Unsupervised Learning of Hybrid Latent Dynamics: A Learn-to-Identify Framework
Not on the map. Why?
Abstract (arXiv)
Modern applications increasingly require unsupervised learning of latent dynamics from high-dimensional time-series. This presents a significant challenge of identifiability: many abstract latent representations may reconstruct observations, yet do they guarantee an adequate identification of the governing dynamics? This paper investigates this challenge from two angles: the use of physics inductive bias specific to the data being modeled, and a learn-to-identify strategy that separates forecasting objectives from the data used for the identification. We combine these two strategies in a novel framework for unsupervised meta-learning of hybrid latent dynamics (Meta-HyLaD) with: 1) a latent dynamic function that hybridize known mathematical expressions of prior physics with neural functions describing its unknown errors, and 2) a meta-learning formulation to learn to separately identify both components of the hybrid dynamics. Through extensive experiments on five physics and one biomedical systems, we provide strong evidence for the benefits of Meta-HyLaD to integrate rich prior knowledge while identifying their gap to observed data.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 29 of its 39 references that are in the map. First posted on arXiv 13 Mar 2024.
Builds on
29 of 39 references.
References · 29
- 2017Learning to Compare: Relation Network for Few-Shot Learningcited by 1,143
- 2019Hamiltonian Neural Networkscited by 712
- 2019Learning agile and dynamic motor skills for legged robotscited by 504
- 2020Lagrangian Neural Networkscited by 398
- 2015A Recurrent Latent Variable Model for Sequential Datacited by 362
- 2020Universal Differential Equations for Scientific Machine Learningcited by 353
- 2019Deep Lagrangian Networks: Using Physics as Model Prior for Deep Learningcited by 212
- 2019Symplectic ODE-Net: Learning Hamiltonian Dynamics with Controlcited by 184
- 2016Structured Inference Networks for Nonlinear State Space Modelscited by 182
- 2016Deep Variational Bayes Filters: Unsupervised Learning of State Space Models from Raw Datacited by 156
- 2019TossingBot: Learning to Throw Arbitrary Objects with Residual Physicscited by 137
- 2019Hamiltonian Generative Networkscited by 137
17 more
- 2020Dynamical Variational Autoencoders: A Comprehensive Reviewcited by 128
- 2017A Disentangled Recognition and Nonlinear Dynamics Model for Unsupervised Learningcited by 107
- 2021gradSim: Differentiable simulation for system identification and visuomotor controlcited by 100
- 2020NeuralSim: Augmenting Differentiable Simulators with Neural Networkscited by 82
- 2020Meta-learning framework with applications to zero-shot time-series forecastingcited by 70
- 2022Generalizing to New Physical Systems via Context-Informed Dynamics Modelcited by 50
- 2019Physics-as-Inverse-Graphics: Unsupervised Physical Parameter Estimation from Videocited by 45
- 2021LEADS: Learning Dynamical Systems that Generalize Across Environmentscited by 40
- 2020Unsupervised Learning of Lagrangian Dynamics from Images for Prediction and Controlcited by 38
- 2021Meta-Learning Dynamics Forecasting Using Task Inferencecited by 38
- 2019Variational Integrator Networks for Physically Structured Embeddingscited by 36
- 2021Which priors matter? Benchmarking models for learning latent dynamicscited by 26
- 2020LagNetViP: A Lagrangian Neural Network for Video Predictioncited by 24
- 2020Forecasting Hamiltonian dynamics without canonical coordinatescited by 17
- 2021Stateful ODE-Nets using Basis Function Expansionscited by 14
- 2019A General Framework for Structured Learning of Mechanical Systemscited by 13
- 2021SyMetric: Measuring the Quality of Learnt Hamiltonian Dynamics Inferred from Visioncited by 5
Built on by
Not cited yet.