mlsanityData to 7 Oct 2026

Identifiability Guarantees for Causal Disentanglement from Purely Observational Data

arXiv 2410.23620 ↗First posted 31 Oct 2024cs.LG stat.MLcited by 7

Not on the map. Why?

Abstract (arXiv)

Causal disentanglement aims to learn about latent causal factors behind data, holding the promise to augment existing representation learning methods in terms of interpretability and extrapolation. Recent advances establish identifiability results assuming that interventions on (single) latent factors are available; however, it remains debatable whether such assumptions are reasonable due to the inherent nature of intervening on latent variables. Accordingly, we reconsider the fundamentals and ask what can be learned using just observational data. We provide a precise characterization of latent factors that can be identified in nonlinear causal models with additive Gaussian noise and linear mixing, without any interventions or graphical restrictions. In particular, we show that the causal variables can be identified up to a layer-wise transformation and that further disentanglement is not possible. We transform these theoretical results into a practical algorithm consisting of solving a quadratic program over the score estimation of the observed data. We provide simulation results to support our theoretical guarantees and demonstrate that our algorithm can derive meaningful causal representations from purely observational data.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 38 of its 60 references that are in the map. First posted on arXiv 31 Oct 2024.

Builds on

38 of 60 references.

References · 38

  1. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  2. 2019Sliced Score Matching: A Scalable Approach to Density and Score Estimationcited by 381
  3. 2013CAM: Causal additive models, high-dimensional order search and penalized regressioncited by 175
  4. 2022Weakly supervised causal representation learningcited by 140
  5. 2022Interventional Causal Representation Learningcited by 140
  6. 2022CITRIS: Causal Identifiability from Temporal Intervened Sequencescited by 132
  7. 2022Score matching enables causal discovery of nonlinear additive noise modelscited by 108
  8. 2022Causal Machine Learning: A Survey and Open Problemscited by 107
  9. 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 105
  10. 2023Learning Linear Causal Representations from Interventions under General Nonlinear Mixingcited by 92
  11. 2022Linear Causal Disentanglement via Interventionscited by 84
  12. 2023Identifiability Guarantees for Causal Disentanglement from Soft Interventionscited by 84
26 more
  1. 2023Multi-View Causal Representation Learning with Partial Observabilitycited by 72
  2. 2022Identifiability of deep generative models without auxiliary informationcited by 70
  3. 2022Weakly Supervised Representation Learning with Sparse Perturbationscited by 64
  4. 2022Diffusion Models for Causal Discovery via Topological Orderingcited by 64
  5. 2021Estimating High Order Gradients of the Data Distribution by Denoisingcited by 63
  6. 2023Causal Component Analysiscited by 56
  7. 2021Learning latent causal graphs via mixture oraclescited by 55
  8. 2020Efficient Learning of Generative Models via Finite-Difference Score Matchingcited by 49
  9. 2023Score-based Causal Representation Learning with Interventionscited by 49
  10. 2023Scalable Causal Discovery with Score Matchingcited by 42
  11. 2023Causal Discovery with Score Matching on Additive Models with Arbitrary Noisecited by 41
  12. 2023Learning nonparametric latent causal graphs with unknown interventionscited by 34
  13. 2023General Identifiability and Achievability for Causal Representation Learningcited by 34
  14. 2024A Sparsity Principle for Partially Observable Causal Representation Learningcited by 34
  15. 2022Identifying Weight-Variant Latent Causal Modelscited by 33
  16. 2023Unpaired Multi-Domain Causal Representation Learningcited by 32
  17. 2024Learning Interpretable Concepts: Unifying Causal Representation Learning and Foundation Modelscited by 29
  18. 2023Identification of Nonlinear Latent Hierarchical Modelscited by 27
  19. 2024Score-based Causal Representation Learning: Linear and General Transformationscited by 27
  20. 2023Multi-Domain Causal Representation Learning via Weak Distributional Invariancescited by 23
  21. 2023Identifying Representations for Intervention Extrapolationcited by 21
  22. 2023Identifying Linearly-Mixed Causal Representations from Multi-Node Interventionscited by 18
  23. 2023Interpretable Reward Redistribution in Reinforcement Learning: A Causal Approachcited by 17
  24. 2023Sample Complexity Bounds for Score-Matching: Causal Discovery and Generative Modelingcited by 15
  25. 2015Anchored Discrete Factor Analysiscited by 9
  26. 2021On the Sample Complexity of Causal Discovery and the Value of Domain Expertisecited by 9

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 4; 2026: 3.

7 papers.

  1. 2025What is causal about causal models and representations?cited by 6
  2. 2025Causal Structure and Representation Learning with Biomedical Applicationscited by 3
  3. 2025Linear Causal Representation Learning by Topological Ordering, Pruning, and Disentanglementcited by 0
  4. 2025Towards Unsupervised Causal Representation Learning via Latent Additive Noise Model Causal Autoencoderscited by 0
  5. 2026Temporal Fusion Nexus: A task-agnostic multi-modal embedding model for clinical narratives and irregular time series in post-kidney transplant carecited by 0
  6. 2026A Dialogue between Causal and Traditional Representation Learning: Toward Mutual Benefits in a Unified Formulationcited by 0
  7. 2026Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shiftscited by 0