich, Pedro G. Lind, and Maria Haase. Reconstruction of complex dynamical systems affected by strong measurement noise. Physical Review Letters, 97:090603, 2006.pubmed.ncbi.nlm.nih.gov/17026351
  34. [34]Rudolf Friedrich, Joachim Peinke, Muhammad Sahimi, and M. Reza Rahimi Tabar. Approaching complexity by stochastic methods: From biological systems to turbulence. Physics Reports, 506(5):87–162, 2011.
  35. [35]J. D. Sargan. The estimation of economic relationships using instrumental variables. Econometrica, 26(3):393–415, 1958.DOI
  36. [36]Lars Peter Hansen. Large sample properties of generalized method of moments estimators. Econometrica, 50(4):1029–1054, 1982.
  37. [37]Peter C. Young. An instrumental variable method for real-time identification of a noisy process. Automatica, 6(2):271–287, 1970.
  38. [38]Torsten Söderström and Petre Stoica. Instrumental Variable Methods for System Identification, volume 57 of Lect. Notes Control Inf. Sci. Springer-Verlag, Berlin, 1983.
  39. [39]Torsten Söderström and Petre Stoica. System Identification. Prentice Hall, Hemel Hempstead, UK, 1989.
  40. [40]Thomas Zeiringer, Hugues Garnier, and Martin Horn. Instrumental variable framework for nonlinear continuous-time system identification. Journal of the Franklin Institute, 363:108619, 2026.
  41. [41]Simon Kuang and Xinfan Lin. Instrumental variables system identification with Iᵖ consistency. In Proceedings of the 8th Annual Conference on Learning for Dynamics and Control, volume 331 of Proceedings of Machine Learning Research, pages 1713–1740, 2026.
  42. [42]Bo Martin Bibby and Michael Sørensen. Martingale estimation functions for discretely observed diffusion processes. Bernoulli, 1(1/2):17–39, 1995.DOI
  43. [43]Lars Peter Hansen and José A. Scheinkman. Back to the future: generating moment implications for continuous-time Markov processes. Econometrica, 63(4):767–804, 1995.
  44. [44]Jeffrey M. Wooldridge. Econometric Analysis of Cross Section and Panel Data. MIT Press, Cambridge, MA, 2nd edition, 2010.
  45. [45]Grigorios A. Pavliotis. Stochastic Processes and Applications: Diffusion Processes, the Fokker–Planck and Langevin Equations, volume 60 of Texts in Applied Mathematics. Springer, New York, 2014.DOI
  46. [46]Douglas Staiger and James H. Stock. Instrumental variables regression with weak instruments. Econometrica, 65(3):557–586, 1997.
  47. [47]Leonard A. Stefanski. Unbiased estimation of a nonlinear function of a normal mean with application to measurement error models. Communications in Statistics—Theory and Methods, 18(12):4335–4358, 1989.
  48. [48]Tsuyoshi Nakamura. Corrected score function for errors-in-variables models: methodology and application to generalized linear models. Biometrika, 77(1):127–137, 1990.DOI
  49. [49]Raymond J. Carroll, David Ruppert, Leonard A. Stefanski, and Ciprian M. Crainiceanu. Measurement Error in Nonlinear Models: A Modern Perspective. Chapman & Hall/CRC, Boca Raton, 2nd edition, 2006.
  50. [50]John G. Cragg and Stephen G. Donald. Testing identifiability and specification in instrumental variable models. Econometric Theory, 9(2):222–240, 1993.DOI
  51. [51]James H. Stock and Motohiro Yogo. Testing for weak instruments in linear IV regression. In Donald W. K. Andrews and James H. Stock, editors, Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg, pages 80–108. Cambridge University Press, Cambridge, 2005.
  52. [52]Robert H. Shumway and David S. Stoffer. An approach to time series smoothing and forecasting using the EM algorithm. Journal of Time Series Analysis, 3(4):253–264, 1982.DOI
  53. [53]George Marsaglia. Ratios of normal variables. J. Stat. Softw., 16(4):1–10, 2006.DOI
  54. [54]D. V. Hinkley. On the ratio of two correlated normal random variables. Biometrika, 56(3):635–639, 1969.
  55. [55]E. C. Fieller. Some problems in interval estimation. J. R. Stat. Soc. B, 16(2):175–185, 1954.DOI
  56. [56]David B. Enfield, Alberto M. Mestas-Nuñez, and Paul J. Trimble. The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental U.S. Geophysical Research Letters, 28(10):2077–2080, 2001.DOI
  57. [57]Alexey Kaplan, Mark A. Cane, Yochanan Kushnir, Amy C. Clement, M. Benno Blumenthal, and Balaji Rajagopalan. Analyses of global sea surface temperature 1856–1991. Journal of Geophysical Research: Oceans, 103(C9):18567–18589, 1998.
  58. [58]Jan Gieseler, Juan Ruben Gomez-Solano, Alessandro Magazzù, Isaac Pérez Castillo, Laura Pérez García, Marta Gironella-Torrent, Xavier Viader-Godoy, Felix Ritort, Giuseppe Pesce, Alejandro V. Arzola, Karen Volke-Sepúlveda, and Giovanni Volpe. Optical tweezers — from calibration to applications: a tutorial. Adv. Opt. Photonics, 13(1):74–241, 2021.arxiv.org/abs/2004.05246
  59. [59]Niklas Boers. Observation-based early-warning signals for a collapse of the Atlantic Meridional Overturning Circulation. Nature Climate Change, 11:680–688, 2021.DOI
  60. [60]Peter Ditlevsen and Susanne Ditlevsen. Warning of a forthcoming collapse of the Atlantic meridional overturning circulation. Nature Communications, 14:4254, 2023.
  61. [61]Andreas Morr, Maya Ben-Yami, Brian Groenke, Christof Schötz, Alessandro Cotronei, Eirik Myrvoll-Nilsen, Sebastian Bathiany, Martin Rypdal, and Niklas Boers. Extrapolation from historical data cannot reliably predict the time of a potential AMOC collapse, 2026. arXiv preprint, arXiv:2604.20341 [physics.geo-ph].arxiv.org/abs/2604.20341
  62. [62]Maya Ben-Yami, Andreas Morr, Sebastian Bathiany, and Niklas Boers. Uncertainties too large to predict tipping times of major Earth system components from historical data. Science Advances, 10(31):eadl4841, 2024.
  63. [63]Alessandro Cotronei, Eirik Myrvoll-Nilsen, and Martin Rypdal. Evaluating model uncertainty in critical threshold estimations from time series data: Application to the Atlantic meridional Overturning Circulation. Frontiers in Climate, 8:1761461, 2026.DOI
  64. [64]Michael E. Mann, Byron A. Steinman, Daniel J. Brouillette, and Sonya K. Miller. Multidecadal climate oscillations during the past millennium driven by volcanic forcing. Science, 371(6533):1014–1019, 2021.DOI
  65. [65]Donald W. K. Andrews and Biao Lu. Consistent model and moment selection procedures for GMM estimation with application to dynamic panel data models. Journal of Econometrics, 101(1):123–164, 2001.DOI
  66. [66]Whitney K. Newey and Kenneth D. West. Hypothesis testing with efficient method of moments estimation. International Economic Review, 28(3):777–787, 1987.DOI
  67. [67]Kirstine Berg-Sørensen, Erwin J. G. Peterman, Tom Weber, Christoph F. Schmidt, and Henrik Flyvbjerg. Power spectrum analysis for optical tweezers. II: Laser wavelength dependence of parasitic filtering, and how to achieve high bandwidth. Rev. Sci. Instrum., 77(6):063106, 2006.
  68. [68]Arnaud Gloter and Jean Jacod. Diffusions with measurement errors. I. local asymptotic normality. ESAIM: Probability and Statistics, 5:225–242, 2001.
  69. [69]A. W. van der Vaart. Asymptotic Statistics. Cambridge University Press, Cambridge, 1998.
  70. [70]Patrick Billingsley. The Lindeberg–Lévy theorem for martingales. Proceedings of the American Mathematical Society, 12(5):788–792, 1961.
  71. [71]Peter Hall and Christopher C. Heyde. Martingale Limit Theory and Its Application. Academic Press, New York, 1980.
  72. [72]Jonathan C. Mattingly, Andrew M. Stuart, and Desmond J. Higham. Ergodicity for SDEs and approximations: locally Lipschitz vector fields and degenerate noise. Stochastic Processes and their Applications, 101(2):185–232, 2002.DOI
  73. [73]Whitney K. Newey and Kenneth D. West. Automatic lag selection in covariance matrix estimation. The Review of Economic Studies, 61(4):631–653, 1994.
  74. [74]Hans R. Künsch. The jackknife and the bootstrap for general stationary observations. Annals of Statistics, 17(3):1217–1241, 1989.DOI
  75. [75]Terrence W. Kinal. The existence of moments of k-class estimators. Econometrica, 48(1):241–249, 1980.DOI

Paper details

vlnc:W_phhsuwpd
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Claude, ChatGPT

View statement
Looking for related papers…
Contents