mlsanityData to 13 Oct 2026

Learning to Assimilate in Chaotic Dynamical Systems

arXiv 2111.01058 ↗First posted 1 Nov 2021cs.LG nlin.CDcited by 10

Not on the map. Why?

Abstract (arXiv)

The accuracy of simulation-based forecasting in chaotic systems is heavily dependent on high-quality estimates of the system state at the time the forecast is initialized. Data assimilation methods are used to infer these initial conditions by systematically combining noisy, incomplete observations and numerical models of system dynamics to produce effective estimation schemes. We introduce amortized assimilation, a framework for learning to assimilate in dynamical systems from sequences of noisy observations with no need for ground truth data. We motivate the framework by extending powerful results from self-supervised denoising to the dynamical systems setting through the use of differentiable simulation. Experimental results across several benchmark systems highlight the improved effectiveness of our approach over widely-used data assimilation methods.

Key facts. Cited by 10 papers in the map (data through 13 Oct 2026). Builds on 22 of its 79 references that are in the map. First posted on arXiv 1 Nov 2021.

Builds on

22 of 79 references.

References · 22

  1. 2015Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learningcited by 4,430
  2. 2018Neural Ordinary Differential Equationscited by 3,750
  3. 2020Prevalence of Neural Collapse during the terminal phase of deep learning trainingcited by 716
  4. 2018Noise2Noise: Learning Image Restoration without Clean Datacited by 529
  5. 2018Noise2Void - Learning Denoising from Single Noisy Imagescited by 361
  6. 2019Noise2Self: Blind Denoising by Self-Supervisioncited by 275
  7. 2020Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solverscited by 205
  8. 2015Deep Kalman Filterscited by 165
  9. 2019High-Quality Self-Supervised Deep Image Denoisingcited by 161
  10. 2016Deep Variational Bayes Filters: Unsupervised Learning of State Space Models from Raw Datacited by 154
  11. 2020Combining data assimilation and machine learning to emulate a dynamical model from sparse and noisy observations: a case study with the Lorenz 96 modelcited by 72
  12. 2019Recurrent Kalman Networks: Factorized Inference in High-Dimensional Deep Feature Spacescited by 66
10 more
  1. 2017Auto-Encoding Sequential Monte Carlocited by 64
  2. 2020Deep Learning for Post-Processing Ensemble Weather Forecastscited by 52
  3. 2016Backprop KF: Learning Discriminative Deterministic State Estimatorscited by 47
  4. 2019Imputing Missing Events in Continuous-Time Event Streamscited by 43
  5. 2020A case for new neural network smoothness constraintscited by 43
  6. 2017Long Short-Term Memory Kalman Filters:Recurrent Neural Estimators for Pose Regularizationcited by 36
  7. 2019Particle Filter Recurrent Neural Networkscited by 36
  8. 2015Neural Adaptive Sequential Monte Carlocited by 34
  9. 2020Patch2Self: Denoising Diffusion MRI with Self-Supervised Learningcited by 28
  10. 2021Variational Data Assimilation with a Learned Inverse Observation Operatorcited by 28

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2022: 1; 2023: 1; 2024: 4; 2025: 2; 2026: 2.

10 papers.

  1. 2022Bayesian Spline Learning for Equation Discovery of Nonlinear Dynamics with Quantified Uncertaintycited by 13
  2. 2025Learning Enhanced Ensemble Filterscited by 11
  3. 2024Learning Optimal Filters Using Variational Inferencecited by 9
  4. 2024Accurate deep learning-based filtering for chaotic dynamics by identifying instabilities without an ensemblecited by 7
  5. 2023Recurrences reveal shared causal drivers of complex time seriescited by 5
  6. 2025Ensemble Kalman filter in latent space using a variational autoencoder paircited by 5
  7. 2024Benchmarking AI-based data assimilation to advance data-driven global weather forecastingcited by 3
  8. 2026Closed-form conditional diffusion models for data assimilationcited by 3
  9. 2026Learning Developmental Scaffoldings to Guide Self-Organisationcited by 2
  10. 2024Ensemble data assimilation to diagnose AI-based weather prediction model: A case with ClimaX version 0.3.1cited by 1