mlsanityData to 14 Oct 2026

Causal Representation Learning Made Identifiable by Grouping of Observational Variables

arXiv 2310.15709 ↗First posted 24 Oct 2023stat.ML cs.LGcited by 32

Not on the map. Why?

Abstract (arXiv)

A topic of great current interest is Causal Representation Learning (CRL), whose goal is to learn a causal model for hidden features in a data-driven manner. Unfortunately, CRL is severely ill-posed since it is a combination of the two notoriously ill-posed problems of representation learning and causal discovery. Yet, finding practical identifiability conditions that guarantee a unique solution is crucial for its practical applicability. Most approaches so far have been based on assumptions on the latent causal mechanisms, such as temporal causality, or existence of supervision or interventions; these can be too restrictive in actual applications. Here, we show identifiability based on novel, weak constraints, which requires no temporal structure, intervention, nor weak supervision. The approach is based on assuming the observational mixing exhibits a suitable grouping of the observational variables. We also propose a novel self-supervised estimation framework consistent with the model, prove its statistical consistency, and experimentally show its superior CRL performances compared to the state-of-the-art baselines. We further demonstrate its robustness against latent confounders and causal cycles.

Key facts. Cited by 32 papers in the map (data through 14 Oct 2026). Builds on 58 of its 111 references that are in the map. First posted on arXiv 24 Oct 2023.

Builds on

58 of 111 references.

References · 58

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 941
  3. 2017AI2-THOR: An Interactive 3D Environment for Visual AIcited by 869
  4. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 752
  5. 2018Disentangling by Factorisingcited by 616
  6. 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 503
  7. 2021Self-Supervised Learning with Data Augmentations Provably Isolates Content from Stylecited by 312
  8. 2018Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learningcited by 282
  9. 2016Unsupervised Feature Extraction by Time-Contrastive Learning and Nonlinear ICAcited by 273
  10. 2013Causal Discovery with Continuous Additive Noise Modelscited by 255
  11. 2020Weakly-Supervised Disentanglement Without Compromisescited by 239
  12. 2021Contrastive Learning Inverts the Data Generating Processcited by 221
38 more
  1. 2016Variational Deep Embedding: An Unsupervised and Generative Approach to Clusteringcited by 220
  2. 2021Towards Causal Representation Learningcited by 209
  3. 2019Learning Sparse Nonparametric DAGscited by 207
  4. 2020On the Role of Sparsity and DAG Constraints for Learning Linear DAGscited by 191
  5. 2013CAM: Causal additive models, high-dimensional order search and penalized regressioncited by 178
  6. 2021Beware of the Simulated DAG! Causal Discovery Benchmarks May Be Easy To Gamecited by 161
  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. 2022CITRIS: Causal Identifiability from Temporal Intervened Sequencescited by 133
  11. 2020Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Codingcited by 122
  12. 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 106
  13. 2020On Linear Identifiability of Learned Representationscited by 105
  14. 2020ICE-BeeM: Identifiable Conditional Energy-Based Deep Models Based on Nonlinear ICAcited by 102
  15. 2021Learning Temporally Causal Latent Processes from General Temporal Datacited by 98
  16. 2021Independent mechanism analysis, a new concept?cited by 97
  17. 2020Disentanglement by Nonlinear ICA with General Incompressible-flow Networks (GIN)cited by 92
  18. 2023Learning Linear Causal Representations from Interventions under General Nonlinear Mixingcited by 92
  19. 2019The Incomplete Rosetta Stone Problem: Identifiability Results for Multi-View Nonlinear ICAcited by 86
  20. 2022On the Identifiability of Nonlinear ICA: Sparsity and Beyondcited by 85
  21. 2022Linear Causal Disentanglement via Interventionscited by 84
  22. 2020Hidden Markov Nonlinear ICA: Unsupervised Learning from Nonstationary Time Seriescited by 77
  23. 2020Causal Discovery in Physical Systems from Videoscited by 77
  24. 2022Identifiability of deep generative models without auxiliary informationcited by 70
  25. 2020Weakly Supervised Disentangled Generative Causal Representation Learningcited by 68
  26. 2022Weakly Supervised Representation Learning with Sparse Perturbationscited by 65
  27. 2021Disentangling Identifiable Features from Noisy Data with Structured Nonlinear ICAcited by 64
  28. 2013Learning Gaussian Networkscited by 62
  29. 2023Identifiability Results for Multimodal Contrastive Learningcited by 61
  30. 2019Causal Discovery with General Non-Linear Relationships Using Non-Linear ICAcited by 58
  31. 2023Self-Supervised Multimodal Learning: A Surveycited by 58
  32. 2023Posterior Collapse and Latent Variable Non-identifiabilitycited by 57
  33. 2021Learning latent causal graphs via mixture oraclescited by 55
  34. 2021Identifiable Deep Generative Models via Sparse Decodingcited by 54
  35. 2022Causal Representation Learning for Instantaneous and Temporal Effects in Interactive Systemscited by 51
  36. 2023Score-based Causal Representation Learning with Interventionscited by 49
  37. 2013Density Estimation in Infinite Dimensional Exponential Familiescited by 42
  38. 2021Understanding Latent Correlation-Based Multiview Learning and Self-Supervision: An Identifiability Perspectivecited by 37

50 most cited of 58 All 58 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 1; 2024: 4; 2025: 14; 2026: 13.

32 papers.

Sort:

By area:

  1. 2023Multi-View Causal Representation Learning with Partial Observabilitycited by 74
  2. 2025Everything, Everywhere, All at Once: Is Mechanistic Interpretability Identifiable?cited by 34
  3. 2024On the Identification of Temporally Causal Representation with Instantaneous Dependencecited by 16
  4. 2024Linear Causal Representation Learning from Unknown Multi-node Interventionscited by 14
  5. 2025SmartCLIP: Modular Vision-language Alignment with Identification Guaranteescited by 14
  6. 2024Causal Representation Learning from Multimodal Biomedical Observationscited by 13
  7. 2025Position: An Empirically Grounded Identifiability Theory Will Accelerate Self-Supervised Learning Researchcited by 9
  8. 2025Towards Identifiability of Hierarchical Temporal Causal Representation Learningcited by 9
  9. 2025Learning General Causal Structures with Hidden Dynamic Process for Climate Analysiscited by 7
  10. 2025Mechanistic Interpretability Needs Philosophycited by 6
  11. 2025LLM Interpretability with Identifiable Temporal-Instantaneous Representationcited by 6
  12. 2026Causal Representation Learning from General Environments under Nonparametric Mixingcited by 5
20 more
  1. 2025Synergy Between Sufficient Changes and Sparse Mixing Procedure for Disentangled Representation Learningcited by 4
  2. 2025Modeling Unseen Environments with Language-guided Composable Causal Components in Reinforcement Learningcited by 4
  3. 2025Diverse Influence Component Analysis: A Geometric Approach to Nonlinear Mixture Identifiabilitycited by 4
  4. 2026Distribution Shift Is Key to Learning Invariant Predictioncited by 2
  5. 2024Discovering Latent Causal Graphs from Spatiotemporal Datacited by 1
  6. 2025Causal Representation Learning with Observational Grouping for CXR Classificationcited by 1
  7. 2025Intervening to Learn and Compose Causally Disentangled Representationscited by 1
  8. 2025Understanding Catastrophic Interference: On the Identifibility of Latent Representationscited by 1
  9. 2026Identifiability of Potentially Degenerate Gaussian Mixture Models With Piecewise Affine Mixingcited by 1
  10. 2025Operationalizing Quantized Disentanglementcited by 0
  11. 2026CURVE: Learning Causality-Inspired Invariant Representations for Robust Scene Understanding via Uncertainty-Guided Regularizationcited by 0
  12. 2026Who Guards the Guardians? The Challenges of Evaluating Identifiability of Learned Representationscited by 0
  13. 2026Diverse Dictionary Learningcited by 0
  14. 2026Identifiable Multimodal Causal Representation Learning under Partial Latent Sharingcited by 0
  15. 2026Position: Prioritize Identifying Structure, Not Complex Models, for Scientific Discoverycited by 0
  16. 2026Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shiftscited by 0
  17. 2026MoVA: Learning Asymmetric Dual Projections for Modular Long Video-Text Alignmentcited by 0
  18. 2026Equilibrium Causal Games: Separation, Identification, and the Identifiability of Cyclic Latent Statescited by 0
  19. 2026Causal Representation Learning with Instantaneous and Lagged Relations via Nonstationaritycited by 0
  20. 2026DSReg: Provably Recovering Individual World Latents without Reconstructioncited by 0