mlsanityData to 14 Oct 2026

General Identifiability and Achievability for Causal Representation Learning

arXiv 2310.15450 ↗First posted 24 Oct 2023cs.LG stat.MLcited by 34

Not on the map. Why?

Abstract (arXiv)

This paper focuses on causal representation learning (CRL) under a general nonparametric latent causal model and a general transformation model that maps the latent data to the observational data. It establishes identifiability and achievability results using two hard uncoupled interventions per node in the latent causal graph. Notably, one does not know which pair of intervention environments have the same node intervened (hence, uncoupled). For identifiability, the paper establishes that perfect recovery of the latent causal model and variables is guaranteed under uncoupled interventions. For achievability, an algorithm is designed that uses observational and interventional data and recovers the latent causal model and variables with provable guarantees. This algorithm leverages score variations across different environments to estimate the inverse of the transformer and, subsequently, the latent variables. The analysis, additionally, recovers the identifiability result for two hard coupled interventions, that is when metadata about the pair of environments that have the same node intervened is known. This paper also shows that when observational data is available, additional faithfulness assumptions that are adopted by the existing literature are unnecessary.

Key facts. Cited by 34 papers in the map (data through 14 Oct 2026). Builds on 28 of its 48 references that are in the map. First posted on arXiv 24 Oct 2023.

Builds on

28 of 48 references.

References · 28

  1. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 941
  2. 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 503
  3. 2019Sliced Score Matching: A Scalable Approach to Density and Score Estimationcited by 384
  4. 2021Self-Supervised Learning with Data Augmentations Provably Isolates Content from Stylecited by 312
  5. 2018Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learningcited by 282
  6. 2020Weakly-Supervised Disentanglement Without Compromisescited by 239
  7. 2021Disentanglement via Mechanism Sparsity Regularization: A New Principle for Nonlinear ICAcited by 160
  8. 2022Interventional Causal Representation Learningcited by 142
  9. 2022Weakly supervised causal representation learningcited by 140
  10. 2022Score matching enables causal discovery of nonlinear additive noise modelscited by 111
  11. 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 106
  12. 2020ICE-BeeM: Identifiable Conditional Energy-Based Deep Models Based on Nonlinear ICAcited by 102
16 more
  1. 2021Learning Temporally Causal Latent Processes from General Temporal Datacited by 98
  2. 2023Learning Linear Causal Representations from Interventions under General Nonlinear Mixingcited by 92
  3. 2022Linear Causal Disentanglement via Interventionscited by 84
  4. 2023Identifiability Guarantees for Causal Disentanglement from Soft Interventionscited by 84
  5. 2019Weakly Supervised Disentanglement with Guaranteescited by 82
  6. 2020Hidden Markov Nonlinear ICA: Unsupervised Learning from Nonstationary Time Seriescited by 77
  7. 2022Diffusion Models for Causal Discovery via Topological Orderingcited by 66
  8. 2022Weakly Supervised Representation Learning with Sparse Perturbationscited by 65
  9. 2023Causal Component Analysiscited by 56
  10. 2022Causal Representation Learning for Instantaneous and Temporal Effects in Interactive Systemscited by 51
  11. 2023Score-based Causal Representation Learning with Interventionscited by 49
  12. 2023Scalable Causal Discovery with Score Matchingcited by 45
  13. 2020On Implicit Regularization in β-VAEscited by 34
  14. 2022Identifying Weight-Variant Latent Causal Modelscited by 34
  15. 2023Learning nonparametric latent causal graphs with unknown interventionscited by 34
  16. 2021Properties from Mechanisms: An Equivariance Perspective on Identifiable Representation Learningcited by 30

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 1; 2024: 11; 2025: 14; 2026: 8.

34 papers.

Sort:

By area:

  1. 2024Unifying Causal Representation Learning with the Invariance Principlecited by 31
  2. 2024Marrying Causal Representation Learning with Dynamical Systems for Sciencecited by 23
  3. 2024Interaction Asymmetry: A General Principle for Learning Composable Abstractionscited by 20
  4. 2023Learning Causal Representations from General Environments: Identifiability and Intrinsic Ambiguitycited by 15
  5. 2024Linear Causal Representation Learning from Unknown Multi-node Interventionscited by 14
  6. 2024All or None: Identifiable Linear Properties of Next-token Predictors in Language Modelingcited by 14
  7. 2025Towards Interpretable Deep Generative Models via Causal Representation Learningcited by 13
  8. 2024Identifiable Causal Representation Learning: Unsupervised, Multi-View, and Multi-Environmentcited by 12
  9. 2025Prediction-Powered Causal Inferencescited by 9
  10. 2026On the Identifiability of Controlled World Modelscited by 9
  11. 2024Identifiability Guarantees for Causal Disentanglement from Purely Observational Datacited by 7
  12. 2025What is causal about causal models and representations?cited by 6
22 more
  1. 2024Towards Identifiable Latent Additive Noise Modelscited by 5
  2. 2024Identifying General Mechanism Shifts in Linear Causal Representationscited by 4
  3. 2025The Third Pillar of Causal Analysis? A Measurement Perspective on Causal Representationscited by 4
  4. 2025When Does Closeness in Distribution Imply Representational Similarity? An Identifiability Perspectivecited by 4
  5. 2025Exploratory Causal Inference in SAEncecited by 4
  6. 2025Contextures: Representations from Contextscited by 3
  7. 2025From Pixels to Factors: Learning Independently Controllable State Variables for Reinforcement Learningcited by 3
  8. 2024Linear causal disentanglement via higher-order cumulantscited by 1
  9. 2024Identifiable Representation and Model Learning for Latent Dynamic Systemscited by 1
  10. 2025Intervening to Learn and Compose Causally Disentangled Representationscited by 1
  11. 2026Multi-Domain Empirical Bayes for Linearly-Mixed Causal Representationscited by 1
  12. 2026Concept Modulation Models: A Unified Framework for Identifiability and Extrapolationcited by 1
  13. 2026Catastrophic Compositional Generation: Why Vanilla Diffusion Models Fail to Extrapolatecited by 1
  14. 2025A Generative Framework for Causal Estimation via Importance-Weighted Diffusion Distillationcited by 0
  15. 2025Learning Joint Interventional Effects from Single-Variable Interventions in Additive Modelscited by 0
  16. 2025Incorporating Interventional Independence Improves Robustness against Interventional Distribution Shiftcited by 0
  17. 2025Linear Causal Representation Learning by Topological Ordering, Pruning, and Disentanglementcited by 0
  18. 2025ROPES: Robotic Pose Estimation via Score-Based Causal Representation Learningcited by 0
  19. 2026From Causal Graphs to Markovian States: Multi-Order State Exposure for Deep RLcited by 0
  20. 2026A Dialogue between Causal and Traditional Representation Learning: Toward Mutual Benefits in a Unified Formulationcited by 0
  21. 2026Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shiftscited by 0
  22. 2026Jigsaw-CRL: Recovering Global Latent Causal Order from Fragmented Multi-Client Interventionscited by 0