This work empirically verify that the weak formulation, with a proper choice of test function and integration domain, effectively filters noisy data and explains why a weak form loss function is analogous to fitting a model to filtered data and provides a practical way to parameterize the weak form.
Abstract
The accurate forecasting of complex, high-dimensional dynamical systems from observational data is a fundamental task across numerous scientific and engineering disciplines. A significant challenge arises from noisy observations of deterministic dynamics, which severely degrade the performance of data-driven models. In chaotic dynamical systems, where small initial errors amplify exponentially, it is particularly difficult to develop a model from noisy data that achieves short-term accuracy while preserving long-term invariant properties. To overcome this, we consider the weak formulation as a complementary approach to the classical L2-loss function for training models of dynamical systems. We empirically verify that the weak formulation, with a proper choice of test function and integration domain, effectively filters noisy data. This insight explains why a weak form loss function is analogous to fitting a model to filtered data and provides a practical way to parameterize the weak form. Subsequently, we demonstrate how this approach overcomes the instability and inaccuracy of standard Neural ODE (NODE) in modeling chaotic systems. Through numerical examples, we show that our proposed training strategy, the Weak Penalty NODE, is computationally efficient, solver-agnostic, and yields accurate and robust forecasts across benchmark chaotic systems and a real-world climate dataset.
Accurate prediction of complex dynamical systems from noisy measurements remains a significant challenge in scientific computing. Kernel ridge regression learning strategies are often effective when applied to clean data, but have limited success with noisy data. Recent work has observed that a weak formulation can act to filter noisy data, and different learning strategies have achieved increased noise robustness with a weak-form framework. In this manuscript, we give an overview of the filtering mechanism behind the weak formulation and provide a bias-variance error decomposition. Using these insights, we combine a weak formulation with a kernel learning strategy to propose Weak-form Kernel Ridge Regression (WKRR) for learning dynamical systems. The proposed framework is simple to implement, effective for both clean and noisy data, and outperforms several baseline methods. We demonstrate the performance of WKRR on chaotic benchmark systems in up to 64 dimensions, as well as 15,000-dimensional real-world fluid data.
Max Kreider, John Harlim, Daning Huang· 2 citations· ⚡1
We study the reconstruction of an unknown dynamical system from a single noisy scalar time series. The goal is to recover the underlying dynamics for forecasting. We introduce a method that uses differential embedding coordinates to identify a rational closure of the embedding dynamics directly from data. The closure is identified through a weak-form regression pipeline, which avoids unstable pointwise differentiation of noisy data. When applied to noise-free Lorenz and R\"ossler systems, the method recovers closures that support long forecasts across a broad ensemble of realizations ($18.1$ and $7.1$ Lyapunov times respectively). Under $15$--$30\%$ additive Gaussian noise, performance becomes system-dependent. For the Lorenz system, forecast horizons remain short even in the best cases, whereas the R\"ossler system generally performs better in absolute terms, though not once normalized by the Lyapunov time. Our proposed method recovers directly interpretable closure coefficients which we compared against the known analytic closures of the Lorenz and R\"ossler systems.
Distinguishing chaos from noise using time-series data is fundamentally challenging because both exhibit irregular fluctuations and share many statistical and dynamical characteristics. Existing methods face two key limitations: temporally correlated noise can yield spurious signatures of chaos, and analyses of scalar time series often require explicit choices of embedding parameters. Here, we propose a purely data-driven method for distinguishing chaos and noise based on a reservoir-computing framework with a cross-prediction scheme. In the proposed approach, the model is trained to predict the future change of a variable from its current value, thereby combining a short-term predictability test with a test of the smoothness of deterministic flows. The recurrent structure of reservoir computing enables effective prediction of high-dimensional chaotic dynamics even from scalar time series without explicit delay-coordinate reconstruction, while the cross-prediction framework strongly suppresses spurious predictive correlations arising from noise. We apply the proposed method to diverse synthetic and empirical time series. Chaotic systems consistently yield strong correlations between the true and predicted future changes, whereas noise processes remain clearly separated in a low-correlation regime. The method also exhibits substantial robustness to practical limitations in empirical data, including measurement noise, limited data length, and increasing prediction lag. These results demonstrate that the squared Pearson correlation coefficient provides a simple quantitative criterion for distinguishing chaos from noise directly from observed time-series data.
Discovering governing equations directly from observational data remains a fundamental challenge in science and engineering, particularly when measurements are noisy, high-dimensional, or multi-scale. Existing approaches often cast equation discovery as a regression problem that selects candidate terms to fit observed trajectories, which can limit structural stability and identifiability under realistic data conditions. We propose a structured operator-learning framework that reformulates equation discovery as a constrained dynamical inference problem integrating spectral decomposition, physics-guided sparse projection, and cross-view consistency regularization within a unified architecture. By decomposing dynamics into scale-resolved components and enforcing invariance across perturbed observations, the framework promotes stable and interpretable equation recovery. Here, we show that the method consistently identifies compact governing equations while maintaining strong long-horizon predictive accuracy across canonical nonlinear systems and representative industrial processes, even under noisy and distribution-shifted data. Here, the authors propose structured operator learning with spectral decomposition, sparse regression, and cross-view regularization to recover stable, interpretable governing equations under noisy, high-dimensional, and distribution-shifted data.
We introduce a data-driven framework that maps noisy oscillatory time series directly onto the Hopf normal form, enabling inference of underlying dynamics without knowledge of governing equations. By embedding the normal form in a probabilistic state-space model, the method jointly infers latent states and system parameters, yielding robust estimates of the natural frequency, Floquet exponent, and asymptotic phase even far from the bifurcation point and under strong noise. Combined with complex Gaussian process regression, the approach further reconstructs phase and amplitude sensitivity functions from data. Benchmarks on the van der Pol oscillator demonstrate substantially improved accuracy and noise robustness compared with existing phase-based and regression methods. This work establishes a direct bridge between normal-form theory and statistical inference, providing a general and practical route to low-dimensional descriptions of oscillatory dynamics in complex systems.
Chaotic systems exhibit sensitivity to initial conditions and external parameters, posing challenges for long-term prediction. This study investigates the capability of deep neural networks (DNNs) to infer the time-evolution of the Rössler system a canonical chaotic oscillator by leveraging initial conditions (x0,y0,z0) and forcing parameters (a,b,c) as input variables. A 3D convolutional neural network (3D-CNN) architecture is designed to map these inputs to future states of the system. Results demonstrate that the DNN achieves high accuracy in short-term predictions (<50 time units) but faces exponential error growth beyond this horizon due to chaos. Notably, parameter variations (a,b,c) induce systematic shifts in attractor topology, while initial conditions amplify prediction uncertainty. The study highlights DNNs as viable tools for short-term chaotic forecasting but underscores the need for hybrid approaches to address long-term instability.
A. Fateh, Harrag Abdelmalek, F. Mohamed et al.· International Journal of App...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduAug 18, 2026
A new method for surgically removing training examples from a model reveals that as datasets grow, the link between what a model learns and what it produces dissolves.
