mlsanityData to 7 Oct 2026

Meta-Learning Dynamics Forecasting Using Task Inference

arXiv 2102.10271 ↗First posted 20 Feb 2021cs.LGcited by 37

Not on the map. Why?

Abstract (arXiv)

Current deep learning models for dynamics forecasting struggle with generalization. They can only forecast in a specific domain and fail when applied to systems with different parameters, external forces, or boundary conditions. We propose a model-based meta-learning method called DyAd which can generalize across heterogeneous domains by partitioning them into different tasks. DyAd has two parts: an encoder which infers the time-invariant hidden features of the task with weak supervision, and a forecaster which learns the shared dynamics of the entire domain. The encoder adapts and controls the forecaster during inference using adaptive instance normalization and adaptive padding. Theoretically, we prove that the generalization error of such procedure is related to the task relatedness in the source domain, as well as the domain differences between source and target. Experimentally, we demonstrate that our model outperforms state-of-the-art approaches on both turbulent flow and real-world ocean data forecasting tasks.

Key facts. Cited by 37 papers in the map (data through 7 Oct 2026). Builds on 60 of its 94 references that are in the map. First posted on arXiv 20 Feb 2021.

Builds on

60 of 94 references.

References · 60

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  3. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  4. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  5. 2018Neural Ordinary Differential Equationscited by 3,733
  6. 2017Prototypical Networks for Few-shot Learningcited by 3,001
  7. 2020Fourier Neural Operator for Parametric Partial Differential Equationscited by 2,806
  8. 2016Matching Networks for One Shot Learningcited by 2,246
  9. 2017Arbitrary Style Transfer in Real-time with Adaptive Instance Normalizationcited by 1,998
  10. 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,345
  11. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 929
  12. 2020Learning Mesh-Based Simulation with Graph Networkscited by 840
38 more
  1. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 744
  2. 2019Hamiltonian Neural Networkscited by 702
  3. 2016Learning to learn by gradient descent by gradient descentcited by 616
  4. 2018End-to-End Multi-Task Learning with Attentioncited by 533
  5. 2018Meta-Learning with Latent Embedding Optimizationcited by 476
  6. 2020Lagrangian Neural Networkscited by 388
  7. 2017Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equationscited by 362
  8. 2016A Learned Representation For Artistic Stylecited by 340
  9. 2015Deep multi-scale video prediction beyond mean square errorcited by 297
  10. 2017Meta Networkscited by 268
  11. 2019Cormorant: Covariant Molecular Neural Networkscited by 262
  12. 2016Unsupervised Learning for Physical Interaction through Video Predictioncited by 256
  13. 2018PAD-Net: Multi-Tasks Guided Prediction-and-Distillation Network for Simultaneous Depth Estimation and Scene Parsingcited by 224
  14. 2019Deep Lagrangian Networks: Using Physics as Model Prior for Deep Learningcited by 205
  15. 2019Towards Physics-informed Deep Learning for Turbulent Flow Predictioncited by 203
  16. 2017Neural Style Transfer: A Reviewcited by 199
  17. 2015The Benefit of Multitask Representation Learningcited by 194
  18. 2018Bayesian Model-Agnostic Meta-Learningcited by 194
  19. 2020Disentangling Physical Dynamics from Unknown Factors for Unsupervised Video Predictioncited by 192
  20. 2017Deep Learning for Precipitation Nowcasting: A Benchmark and A New Modelcited by 180
  21. 2018Recasting Gradient-Based Meta-Learning as Hierarchical Bayescited by 180
  22. 2015Action-Conditional Video Prediction using Deep Networks in Atari Gamescited by 171
  23. 2017Decomposing Motion and Content for Natural Video Sequence Predictioncited by 168
  24. 2016Accelerating Eulerian Fluid Simulation With Convolutional Networkscited by 128
  25. 2018Meta-Reinforcement Learning of Structured Exploration Strategiescited by 128
  26. 2018Style Transfer Through Back-Translationcited by 124
  27. 2020Incorporating Symmetry into Deep Dynamics Models for Improved Generalizationcited by 123
  28. 2020Forecasting Sequential Data using Consistent Koopman Autoencoderscited by 114
  29. 2020A Review on Deep Learning Techniques for Video Predictioncited by 114
  30. 2020Learning Stable Deep Dynamics Modelscited by 110
  31. 2016Theoretical Analysis of Domain Adaptation with Optimal Transportcited by 105
  32. 2017Deep Learning for Physical Processes: Incorporating Prior Scientific Knowledgecited by 95
  33. 2017Exploring the structure of a real-time, arbitrary neural artistic stylization networkcited by 89
  34. 2020Multi-Source Deep Domain Adaptation with Weak Supervision for Time-Series Sensor Datacited by 82
  35. 2020Meta-Learning Symmetries by Reparameterizationcited by 75
  36. 2018Deep Dynamical Modeling and Control of Unsteady Fluid Flowscited by 67
  37. 2019Hierarchically Structured Meta-learningcited by 67
  38. 2016Artistic style transfer for videoscited by 54

50 most cited of 60 All 60 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 6; 2023: 10; 2024: 9; 2025: 8; 2026: 3.

37 papers.

Sort:

By area:

  1. 2023DYffusion: A Dynamics-informed Diffusion Model for Spatiotemporal Forecastingcited by 105
  2. 2023On the Connection Between MPNN and Graph Transformercited by 60
  3. 2023A Survey on Time-Series Pre-Trained Modelscited by 52
  4. 2022Generalizing to New Physical Systems via Context-Informed Dynamics Modelcited by 49
  5. 2021Physics-Guided Deep Learning for Dynamical Systems: A Surveycited by 43
  6. 2023Training neural operators to preserve invariant measures of chaotic attractorscited by 43
  7. 2022Multi-Task Imitation Learning for Linear Dynamical Systemscited by 22
  8. 2023Physics Informed Token Transformer for Solving Partial Differential Equationscited by 21
  9. 2025DISCO: learning to DISCover an evolution Operator for multi-physics-agnostic predictioncited by 21
  10. 2022Data Augmentation vs. Equivariant Networks: A Theory of Generalization on Dynamics Forecastingcited by 17
  11. 2023Deciphering and integrating invariants for neural operator learning with various physical mechanismscited by 15
  12. 2024Zero-shot Imputation with Foundation Inference Models for Dynamical Systemscited by 13
25 more
  1. 2024SineNet: Learning Temporal Dynamics in Time-Dependent Partial Differential Equationscited by 13
  2. 2023Interpretable Meta-Learning of Physical Systemscited by 11
  3. 2024Masked Autoencoders are PDE Learnerscited by 11
  4. 2023MetaNO: How to Transfer Your Knowledge on Learning Hidden Physicscited by 10
  5. 2024Neural Context Flows for Meta-Learning of Dynamical Systemscited by 9
  6. 2026Test-time Generalization for Physics through Neural Operator Splittingcited by 8
  7. 2024Meta-Dynamical State Space Models for Integrative Neural Data Analysiscited by 7
  8. 2024GEPS: Boosting Generalization in Parametric PDE Neural Solvers through Adaptive Conditioningcited by 7
  9. 2022Continual Learning of Dynamical Systems with Competitive Federated Reservoir Computingcited by 6
  10. 2023MetaPhysiCa: OOD Robustness in Physics-informed Machine Learningcited by 6
  11. 2025Discovering Physics Laws of Dynamical Systems via Invariant Function Learningcited by 5
  12. 2025Towards Foundational Models for Dynamical System Reconstruction: Hierarchical Meta-Learning via Mixture of Expertscited by 5
  13. 2022Towards Cross Domain Generalization of Hamiltonian Representation via Meta Learningcited by 3
  14. 2022Gaussian Process Latent Variable Modeling for Few-shot Time Series Forecastingcited by 3
  15. 2023Inductive biases in deep learning models for weather predictioncited by 3
  16. 2024The Power of Next-Frame Prediction for Learning Physical Lawscited by 2
  17. 2024Learning to learn ecosystems from limited data -- a meta-learning approachcited by 2
  18. 2025Tailored Forecasting from Short Time Series via Meta-learningcited by 2
  19. 2025Environment Inference for Learning Generalizable Dynamical Systemcited by 2
  20. 2025Towards Generalizable PDE Dynamics Forecasting via Physics-Guided Invariant Learningcited by 1
  21. 2024Unsupervised Learning of Hybrid Latent Dynamics: A Learn-to-Identify Frameworkcited by 0
  22. 2025Generalizing to New Dynamical Systems via Frequency Domain Adaptationcited by 0
  23. 2025Video Prediction of Dynamic Physical Simulations With Pixel-Space Spatiotemporal Transformerscited by 0
  24. 2026Generalizing Dynamics Modeling More Easily from Representation Perspectivecited by 0
  25. 2026Conservation Buys Stability and Factoring Buys Counterfactuals in Physical World Modelscited by 0