mlsanityData to 14 Oct 2026

Identifiable Exchangeable Mechanisms for Causal Structure and Representation Learning

arXiv 2406.14302 ↗First posted 20 Jun 2024stat.ML cs.AI cs.LGcited by 16

Not on the map. Why?

Abstract (arXiv)

Identifying latent representations or causal structures is important for good generalization and downstream task performance. However, both fields have been developed rather independently. We observe that several methods in both representation and causal structure learning rely on the same data-generating process (DGP), namely, exchangeable but not i.i.d. (independent and identically distributed) data. We provide a unified framework, termed Identifiable Exchangeable Mechanisms (IEM), for representation and structure learning under the lens of exchangeability. IEM provides new insights that let us relax the necessary conditions for causal structure identification in exchangeable non--i.i.d. data. We also demonstrate the existence of a duality condition in identifiable representation learning, leading to new identifiability results. We hope this work will pave the way for further research in causal representation learning.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Builds on 45 of its 86 references that are in the map. First posted on arXiv 20 Jun 2024.

Builds on

45 of 86 references.

References · 45

  1. 2019Invariant Risk Minimizationcited by 1,854
  2. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 941
  3. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 752
  4. 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 503
  5. 2018Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learningcited by 282
  6. 2016Unsupervised Feature Extraction by Time-Contrastive Learning and Nonlinear ICAcited by 273
  7. 2019Causality for Machine Learningcited by 243
  8. 2019Gradient-Based Neural DAG Learningcited by 227
  9. 2021Contrastive Learning Inverts the Data Generating Processcited by 221
  10. 2021Towards Causal Representation Learningcited by 209
  11. 2013Causal Inference in the Presence of Latent Variables and Selection Biascited by 166
  12. 2021Disentanglement via Mechanism Sparsity Regularization: A New Principle for Nonlinear ICAcited by 160
33 more
  1. 2020Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Codingcited by 122
  2. 2021Independent mechanism analysis, a new concept?cited by 97
  3. 2019Masked Gradient-Based Causal Structure Learningcited by 89
  4. 2019The Incomplete Rosetta Stone Problem: Identifiability Results for Multi-View Nonlinear ICAcited by 86
  5. 2024Robust agents learn causal world modelscited by 81
  6. 2020Hidden Markov Nonlinear ICA: Unsupervised Learning from Nonstationary Time Seriescited by 77
  7. 2022Weakly Supervised Representation Learning with Sparse Perturbationscited by 65
  8. 2021Disentangling Identifiable Features from Noisy Data with Structured Nonlinear ICAcited by 64
  9. 2021Out of Distribution Generalization in Machine Learningcited by 60
  10. 2019Causal Discovery with General Non-Linear Relationships Using Non-Linear ICAcited by 58
  11. 2023Identifiability of latent-variable and structural-equation models: from linear to nonlinearcited by 58
  12. 2023Additive Decoders for Latent Variables Identification and Cartesian-Product Extrapolationcited by 58
  13. 2022Causal Discovery in Heterogeneous Environments Under the Sparse Mechanism Shift Hypothesiscited by 55
  14. 2018Structural Agnostic Modeling: Adversarial Learning of Causal Graphscited by 48
  15. 2023Scalable Causal Discovery with Score Matchingcited by 45
  16. 2021Variational Causal Networks: Approximate Bayesian Inference over Causal Structurescited by 43
  17. 2023Causal Discovery with Score Matching on Additive Models with Arbitrary Noisecited by 43
  18. 2022Causal de Finetti: On the Identification of Invariant Causal Structure in Exchangeable Datacited by 42
  19. 2021Relating Graph Neural Networks to Structural Causal Modelscited by 41
  20. 2023Provable Compositional Generalization for Object-Centric Learningcited by 41
  21. 2023Leveraging sparse and shared feature activations for disentangled representation learningcited by 32
  22. 2023Probabilistic Contrastive Learning Recovers the Correct Aleatoric Uncertainty of Ambiguous Inputscited by 29
  23. 2022Discovering Agentscited by 28
  24. 2023Object-centric architectures enable efficient causal representation learningcited by 21
  25. 2020Independent Innovation Analysis for Nonlinear Vector Autoregressive Processcited by 19
  26. 2023Self-Supervised Disentanglement by Leveraging Structure in Data Augmentationscited by 17
  27. 2023Learning Causal Representations from General Environments: Identifiability and Intrinsic Ambiguitycited by 15
  28. 2018Direct Estimation of Differences in Causal Graphscited by 14
  29. 2024Do Finetti: On Causal Effects for Exchangeable Datacited by 11
  30. 2023An Interventional Perspective on Identifiability in Gaussian LTI Systems with Independent Component Analysiscited by 9
  31. 2022On Finite-Sample Identifiability of Contrastive Learning-Based Nonlinear Independent Component Analysiscited by 7
  32. 2024Towards Identifiable Latent Additive Noise Modelscited by 5
  33. 2023Disentanglement of Latent Representations via Causal Interventionscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 9; 2026: 6.

16 papers.

Sort:

By area:

  1. 2025On the Value of Cross-Modal Misalignment in Multimodal Representation Learningcited by 10
  2. 2025Position: An Empirically Grounded Identifiability Theory Will Accelerate Self-Supervised Learning Researchcited by 9
  3. 2026Causality is Key for Interpretability Claims to Generalisecited by 8
  4. 2025From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?cited by 7
  5. 2025Multi-Domain Causal Discovery in Bijective Causal Modelscited by 4
  6. 2025Counterfactual reasoning: an analysis of in-context emergencecited by 4
  7. 2025Skill Learning via Policy Diversity Yields Identifiable Representations for Reinforcement Learningcited by 4
  8. 2024Analyzing Generative Models by Manifold Entropic Metricscited by 2
  9. 2025Probabilistic Modelling is Sufficient for Causal Inferencecited by 2
  10. 2026The Geometric Mechanics of Contrastive Representation Learning: Alignment Potentials, Entropic Dispersion, and Cross-modal Divergencecited by 2
  11. 2025A Roadmap Towards Improving Multi-Agent Reinforcement Learning With Causal Discovery And Inferencecited by 1
  12. 2026On the Emergence and Test-Time Use of Structural Information in Large Language Modelscited by 1
4 more
  1. 2026Concept Modulation Models: A Unified Framework for Identifiability and Extrapolationcited by 1
  2. 2025Estimating Treatment Effects with Independent Component Analysiscited by 0
  3. 2026Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shiftscited by 0
  4. 2026Self-Reports Do Not Identify Self-Models: An Identifiability Test for Counterfactual Reportscited by 0