Identifiable Representation and Model Learning for Latent Dynamic Systems
Not on the map. Why?
Abstract (arXiv)
Learning identifiable representations and models from low-level observations is helpful for an intelligent spacecraft to complete downstream tasks reliably. For temporal observations, to ensure that the data generating process is provably inverted, most existing works either assume the noise variables in the dynamic mechanisms are (conditionally) independent or require that the interventions can directly affect each latent variable. However, in practice, the relationship between the exogenous inputs/interventions and the latent variables may follow some complex deterministic mechanisms. In this work, we study the problem of identifiable representation and model learning for latent dynamic systems. The key idea is to use an inductive bias inspired by controllable canonical forms, which are sparse and input-dependent by definition. We prove that, for linear and affine nonlinear latent dynamic systems with sparse input matrices, it is possible to identify the latent variables up to scaling and determine the dynamic models up to some simple transformations. The results have the potential to provide some theoretical guarantees for developing more trustworthy decision-making and control methods for intelligent spacecrafts.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 39 of its 54 references that are in the map. First posted on arXiv 23 Oct 2024.
Builds on
39 of 54 references.
References · 39
- 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 1,653
- 2018Learning Latent Dynamics for Planning from Pixelscited by 1,264
- 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 941
- 2020CURL: Contrastive Unsupervised Representations for Reinforcement Learningcited by 760
- 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 503
- 2020Learning Invariant Representations for Reinforcement Learning without Reconstructioncited by 436
- 2016Reinforcement Learning with Unsupervised Auxiliary Taskscited by 404
- 2022DayDreamer: World Models for Physical Robot Learningcited by 400
- 2020Data-Efficient Reinforcement Learning with Self-Predictive Representationscited by 318
- 2015Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Imagescited by 274
- 2016Unsupervised Feature Extraction by Time-Contrastive Learning and Nonlinear ICAcited by 273
- 2020Weakly-Supervised Disentanglement Without Compromisescited by 239
27 more
- 2021Contrastive Learning Inverts the Data Generating Processcited by 221
- 2019DeepMDP: Learning Continuous Latent Space Models for Representation Learningcited by 207
- 2021Disentanglement via Mechanism Sparsity Regularization: A New Principle for Nonlinear ICAcited by 160
- 2022Interventional Causal Representation Learningcited by 142
- 2022Weakly supervised causal representation learningcited by 140
- 2022CITRIS: Causal Identifiability from Temporal Intervened Sequencescited by 133
- 2020Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Codingcited by 122
- 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 106
- 2021Learning Temporally Causal Latent Processes from General Temporal Datacited by 98
- 2020On Disentangled Representations Learned From Correlated Datacited by 94
- 2022Linear Causal Disentanglement via Interventionscited by 84
- 2023Identifiability Guarantees for Causal Disentanglement from Soft Interventionscited by 84
- 2020Hidden Markov Nonlinear ICA: Unsupervised Learning from Nonstationary Time Seriescited by 77
- 2022Identifiability of deep generative models without auxiliary informationcited by 70
- 2020Weakly Supervised Disentangled Generative Causal Representation Learningcited by 68
- 2022Denoised MDPs: Learning World Models Better Than the World Itselfcited by 68
- 2022Weakly Supervised Representation Learning with Sparse Perturbationscited by 65
- 2022Temporally Disentangled Representation Learningcited by 61
- 2023Causal Component Analysiscited by 56
- 2022Causal Representation Learning for Instantaneous and Temporal Effects in Interactive Systemscited by 51
- 2023Nonlinear Independent Component Analysis for Principled Disentanglement in Unsupervised Deep Learningcited by 48
- 2024Nonparametric Partial Disentanglement via Mechanism Sparsity: Sparse Actions, Interventions and Sparse Temporal Dependenciescited by 41
- 2023General Identifiability and Achievability for Causal Representation Learningcited by 34
- 2021Action-Sufficient State Representation Learning for Control with Structural Constraintscited by 31
- 2021Properties from Mechanisms: An Equivariance Perspective on Identifiable Representation Learningcited by 30
- 2021Learning Representations for Pixel-based Control: What Matters and Why?cited by 24
- 2023Learning World Models with Identifiable Factorizationcited by 23
Built on by
Citing papers per year: 2026: 1.
1 paper.