Linear causal disentanglement via higher-order cumulants
Not on the map. Why?
Abstract (arXiv)
Linear causal disentanglement is a recent method in causal representation learning to describe a collection of observed variables via latent variables with causal dependencies between them. It can be viewed as a generalization of both independent component analysis and linear structural equation models. We study the identifiability of linear causal disentanglement, assuming access to data under multiple contexts, each given by an intervention on a latent variable. We show that one perfect intervention on each latent variable is sufficient and in the worst case necessary to recover parameters under perfect interventions, generalizing previous work to allow more latent than observed variables. We give a constructive proof that computes parameters via a coupled tensor decomposition. For soft interventions, we find the equivalence class of latent graphs and parameters that are consistent with observed data, via the study of a system of polynomial equations. Our results hold assuming the existence of non-zero higher-order cumulants, which implies non-Gaussianity of variables.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 27 of its 72 references that are in the map. First posted on arXiv 5 Jul 2024.
Builds on
27 of 72 references.
References · 27
- 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 941
- 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 503
- 2021Contrastive Learning Inverts the Data Generating Processcited by 221
- 2022Interventional Causal Representation Learningcited by 142
- 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 106
- 2023Learning Linear Causal Representations from Interventions under General Nonlinear Mixingcited by 92
- 2022Linear Causal Disentanglement via Interventionscited by 84
- 2023Identifiability Guarantees for Causal Disentanglement from Soft Interventionscited by 84
- 2024Causal Representation Learning from Multiple Distributions: A General Settingcited by 68
- 2023Additive Decoders for Latent Variables Identification and Cartesian-Product Extrapolationcited by 58
- 2021Learning latent causal graphs via mixture oraclescited by 55
- 2021Identifiable Deep Generative Models via Sparse Decodingcited by 54
15 more
- 2023Score-based Causal Representation Learning with Interventionscited by 49
- 2024Nonparametric Partial Disentanglement via Mechanism Sparsity: Sparse Actions, Interventions and Sparse Temporal Dependenciescited by 41
- 2022Identifying Weight-Variant Latent Causal Modelscited by 34
- 2023Learning nonparametric latent causal graphs with unknown interventionscited by 34
- 2023General Identifiability and Achievability for Causal Representation Learningcited by 34
- 2021Properties from Mechanisms: An Equivariance Perspective on Identifiable Representation Learningcited by 30
- 2018High-Dimensional Causal Discovery Under non-Gaussianitycited by 24
- 2023Identifiable Latent Polynomial Causal Models Through the Lens of Changecited by 23
- 2023Targeted Reduction of Causal Modelscited by 15
- 2019Subspace power method for symmetric tensor decompositioncited by 14
- 2020A Commentary on the Unsupervised Learning of Disentangled Representationscited by 13
- 2015Anchored Discrete Factor Analysiscited by 9
- 2024Propensity Score Alignment of Unpaired Multimodal Datacited by 6
- 2023Invariance & Causal Representation Learning: Prospects and Limitationscited by 5
- 2024Towards Identifiable Latent Additive Noise Modelscited by 5
Built on by
Citing papers per year: 2026: 1.
1 paper.