mlsanityData to 7 Oct 2026

Beyond identifiability: Learning causal representations with few environments and finite samples

arXiv 2603.25796 ↗First posted 26 Mar 2026stat.ML cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

We provide explicit, finite-sample guarantees for learning causal representations from data with a sublinear number of environments. Causal representation learning seeks to provide a rigourous foundation for the general representation learning problem by bridging causal models with latent factor models in order to learn interpretable representations with causal semantics. Despite a blossoming theory of identifiability in causal representation learning, estimation and finite-sample bounds are less well understood. We show that causal representations can be learned with only a logarithmic number of unknown, multi-node interventions, and that the intervention targets need not be carefully designed in advance. Through a careful perturbation analysis, we provide a new analysis of this problem that guarantees consistent recovery of (a) the latent causal graph, (b) the mixing matrix and representations, and (c) unknown intervention targets.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 35 of its 68 references that are in the map. First posted on arXiv 26 Mar 2026.

Builds on

35 of 68 references.

References · 35

  1. 2015Causal inference using invariant prediction: identification and confidence intervalscited by 533
  2. 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 494
  3. 2016Unsupervised Feature Extraction by Time-Contrastive Learning and Nonlinear ICAcited by 270
  4. 2020The Risks of Invariant Risk Minimizationcited by 262
  5. 2022Interventional Causal Representation Learningcited by 140
  6. 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 105
  7. 2023Learning Linear Causal Representations from Interventions under General Nonlinear Mixingcited by 92
  8. 2022On the Identifiability of Nonlinear ICA: Sparsity and Beyondcited by 84
  9. 2022Linear Causal Disentanglement via Interventionscited by 84
  10. 2022Identifiability of deep generative models without auxiliary informationcited by 70
  11. 2021Learning latent causal graphs via mixture oraclescited by 55
  12. 2021Identifiable Deep Generative Models via Sparse Decodingcited by 54
23 more
  1. 2022Latent Hierarchical Causal Structure Discovery with Rank Constraintscited by 50
  2. 2023BISCUIT: Causal Representation Learning from Binary Interactionscited by 38
  3. 2015backShift: Learning causal cyclic graphs from unknown shift interventionscited by 37
  4. 2023Learning nonparametric latent causal graphs with unknown interventionscited by 34
  5. 2021Iterative Feature Matching: Toward Provable Domain Generalization with Logarithmic Environmentscited by 30
  6. 2023Multi-Domain Causal Representation Learning via Weak Distributional Invariancescited by 23
  7. 2017Adaptive Estimation in Structured Factor Models with Applications to Overlapping Clusteringcited by 19
  8. 2023Identifying Linearly-Mixed Causal Representations from Multi-Node Interventionscited by 18
  9. 2023Environment Invariant Linear Least Squarescited by 17
  10. 2023Learning Causal Representations from General Environments: Identifiability and Intrinsic Ambiguitycited by 14
  11. 2023Large Dimensional Independent Component Analysis: Statistical Optimality and Computational Tractabilitycited by 12
  12. 2025Towards Interpretable Deep Generative Models via Causal Representation Learningcited by 12
  13. 2022Provable concept learning for interpretable predictions using variational autoencoderscited by 9
  14. 2023Deflated HeteroPCA: Overcoming the curse of ill-conditioning in heteroskedastic PCAcited by 7
  15. 2024Causality Pursuit from Heterogeneous Environments via Neural Adversarial Invariance Learningcited by 7
  16. 2025Robustness of Nonlinear Representation Learningcited by 7
  17. 2025Sample-efficient Learning of Concepts with Theoretical Guarantees: from Data to Concepts without Interventionscited by 6
  18. 2020Optimal estimation of sparse topic modelscited by 5
  19. 2024Towards the Reusability and Compositionality of Causal Representationscited by 5
  20. 2025Discovering Hierarchical Latent Capabilities of Language Models via Causal Representation Learningcited by 4
  21. 2022Structure Learning of Latent Factors via Clique Search on Correlation Thresholded Graphscited by 3
  22. 2024Learning Discrete Latent Variable Structures with Tensor Rank Conditionscited by 3
  23. 2019Rotational Uniqueness Conditions Under Oblique Factor Correlation Metriccited by 1

Built on by

Not cited yet.